2019-01-18 02:23:39 +08:00
|
|
|
|
// SPDX-License-Identifier: GPL-2.0+
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
/*
|
2020-01-07 04:08:02 +08:00
|
|
|
|
* Read-Copy Update mechanism for mutual exclusion (tree-based version)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
*
|
|
|
|
|
* Copyright IBM Corporation, 2008
|
|
|
|
|
*
|
|
|
|
|
* Authors: Dipankar Sarma <dipankar@in.ibm.com>
|
|
|
|
|
* Manfred Spraul <manfred@colorfullife.com>
|
2020-01-07 04:08:02 +08:00
|
|
|
|
* Paul E. McKenney <paulmck@linux.ibm.com>
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
*
|
2019-01-18 02:23:39 +08:00
|
|
|
|
* Based on the original work by Paul McKenney <paulmck@linux.ibm.com>
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
* and inputs from Rusty Russell, Andrea Arcangeli and Andi Kleen.
|
|
|
|
|
*
|
|
|
|
|
* For detailed explanation of Read-Copy Update mechanism see -
|
2009-09-19 01:28:19 +08:00
|
|
|
|
* Documentation/RCU
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
*/
|
2018-05-15 04:27:33 +08:00
|
|
|
|
|
|
|
|
|
#define pr_fmt(fmt) "rcu: " fmt
|
|
|
|
|
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
#include <linux/types.h>
|
|
|
|
|
#include <linux/kernel.h>
|
|
|
|
|
#include <linux/init.h>
|
|
|
|
|
#include <linux/spinlock.h>
|
|
|
|
|
#include <linux/smp.h>
|
2017-02-06 16:50:49 +08:00
|
|
|
|
#include <linux/rcupdate_wait.h>
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
#include <linux/interrupt.h>
|
|
|
|
|
#include <linux/sched.h>
|
2017-02-09 01:51:35 +08:00
|
|
|
|
#include <linux/sched/debug.h>
|
2009-08-02 17:28:21 +08:00
|
|
|
|
#include <linux/nmi.h>
|
rcu: Avoid acquiring rcu_node locks in timer functions
This commit switches manipulations of the rcu_node ->wakemask field
to atomic operations, which allows rcu_cpu_kthread_timer() to avoid
acquiring the rcu_node lock. This should avoid the following lockdep
splat reported by Valdis Kletnieks:
[ 12.872150] usb 1-4: new high speed USB device number 3 using ehci_hcd
[ 12.986667] usb 1-4: New USB device found, idVendor=413c, idProduct=2513
[ 12.986679] usb 1-4: New USB device strings: Mfr=0, Product=0, SerialNumber=0
[ 12.987691] hub 1-4:1.0: USB hub found
[ 12.987877] hub 1-4:1.0: 3 ports detected
[ 12.996372] input: PS/2 Generic Mouse as /devices/platform/i8042/serio1/input/input10
[ 13.071471] udevadm used greatest stack depth: 3984 bytes left
[ 13.172129]
[ 13.172130] =======================================================
[ 13.172425] [ INFO: possible circular locking dependency detected ]
[ 13.172650] 2.6.39-rc6-mmotm0506 #1
[ 13.172773] -------------------------------------------------------
[ 13.172997] blkid/267 is trying to acquire lock:
[ 13.173009] (&p->pi_lock){-.-.-.}, at: [<ffffffff81032d8f>] try_to_wake_up+0x29/0x1aa
[ 13.173009]
[ 13.173009] but task is already holding lock:
[ 13.173009] (rcu_node_level_0){..-...}, at: [<ffffffff810901cc>] rcu_cpu_kthread_timer+0x27/0x58
[ 13.173009]
[ 13.173009] which lock already depends on the new lock.
[ 13.173009]
[ 13.173009]
[ 13.173009] the existing dependency chain (in reverse order) is:
[ 13.173009]
[ 13.173009] -> #2 (rcu_node_level_0){..-...}:
[ 13.173009] [<ffffffff810679b9>] check_prevs_add+0x8b/0x104
[ 13.173009] [<ffffffff81067da1>] validate_chain+0x36f/0x3ab
[ 13.173009] [<ffffffff8106846b>] __lock_acquire+0x369/0x3e2
[ 13.173009] [<ffffffff81068a0f>] lock_acquire+0xfc/0x14c
[ 13.173009] [<ffffffff815697f1>] _raw_spin_lock+0x36/0x45
[ 13.173009] [<ffffffff81090794>] rcu_read_unlock_special+0x8c/0x1d5
[ 13.173009] [<ffffffff8109092c>] __rcu_read_unlock+0x4f/0xd7
[ 13.173009] [<ffffffff81027bd3>] rcu_read_unlock+0x21/0x23
[ 13.173009] [<ffffffff8102cc34>] cpuacct_charge+0x6c/0x75
[ 13.173009] [<ffffffff81030cc6>] update_curr+0x101/0x12e
[ 13.173009] [<ffffffff810311d0>] check_preempt_wakeup+0xf7/0x23b
[ 13.173009] [<ffffffff8102acb3>] check_preempt_curr+0x2b/0x68
[ 13.173009] [<ffffffff81031d40>] ttwu_do_wakeup+0x76/0x128
[ 13.173009] [<ffffffff81031e49>] ttwu_do_activate.constprop.63+0x57/0x5c
[ 13.173009] [<ffffffff81031e96>] scheduler_ipi+0x48/0x5d
[ 13.173009] [<ffffffff810177d5>] smp_reschedule_interrupt+0x16/0x18
[ 13.173009] [<ffffffff815710f3>] reschedule_interrupt+0x13/0x20
[ 13.173009] [<ffffffff810b66d1>] rcu_read_unlock+0x21/0x23
[ 13.173009] [<ffffffff810b739c>] find_get_page+0xa9/0xb9
[ 13.173009] [<ffffffff810b8b48>] filemap_fault+0x6a/0x34d
[ 13.173009] [<ffffffff810d1a25>] __do_fault+0x54/0x3e6
[ 13.173009] [<ffffffff810d447a>] handle_pte_fault+0x12c/0x1ed
[ 13.173009] [<ffffffff810d48f7>] handle_mm_fault+0x1cd/0x1e0
[ 13.173009] [<ffffffff8156cfee>] do_page_fault+0x42d/0x5de
[ 13.173009] [<ffffffff8156a75f>] page_fault+0x1f/0x30
[ 13.173009]
[ 13.173009] -> #1 (&rq->lock){-.-.-.}:
[ 13.173009] [<ffffffff810679b9>] check_prevs_add+0x8b/0x104
[ 13.173009] [<ffffffff81067da1>] validate_chain+0x36f/0x3ab
[ 13.173009] [<ffffffff8106846b>] __lock_acquire+0x369/0x3e2
[ 13.173009] [<ffffffff81068a0f>] lock_acquire+0xfc/0x14c
[ 13.173009] [<ffffffff815697f1>] _raw_spin_lock+0x36/0x45
[ 13.173009] [<ffffffff81027e19>] __task_rq_lock+0x8b/0xd3
[ 13.173009] [<ffffffff81032f7f>] wake_up_new_task+0x41/0x108
[ 13.173009] [<ffffffff810376c3>] do_fork+0x265/0x33f
[ 13.173009] [<ffffffff81007d02>] kernel_thread+0x6b/0x6d
[ 13.173009] [<ffffffff8153a9dd>] rest_init+0x21/0xd2
[ 13.173009] [<ffffffff81b1db4f>] start_kernel+0x3bb/0x3c6
[ 13.173009] [<ffffffff81b1d29f>] x86_64_start_reservations+0xaf/0xb3
[ 13.173009] [<ffffffff81b1d393>] x86_64_start_kernel+0xf0/0xf7
[ 13.173009]
[ 13.173009] -> #0 (&p->pi_lock){-.-.-.}:
[ 13.173009] [<ffffffff81067788>] check_prev_add+0x68/0x20e
[ 13.173009] [<ffffffff810679b9>] check_prevs_add+0x8b/0x104
[ 13.173009] [<ffffffff81067da1>] validate_chain+0x36f/0x3ab
[ 13.173009] [<ffffffff8106846b>] __lock_acquire+0x369/0x3e2
[ 13.173009] [<ffffffff81068a0f>] lock_acquire+0xfc/0x14c
[ 13.173009] [<ffffffff815698ea>] _raw_spin_lock_irqsave+0x44/0x57
[ 13.173009] [<ffffffff81032d8f>] try_to_wake_up+0x29/0x1aa
[ 13.173009] [<ffffffff81032f3c>] wake_up_process+0x10/0x12
[ 13.173009] [<ffffffff810901e9>] rcu_cpu_kthread_timer+0x44/0x58
[ 13.173009] [<ffffffff81045286>] call_timer_fn+0xac/0x1e9
[ 13.173009] [<ffffffff8104556d>] run_timer_softirq+0x1aa/0x1f2
[ 13.173009] [<ffffffff8103e487>] __do_softirq+0x109/0x26a
[ 13.173009] [<ffffffff8157144c>] call_softirq+0x1c/0x30
[ 13.173009] [<ffffffff81003207>] do_softirq+0x44/0xf1
[ 13.173009] [<ffffffff8103e8b9>] irq_exit+0x58/0xc8
[ 13.173009] [<ffffffff81017f5a>] smp_apic_timer_interrupt+0x79/0x87
[ 13.173009] [<ffffffff81570fd3>] apic_timer_interrupt+0x13/0x20
[ 13.173009] [<ffffffff810bd51a>] get_page_from_freelist+0x2aa/0x310
[ 13.173009] [<ffffffff810bdf03>] __alloc_pages_nodemask+0x178/0x243
[ 13.173009] [<ffffffff8101fe2f>] pte_alloc_one+0x1e/0x3a
[ 13.173009] [<ffffffff810d27fe>] __pte_alloc+0x22/0x14b
[ 13.173009] [<ffffffff810d48a8>] handle_mm_fault+0x17e/0x1e0
[ 13.173009] [<ffffffff8156cfee>] do_page_fault+0x42d/0x5de
[ 13.173009] [<ffffffff8156a75f>] page_fault+0x1f/0x30
[ 13.173009]
[ 13.173009] other info that might help us debug this:
[ 13.173009]
[ 13.173009] Chain exists of:
[ 13.173009] &p->pi_lock --> &rq->lock --> rcu_node_level_0
[ 13.173009]
[ 13.173009] Possible unsafe locking scenario:
[ 13.173009]
[ 13.173009] CPU0 CPU1
[ 13.173009] ---- ----
[ 13.173009] lock(rcu_node_level_0);
[ 13.173009] lock(&rq->lock);
[ 13.173009] lock(rcu_node_level_0);
[ 13.173009] lock(&p->pi_lock);
[ 13.173009]
[ 13.173009] *** DEADLOCK ***
[ 13.173009]
[ 13.173009] 3 locks held by blkid/267:
[ 13.173009] #0: (&mm->mmap_sem){++++++}, at: [<ffffffff8156cdb4>] do_page_fault+0x1f3/0x5de
[ 13.173009] #1: (&yield_timer){+.-...}, at: [<ffffffff810451da>] call_timer_fn+0x0/0x1e9
[ 13.173009] #2: (rcu_node_level_0){..-...}, at: [<ffffffff810901cc>] rcu_cpu_kthread_timer+0x27/0x58
[ 13.173009]
[ 13.173009] stack backtrace:
[ 13.173009] Pid: 267, comm: blkid Not tainted 2.6.39-rc6-mmotm0506 #1
[ 13.173009] Call Trace:
[ 13.173009] <IRQ> [<ffffffff8154a529>] print_circular_bug+0xc8/0xd9
[ 13.173009] [<ffffffff81067788>] check_prev_add+0x68/0x20e
[ 13.173009] [<ffffffff8100c861>] ? save_stack_trace+0x28/0x46
[ 13.173009] [<ffffffff810679b9>] check_prevs_add+0x8b/0x104
[ 13.173009] [<ffffffff81067da1>] validate_chain+0x36f/0x3ab
[ 13.173009] [<ffffffff8106846b>] __lock_acquire+0x369/0x3e2
[ 13.173009] [<ffffffff81032d8f>] ? try_to_wake_up+0x29/0x1aa
[ 13.173009] [<ffffffff81068a0f>] lock_acquire+0xfc/0x14c
[ 13.173009] [<ffffffff81032d8f>] ? try_to_wake_up+0x29/0x1aa
[ 13.173009] [<ffffffff810901a5>] ? rcu_check_quiescent_state+0x82/0x82
[ 13.173009] [<ffffffff815698ea>] _raw_spin_lock_irqsave+0x44/0x57
[ 13.173009] [<ffffffff81032d8f>] ? try_to_wake_up+0x29/0x1aa
[ 13.173009] [<ffffffff81032d8f>] try_to_wake_up+0x29/0x1aa
[ 13.173009] [<ffffffff810901a5>] ? rcu_check_quiescent_state+0x82/0x82
[ 13.173009] [<ffffffff81032f3c>] wake_up_process+0x10/0x12
[ 13.173009] [<ffffffff810901e9>] rcu_cpu_kthread_timer+0x44/0x58
[ 13.173009] [<ffffffff810901a5>] ? rcu_check_quiescent_state+0x82/0x82
[ 13.173009] [<ffffffff81045286>] call_timer_fn+0xac/0x1e9
[ 13.173009] [<ffffffff810451da>] ? del_timer+0x75/0x75
[ 13.173009] [<ffffffff810901a5>] ? rcu_check_quiescent_state+0x82/0x82
[ 13.173009] [<ffffffff8104556d>] run_timer_softirq+0x1aa/0x1f2
[ 13.173009] [<ffffffff8103e487>] __do_softirq+0x109/0x26a
[ 13.173009] [<ffffffff8106365f>] ? tick_dev_program_event+0x37/0xf6
[ 13.173009] [<ffffffff810a0e4a>] ? time_hardirqs_off+0x1b/0x2f
[ 13.173009] [<ffffffff8157144c>] call_softirq+0x1c/0x30
[ 13.173009] [<ffffffff81003207>] do_softirq+0x44/0xf1
[ 13.173009] [<ffffffff8103e8b9>] irq_exit+0x58/0xc8
[ 13.173009] [<ffffffff81017f5a>] smp_apic_timer_interrupt+0x79/0x87
[ 13.173009] [<ffffffff81570fd3>] apic_timer_interrupt+0x13/0x20
[ 13.173009] <EOI> [<ffffffff810bd384>] ? get_page_from_freelist+0x114/0x310
[ 13.173009] [<ffffffff810bd51a>] ? get_page_from_freelist+0x2aa/0x310
[ 13.173009] [<ffffffff812220e7>] ? clear_page_c+0x7/0x10
[ 13.173009] [<ffffffff810bd1ef>] ? prep_new_page+0x14c/0x1cd
[ 13.173009] [<ffffffff810bd51a>] get_page_from_freelist+0x2aa/0x310
[ 13.173009] [<ffffffff810bdf03>] __alloc_pages_nodemask+0x178/0x243
[ 13.173009] [<ffffffff810d46b9>] ? __pmd_alloc+0x87/0x99
[ 13.173009] [<ffffffff8101fe2f>] pte_alloc_one+0x1e/0x3a
[ 13.173009] [<ffffffff810d46b9>] ? __pmd_alloc+0x87/0x99
[ 13.173009] [<ffffffff810d27fe>] __pte_alloc+0x22/0x14b
[ 13.173009] [<ffffffff810d48a8>] handle_mm_fault+0x17e/0x1e0
[ 13.173009] [<ffffffff8156cfee>] do_page_fault+0x42d/0x5de
[ 13.173009] [<ffffffff810d915f>] ? sys_brk+0x32/0x10c
[ 13.173009] [<ffffffff810a0e4a>] ? time_hardirqs_off+0x1b/0x2f
[ 13.173009] [<ffffffff81065c4f>] ? trace_hardirqs_off_caller+0x3f/0x9c
[ 13.173009] [<ffffffff812235dd>] ? trace_hardirqs_off_thunk+0x3a/0x3c
[ 13.173009] [<ffffffff8156a75f>] page_fault+0x1f/0x30
[ 14.010075] usb 5-1: new full speed USB device number 2 using uhci_hcd
Reported-by: Valdis Kletnieks <Valdis.Kletnieks@vt.edu>
Signed-off-by: Paul E. McKenney <paul.mckenney@linaro.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2011-05-11 20:41:41 +08:00
|
|
|
|
#include <linux/atomic.h>
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
#include <linux/bitops.h>
|
2011-05-24 02:51:41 +08:00
|
|
|
|
#include <linux/export.h>
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
#include <linux/completion.h>
|
2023-10-01 01:46:56 +08:00
|
|
|
|
#include <linux/kmemleak.h>
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
#include <linux/moduleparam.h>
|
2021-07-01 09:54:59 +08:00
|
|
|
|
#include <linux/panic.h>
|
|
|
|
|
#include <linux/panic_notifier.h>
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
#include <linux/percpu.h>
|
|
|
|
|
#include <linux/notifier.h>
|
|
|
|
|
#include <linux/cpu.h>
|
|
|
|
|
#include <linux/mutex.h>
|
|
|
|
|
#include <linux/time.h>
|
2010-04-03 07:17:17 +08:00
|
|
|
|
#include <linux/kernel_stat.h>
|
2011-01-13 06:10:23 +08:00
|
|
|
|
#include <linux/wait.h>
|
|
|
|
|
#include <linux/kthread.h>
|
2017-02-02 01:07:51 +08:00
|
|
|
|
#include <uapi/linux/sched/types.h>
|
2011-05-21 03:50:29 +08:00
|
|
|
|
#include <linux/prefetch.h>
|
2012-01-24 09:05:46 +08:00
|
|
|
|
#include <linux/delay.h>
|
2012-07-07 20:57:03 +08:00
|
|
|
|
#include <linux/random.h>
|
2015-04-30 02:36:05 +08:00
|
|
|
|
#include <linux/trace_events.h>
|
2013-04-22 06:12:42 +08:00
|
|
|
|
#include <linux/suspend.h>
|
2017-04-06 00:05:18 +08:00
|
|
|
|
#include <linux/ftrace.h>
|
2018-07-26 02:49:47 +08:00
|
|
|
|
#include <linux/tick.h>
|
2018-12-13 04:32:06 +08:00
|
|
|
|
#include <linux/sysrq.h>
|
2019-02-13 00:12:15 +08:00
|
|
|
|
#include <linux/kprobes.h>
|
2019-03-21 05:13:33 +08:00
|
|
|
|
#include <linux/gfp.h>
|
|
|
|
|
#include <linux/oom.h>
|
|
|
|
|
#include <linux/smpboot.h>
|
|
|
|
|
#include <linux/jiffies.h>
|
2019-08-31 00:36:32 +08:00
|
|
|
|
#include <linux/slab.h>
|
2019-03-21 05:13:33 +08:00
|
|
|
|
#include <linux/sched/isolation.h>
|
2019-07-25 09:07:52 +08:00
|
|
|
|
#include <linux/sched/clock.h>
|
2020-05-26 05:47:53 +08:00
|
|
|
|
#include <linux/vmalloc.h>
|
|
|
|
|
#include <linux/mm.h>
|
2020-08-07 14:24:35 +08:00
|
|
|
|
#include <linux/kasan.h>
|
2022-06-08 22:40:33 +08:00
|
|
|
|
#include <linux/context_tracking.h>
|
2019-03-21 05:13:33 +08:00
|
|
|
|
#include "../time/tick-internal.h"
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
2013-10-09 11:23:47 +08:00
|
|
|
|
#include "tree.h"
|
2011-06-18 06:53:19 +08:00
|
|
|
|
#include "rcu.h"
|
2009-08-23 04:56:45 +08:00
|
|
|
|
|
2013-10-09 11:23:47 +08:00
|
|
|
|
#ifdef MODULE_PARAM_PREFIX
|
|
|
|
|
#undef MODULE_PARAM_PREFIX
|
|
|
|
|
#endif
|
|
|
|
|
#define MODULE_PARAM_PREFIX "rcutree."
|
|
|
|
|
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
/* Data structures. */
|
|
|
|
|
|
2018-08-04 12:00:38 +08:00
|
|
|
|
static DEFINE_PER_CPU_SHARED_ALIGNED(struct rcu_data, rcu_data) = {
|
2022-08-05 08:01:55 +08:00
|
|
|
|
.gpwrap = true,
|
2020-11-13 20:13:17 +08:00
|
|
|
|
#ifdef CONFIG_RCU_NOCB_CPU
|
2021-10-19 08:08:07 +08:00
|
|
|
|
.cblist.flags = SEGCBLIST_RCU_CORE,
|
2020-11-13 20:13:17 +08:00
|
|
|
|
#endif
|
2018-08-04 12:00:38 +08:00
|
|
|
|
};
|
2019-10-12 12:40:09 +08:00
|
|
|
|
static struct rcu_state rcu_state = {
|
2018-07-04 05:15:31 +08:00
|
|
|
|
.level = { &rcu_state.node[0] },
|
|
|
|
|
.gp_state = RCU_GP_IDLE,
|
|
|
|
|
.gp_seq = (0UL - 300UL) << RCU_SEQ_CTR_SHIFT,
|
|
|
|
|
.barrier_mutex = __MUTEX_INITIALIZER(rcu_state.barrier_mutex),
|
rcu: Make rcu_barrier() no longer block CPU-hotplug operations
This commit removes the cpus_read_lock() and cpus_read_unlock() calls
from rcu_barrier(), thus allowing CPUs to come and go during the course
of rcu_barrier() execution. Posting of the ->barrier_head callbacks does
synchronize with portions of RCU's CPU-hotplug notifiers, but these locks
are held for short time periods on both sides. Thus, full CPU-hotplug
operations could both start and finish during the execution of a given
rcu_barrier() invocation.
Additional synchronization is provided by a global ->barrier_lock.
Since the ->barrier_lock is only used during rcu_barrier() execution and
during onlining/offlining a CPU, the contention for this lock should
be low. It might be tempting to make use of a per-CPU lock just on
general principles, but straightforward attempts to do this have the
problems shown below.
Initial state: 3 CPUs present, CPU 0 and CPU1 do not have
any callback and CPU2 has callbacks.
1. CPU0 calls rcu_barrier().
2. CPU1 starts offlining for CPU2. CPU1 calls
rcutree_migrate_callbacks(). rcu_barrier_entrain() is called
from rcutree_migrate_callbacks(), with CPU2's rdp->barrier_lock.
It does not entrain ->barrier_head for CPU2, as rcu_barrier()
on CPU0 hasn't started the barrier sequence (by calling
rcu_seq_start(&rcu_state.barrier_sequence)) yet.
3. CPU0 starts new barrier sequence. It iterates over
CPU0 and CPU1, after acquiring their per-cpu ->barrier_lock
and finds 0 segcblist length. It updates ->barrier_seq_snap
for CPU0 and CPU1 and continues loop iteration to CPU2.
for_each_possible_cpu(cpu) {
raw_spin_lock_irqsave(&rdp->barrier_lock, flags);
if (!rcu_segcblist_n_cbs(&rdp->cblist)) {
WRITE_ONCE(rdp->barrier_seq_snap, gseq);
raw_spin_unlock_irqrestore(&rdp->barrier_lock, flags);
rcu_barrier_trace(TPS("NQ"), cpu, rcu_state.barrier_sequence);
continue;
}
4. rcutree_migrate_callbacks() completes execution on CPU1.
Segcblist len for CPU2 becomes 0.
5. The loop iteration on CPU0, checks rcu_segcblist_n_cbs(&rdp->cblist)
for CPU2 and completes the loop iteration after setting
->barrier_seq_snap.
6. As there isn't any ->barrier_head callback entrained; at
this point, rcu_barrier() in CPU0 returns.
7. The callbacks, which migrated from CPU2 to CPU1, execute.
Straightforward per-CPU locking is also subject to the following race
condition noted by Boqun Feng:
1. CPU0 calls rcu_barrier(), starting a new barrier sequence by invoking
rcu_seq_start() and init_completion(), but does not yet initialize
rcu_state.barrier_cpu_count.
2. CPU1 starts offlining for CPU2, calling rcutree_migrate_callbacks(),
which in turn calls rcu_barrier_entrain() holding CPU2's.
rdp->barrier_lock. It then entrains ->barrier_head for CPU2
and atomically increments rcu_state.barrier_cpu_count, which is
unfortunately not yet initialized to the value 2.
3. The just-entrained RCU callback is invoked. It atomically
decrements rcu_state.barrier_cpu_count and sees that it is
now zero. This callback therefore invokes complete().
4. CPU0 continues executing rcu_barrier(), but is not blocked
by its call to wait_for_completion(). This results in rcu_barrier()
returning before all pre-existing callbacks have been invoked,
which is a bug.
Therefore, synchronization is provided by rcu_state.barrier_lock,
which is also held across the initialization sequence, especially the
rcu_seq_start() and the atomic_set() that sets rcu_state.barrier_cpu_count
to the value 2. In addition, this lock is held when entraining the
rcu_barrier() callback, when deciding whether or not a CPU has callbacks
that rcu_barrier() must wait on, when setting the ->qsmaskinitnext for
incoming CPUs, and when migrating callbacks from a CPU that is going
offline.
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Co-developed-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-12-15 05:35:17 +08:00
|
|
|
|
.barrier_lock = __RAW_SPIN_LOCK_UNLOCKED(rcu_state.barrier_lock),
|
2018-07-04 05:15:31 +08:00
|
|
|
|
.name = RCU_NAME,
|
|
|
|
|
.abbr = RCU_ABBR,
|
|
|
|
|
.exp_mutex = __MUTEX_INITIALIZER(rcu_state.exp_mutex),
|
|
|
|
|
.exp_wake_mutex = __MUTEX_INITIALIZER(rcu_state.exp_wake_mutex),
|
2021-02-16 23:04:34 +08:00
|
|
|
|
.ofl_lock = __ARCH_SPIN_LOCK_UNLOCKED,
|
2018-07-04 05:15:31 +08:00
|
|
|
|
};
|
2011-02-08 04:47:15 +08:00
|
|
|
|
|
2015-04-21 02:40:50 +08:00
|
|
|
|
/* Dump rcu_node combining tree at boot to verify correct setup. */
|
|
|
|
|
static bool dump_tree;
|
|
|
|
|
module_param(dump_tree, bool, 0444);
|
2019-03-21 05:13:33 +08:00
|
|
|
|
/* By default, use RCU_SOFTIRQ instead of rcuc kthreads. */
|
2020-12-15 22:16:46 +08:00
|
|
|
|
static bool use_softirq = !IS_ENABLED(CONFIG_PREEMPT_RT);
|
|
|
|
|
#ifndef CONFIG_PREEMPT_RT
|
2019-03-21 05:13:33 +08:00
|
|
|
|
module_param(use_softirq, bool, 0444);
|
2020-12-15 22:16:46 +08:00
|
|
|
|
#endif
|
2015-04-21 01:27:15 +08:00
|
|
|
|
/* Control rcu_node-tree auto-balancing at boot time. */
|
|
|
|
|
static bool rcu_fanout_exact;
|
|
|
|
|
module_param(rcu_fanout_exact, bool, 0444);
|
2015-04-22 00:12:13 +08:00
|
|
|
|
/* Increase (but not decrease) the RCU_FANOUT_LEAF at boot time. */
|
|
|
|
|
static int rcu_fanout_leaf = RCU_FANOUT_LEAF;
|
2012-07-02 06:42:33 +08:00
|
|
|
|
module_param(rcu_fanout_leaf, int, 0444);
|
2012-04-24 06:52:53 +08:00
|
|
|
|
int rcu_num_lvls __read_mostly = RCU_NUM_LVLS;
|
2015-06-03 14:18:30 +08:00
|
|
|
|
/* Number of rcu_nodes at specified level. */
|
2017-03-16 04:11:11 +08:00
|
|
|
|
int num_rcu_lvl[] = NUM_RCU_LVL_INIT;
|
2012-04-24 06:52:53 +08:00
|
|
|
|
int rcu_num_nodes __read_mostly = NUM_RCU_NODES; /* Total # rcu_nodes in use. */
|
|
|
|
|
|
2011-07-11 06:57:35 +08:00
|
|
|
|
/*
|
rcu: Narrow early boot window of illegal synchronous grace periods
The current preemptible RCU implementation goes through three phases
during bootup. In the first phase, there is only one CPU that is running
with preemption disabled, so that a no-op is a synchronous grace period.
In the second mid-boot phase, the scheduler is running, but RCU has
not yet gotten its kthreads spawned (and, for expedited grace periods,
workqueues are not yet running. During this time, any attempt to do
a synchronous grace period will hang the system (or complain bitterly,
depending). In the third and final phase, RCU is fully operational and
everything works normally.
This has been OK for some time, but there has recently been some
synchronous grace periods showing up during the second mid-boot phase.
This code worked "by accident" for awhile, but started failing as soon
as expedited RCU grace periods switched over to workqueues in commit
8b355e3bc140 ("rcu: Drive expedited grace periods from workqueue").
Note that the code was buggy even before this commit, as it was subject
to failure on real-time systems that forced all expedited grace periods
to run as normal grace periods (for example, using the rcu_normal ksysfs
parameter). The callchain from the failure case is as follows:
early_amd_iommu_init()
|-> acpi_put_table(ivrs_base);
|-> acpi_tb_put_table(table_desc);
|-> acpi_tb_invalidate_table(table_desc);
|-> acpi_tb_release_table(...)
|-> acpi_os_unmap_memory
|-> acpi_os_unmap_iomem
|-> acpi_os_map_cleanup
|-> synchronize_rcu_expedited
The kernel showing this callchain was built with CONFIG_PREEMPT_RCU=y,
which caused the code to try using workqueues before they were
initialized, which did not go well.
This commit therefore reworks RCU to permit synchronous grace periods
to proceed during this mid-boot phase. This commit is therefore a
fix to a regression introduced in v4.9, and is therefore being put
forward post-merge-window in v4.10.
This commit sets a flag from the existing rcu_scheduler_starting()
function which causes all synchronous grace periods to take the expedited
path. The expedited path now checks this flag, using the requesting task
to drive the expedited grace period forward during the mid-boot phase.
Finally, this flag is updated by a core_initcall() function named
rcu_exp_runtime_mode(), which causes the runtime codepaths to be used.
Note that this arrangement assumes that tasks are not sent POSIX signals
(or anything similar) from the time that the first task is spawned
through core_initcall() time.
Fixes: 8b355e3bc140 ("rcu: Drive expedited grace periods from workqueue")
Reported-by: "Zheng, Lv" <lv.zheng@intel.com>
Reported-by: Borislav Petkov <bp@alien8.de>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Stan Kain <stan.kain@gmail.com>
Tested-by: Ivan <waffolz@hotmail.com>
Tested-by: Emanuel Castelo <emanuel.castelo@gmail.com>
Tested-by: Bruno Pesavento <bpesavento@infinito.it>
Tested-by: Borislav Petkov <bp@suse.de>
Tested-by: Frederic Bezies <fredbezies@gmail.com>
Cc: <stable@vger.kernel.org> # 4.9.0-
2017-01-10 18:28:26 +08:00
|
|
|
|
* The rcu_scheduler_active variable is initialized to the value
|
|
|
|
|
* RCU_SCHEDULER_INACTIVE and transitions RCU_SCHEDULER_INIT just before the
|
|
|
|
|
* first task is spawned. So when this variable is RCU_SCHEDULER_INACTIVE,
|
|
|
|
|
* RCU can assume that there is but one task, allowing RCU to (for example)
|
2016-04-08 20:00:03 +08:00
|
|
|
|
* optimize synchronize_rcu() to a simple barrier(). When this variable
|
rcu: Narrow early boot window of illegal synchronous grace periods
The current preemptible RCU implementation goes through three phases
during bootup. In the first phase, there is only one CPU that is running
with preemption disabled, so that a no-op is a synchronous grace period.
In the second mid-boot phase, the scheduler is running, but RCU has
not yet gotten its kthreads spawned (and, for expedited grace periods,
workqueues are not yet running. During this time, any attempt to do
a synchronous grace period will hang the system (or complain bitterly,
depending). In the third and final phase, RCU is fully operational and
everything works normally.
This has been OK for some time, but there has recently been some
synchronous grace periods showing up during the second mid-boot phase.
This code worked "by accident" for awhile, but started failing as soon
as expedited RCU grace periods switched over to workqueues in commit
8b355e3bc140 ("rcu: Drive expedited grace periods from workqueue").
Note that the code was buggy even before this commit, as it was subject
to failure on real-time systems that forced all expedited grace periods
to run as normal grace periods (for example, using the rcu_normal ksysfs
parameter). The callchain from the failure case is as follows:
early_amd_iommu_init()
|-> acpi_put_table(ivrs_base);
|-> acpi_tb_put_table(table_desc);
|-> acpi_tb_invalidate_table(table_desc);
|-> acpi_tb_release_table(...)
|-> acpi_os_unmap_memory
|-> acpi_os_unmap_iomem
|-> acpi_os_map_cleanup
|-> synchronize_rcu_expedited
The kernel showing this callchain was built with CONFIG_PREEMPT_RCU=y,
which caused the code to try using workqueues before they were
initialized, which did not go well.
This commit therefore reworks RCU to permit synchronous grace periods
to proceed during this mid-boot phase. This commit is therefore a
fix to a regression introduced in v4.9, and is therefore being put
forward post-merge-window in v4.10.
This commit sets a flag from the existing rcu_scheduler_starting()
function which causes all synchronous grace periods to take the expedited
path. The expedited path now checks this flag, using the requesting task
to drive the expedited grace period forward during the mid-boot phase.
Finally, this flag is updated by a core_initcall() function named
rcu_exp_runtime_mode(), which causes the runtime codepaths to be used.
Note that this arrangement assumes that tasks are not sent POSIX signals
(or anything similar) from the time that the first task is spawned
through core_initcall() time.
Fixes: 8b355e3bc140 ("rcu: Drive expedited grace periods from workqueue")
Reported-by: "Zheng, Lv" <lv.zheng@intel.com>
Reported-by: Borislav Petkov <bp@alien8.de>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Stan Kain <stan.kain@gmail.com>
Tested-by: Ivan <waffolz@hotmail.com>
Tested-by: Emanuel Castelo <emanuel.castelo@gmail.com>
Tested-by: Bruno Pesavento <bpesavento@infinito.it>
Tested-by: Borislav Petkov <bp@suse.de>
Tested-by: Frederic Bezies <fredbezies@gmail.com>
Cc: <stable@vger.kernel.org> # 4.9.0-
2017-01-10 18:28:26 +08:00
|
|
|
|
* is RCU_SCHEDULER_INIT, RCU must actually do all the hard work required
|
|
|
|
|
* to detect real grace periods. This variable is also used to suppress
|
|
|
|
|
* boot-time false positives from lockdep-RCU error checking. Finally, it
|
|
|
|
|
* transitions from RCU_SCHEDULER_INIT to RCU_SCHEDULER_RUNNING after RCU
|
|
|
|
|
* is fully initialized, including all of its kthreads having been spawned.
|
2011-07-11 06:57:35 +08:00
|
|
|
|
*/
|
2010-04-03 07:17:17 +08:00
|
|
|
|
int rcu_scheduler_active __read_mostly;
|
|
|
|
|
EXPORT_SYMBOL_GPL(rcu_scheduler_active);
|
|
|
|
|
|
2011-07-11 06:57:35 +08:00
|
|
|
|
/*
|
|
|
|
|
* The rcu_scheduler_fully_active variable transitions from zero to one
|
|
|
|
|
* during the early_initcall() processing, which is after the scheduler
|
|
|
|
|
* is capable of creating new tasks. So RCU processing (for example,
|
|
|
|
|
* creating tasks for RCU priority boosting) must be delayed until after
|
|
|
|
|
* rcu_scheduler_fully_active transitions from zero to one. We also
|
|
|
|
|
* currently delay invocation of any RCU callbacks until after this point.
|
|
|
|
|
*
|
|
|
|
|
* It might later prove better for people registering RCU callbacks during
|
|
|
|
|
* early boot to take responsibility for these callbacks, but one step at
|
|
|
|
|
* a time.
|
|
|
|
|
*/
|
|
|
|
|
static int rcu_scheduler_fully_active __read_mostly;
|
|
|
|
|
|
2018-07-04 08:22:34 +08:00
|
|
|
|
static void rcu_report_qs_rnp(unsigned long mask, struct rcu_node *rnp,
|
|
|
|
|
unsigned long gps, unsigned long flags);
|
2012-07-16 18:42:35 +08:00
|
|
|
|
static void rcu_boost_kthread_setaffinity(struct rcu_node *rnp, int outgoingcpu);
|
2011-06-16 06:47:09 +08:00
|
|
|
|
static void invoke_rcu_core(void);
|
2018-07-04 08:22:34 +08:00
|
|
|
|
static void rcu_report_exp_rdp(struct rcu_data *rdp);
|
2016-04-16 07:35:29 +08:00
|
|
|
|
static void sync_sched_exp_online_cleanup(int cpu);
|
2019-10-31 02:56:10 +08:00
|
|
|
|
static void check_cb_ovld_locked(struct rcu_data *rdp, struct rcu_node *rnp);
|
2020-11-12 08:51:21 +08:00
|
|
|
|
static bool rcu_rdp_is_offloaded(struct rcu_data *rdp);
|
2022-11-07 08:33:38 +08:00
|
|
|
|
static bool rcu_rdp_cpu_online(struct rcu_data *rdp);
|
|
|
|
|
static bool rcu_init_invoked(void);
|
|
|
|
|
static void rcu_cleanup_dead_rnp(struct rcu_node *rnp_leaf);
|
|
|
|
|
static void rcu_init_new_rnp(struct rcu_node *rnp_leaf);
|
2011-01-13 06:10:23 +08:00
|
|
|
|
|
rcu/nocb: Add option to opt rcuo kthreads out of RT priority
This commit introduces a RCU_NOCB_CPU_CB_BOOST Kconfig option that
prevents rcuo kthreads from running at real-time priority, even in
kernels built with RCU_BOOST. This capability is important to devices
needing low-latency (as in a few milliseconds) response from expedited
RCU grace periods, but which are not running a classic real-time workload.
On such devices, permitting the rcuo kthreads to run at real-time priority
results in unacceptable latencies imposed on the application tasks,
which run as SCHED_OTHER.
See for example the following trace output:
<snip>
<...>-60 [006] d..1 2979.028717: rcu_batch_start: rcu_preempt CBs=34619 bl=270
<snip>
If that rcuop kthread were permitted to run at real-time SCHED_FIFO
priority, it would monopolize its CPU for hundreds of milliseconds
while invoking those 34619 RCU callback functions, which would cause an
unacceptably long latency spike for many application stacks on Android
platforms.
However, some existing real-time workloads require that callback
invocation run at SCHED_FIFO priority, for example, those running on
systems with heavy SCHED_OTHER background loads. (It is the real-time
system's administrator's responsibility to make sure that important
real-time tasks run at a higher priority than do RCU's kthreads.)
Therefore, this new RCU_NOCB_CPU_CB_BOOST Kconfig option defaults to
"y" on kernels built with PREEMPT_RT and defaults to "n" otherwise.
The effect is to preserve current behavior for real-time systems, but for
other systems to allow expedited RCU grace periods to run with real-time
priority while continuing to invoke RCU callbacks as SCHED_OTHER.
As you would expect, this RCU_NOCB_CPU_CB_BOOST Kconfig option has no
effect except on CPUs with offloaded RCU callbacks.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Acked-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Reviewed-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
2022-05-11 16:57:03 +08:00
|
|
|
|
/*
|
|
|
|
|
* rcuc/rcub/rcuop kthread realtime priority. The "rcuop"
|
|
|
|
|
* real-time priority(enabling/disabling) is controlled by
|
|
|
|
|
* the extra CONFIG_RCU_NOCB_CPU_CB_BOOST configuration.
|
|
|
|
|
*/
|
2015-04-22 00:22:14 +08:00
|
|
|
|
static int kthread_prio = IS_ENABLED(CONFIG_RCU_BOOST) ? 1 : 0;
|
2019-02-21 22:13:27 +08:00
|
|
|
|
module_param(kthread_prio, int, 0444);
|
2014-12-12 23:37:48 +08:00
|
|
|
|
|
2015-04-15 10:33:59 +08:00
|
|
|
|
/* Delay in jiffies for grace-period initialization delays, debug only. */
|
2015-03-11 09:33:20 +08:00
|
|
|
|
|
rcu: Remove *_SLOW_* Kconfig options
The RCU_TORTURE_TEST_SLOW_PREINIT, RCU_TORTURE_TEST_SLOW_PREINIT_DELAY,
RCU_TORTURE_TEST_SLOW_PREINIT_DELAY, RCU_TORTURE_TEST_SLOW_INIT,
RCU_TORTURE_TEST_SLOW_INIT_DELAY, RCU_TORTURE_TEST_SLOW_CLEANUP,
and RCU_TORTURE_TEST_SLOW_CLEANUP_DELAY Kconfig options are only
useful for torture testing, and there are the rcutree.gp_cleanup_delay,
rcutree.gp_init_delay, and rcutree.gp_preinit_delay kernel boot parameters
that rcutorture can use instead. The effect of these parameters is to
artificially slow down grace period initialization and cleanup in order
to make some types of race conditions happen more often.
This commit therefore simplifies Tree RCU a bit by removing the Kconfig
options and adding the corresponding kernel parameters to rcutorture's
.boot files instead. However, this commit also leaves out the kernel
parameters for TREE02, TREE04, and TREE07 in order to have about the
same number of tests slowed as not slowed. TREE01, TREE03, TREE05,
and TREE06 are slowed, and the rest are not slowed.
Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-05-11 05:36:55 +08:00
|
|
|
|
static int gp_preinit_delay;
|
|
|
|
|
module_param(gp_preinit_delay, int, 0444);
|
|
|
|
|
static int gp_init_delay;
|
|
|
|
|
module_param(gp_init_delay, int, 0444);
|
|
|
|
|
static int gp_cleanup_delay;
|
|
|
|
|
module_param(gp_cleanup_delay, int, 0444);
|
2015-03-11 09:33:20 +08:00
|
|
|
|
|
2020-08-11 00:58:03 +08:00
|
|
|
|
// Add delay to rcu_read_unlock() for strict grace periods.
|
|
|
|
|
static int rcu_unlock_delay;
|
|
|
|
|
#ifdef CONFIG_RCU_STRICT_GRACE_PERIOD
|
|
|
|
|
module_param(rcu_unlock_delay, int, 0444);
|
|
|
|
|
#endif
|
|
|
|
|
|
2020-05-26 05:47:52 +08:00
|
|
|
|
/*
|
|
|
|
|
* This rcu parameter is runtime-read-only. It reflects
|
|
|
|
|
* a minimum allowed number of objects which can be cached
|
|
|
|
|
* per-CPU. Object size is equal to one page. This value
|
|
|
|
|
* can be changed at boot time.
|
|
|
|
|
*/
|
2020-10-30 00:50:04 +08:00
|
|
|
|
static int rcu_min_cached_objs = 5;
|
2020-05-26 05:47:52 +08:00
|
|
|
|
module_param(rcu_min_cached_objs, int, 0444);
|
|
|
|
|
|
2021-04-16 01:19:56 +08:00
|
|
|
|
// A page shrinker can ask for pages to be freed to make them
|
|
|
|
|
// available for other parts of the system. This usually happens
|
|
|
|
|
// under low memory conditions, and in that case we should also
|
|
|
|
|
// defer page-cache filling for a short time period.
|
|
|
|
|
//
|
|
|
|
|
// The default value is 5 seconds, which is long enough to reduce
|
|
|
|
|
// interference with the shrinker while it asks other systems to
|
|
|
|
|
// drain their caches.
|
|
|
|
|
static int rcu_delay_page_cache_fill_msec = 5000;
|
|
|
|
|
module_param(rcu_delay_page_cache_fill_msec, int, 0444);
|
|
|
|
|
|
2018-07-03 03:15:25 +08:00
|
|
|
|
/* Retrieve RCU kthreads priority for rcutorture */
|
2018-06-20 06:14:18 +08:00
|
|
|
|
int rcu_get_gp_kthreads_prio(void)
|
|
|
|
|
{
|
|
|
|
|
return kthread_prio;
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(rcu_get_gp_kthreads_prio);
|
|
|
|
|
|
2015-04-16 03:08:22 +08:00
|
|
|
|
/*
|
|
|
|
|
* Number of grace periods between delays, normalized by the duration of
|
2017-02-09 06:49:27 +08:00
|
|
|
|
* the delay. The longer the delay, the more the grace periods between
|
2015-04-16 03:08:22 +08:00
|
|
|
|
* each delay. The reason for this normalization is that it means that,
|
|
|
|
|
* for non-zero delays, the overall slowdown of grace periods is constant
|
|
|
|
|
* regardless of the duration of the delay. This arrangement balances
|
|
|
|
|
* the need for long delays to increase some race probabilities with the
|
|
|
|
|
* need for fast grace periods to increase other race probabilities.
|
|
|
|
|
*/
|
2021-03-31 04:47:42 +08:00
|
|
|
|
#define PER_RCU_NODE_PERIOD 3 /* Number of grace periods between delays for debugging. */
|
2015-01-23 10:24:08 +08:00
|
|
|
|
|
2009-09-24 00:50:41 +08:00
|
|
|
|
/*
|
2015-03-04 06:57:58 +08:00
|
|
|
|
* Return true if an RCU grace period is in progress. The READ_ONCE()s
|
2009-09-24 00:50:41 +08:00
|
|
|
|
* permit this function to be invoked without holding the root rcu_node
|
|
|
|
|
* structure's ->lock, but of course results can be subject to change.
|
|
|
|
|
*/
|
2018-07-04 08:22:34 +08:00
|
|
|
|
static int rcu_gp_in_progress(void)
|
2009-09-24 00:50:41 +08:00
|
|
|
|
{
|
2018-07-04 08:22:34 +08:00
|
|
|
|
return rcu_seq_state(rcu_seq_current(&rcu_state.gp_seq));
|
2009-03-13 10:20:49 +08:00
|
|
|
|
}
|
|
|
|
|
|
2018-10-03 07:05:46 +08:00
|
|
|
|
/*
|
|
|
|
|
* Return the number of callbacks queued on the specified CPU.
|
|
|
|
|
* Handles both the nocbs and normal cases.
|
|
|
|
|
*/
|
|
|
|
|
static long rcu_get_n_cbs_cpu(int cpu)
|
|
|
|
|
{
|
|
|
|
|
struct rcu_data *rdp = per_cpu_ptr(&rcu_data, cpu);
|
|
|
|
|
|
2019-05-21 23:28:41 +08:00
|
|
|
|
if (rcu_segcblist_is_enabled(&rdp->cblist))
|
2018-10-03 07:05:46 +08:00
|
|
|
|
return rcu_segcblist_n_cbs(&rdp->cblist);
|
2019-05-21 23:28:41 +08:00
|
|
|
|
return 0;
|
2018-10-03 07:05:46 +08:00
|
|
|
|
}
|
|
|
|
|
|
2018-06-29 05:45:25 +08:00
|
|
|
|
void rcu_softirq_qs(void)
|
2009-03-13 10:20:49 +08:00
|
|
|
|
{
|
2018-07-03 05:30:37 +08:00
|
|
|
|
rcu_qs();
|
2018-06-29 05:45:25 +08:00
|
|
|
|
rcu_preempt_deferred_qs(current);
|
2021-03-25 08:08:48 +08:00
|
|
|
|
rcu_tasks_qs(current, false);
|
2009-03-13 10:20:49 +08:00
|
|
|
|
}
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
2016-11-03 05:23:30 +08:00
|
|
|
|
/*
|
|
|
|
|
* Reset the current CPU's ->dynticks counter to indicate that the
|
|
|
|
|
* newly onlined CPU is no longer in an extended quiescent state.
|
|
|
|
|
* This will either leave the counter unchanged, or increment it
|
|
|
|
|
* to the next non-quiescent value.
|
|
|
|
|
*
|
|
|
|
|
* The non-atomic test/increment sequence works because the upper bits
|
|
|
|
|
* of the ->dynticks counter are manipulated only by the corresponding CPU,
|
|
|
|
|
* or when the corresponding CPU is offline.
|
|
|
|
|
*/
|
|
|
|
|
static void rcu_dynticks_eqs_online(void)
|
|
|
|
|
{
|
2022-06-08 22:40:35 +08:00
|
|
|
|
if (ct_dynticks() & RCU_DYNTICKS_IDX)
|
2016-11-03 05:23:30 +08:00
|
|
|
|
return;
|
2022-06-08 22:40:35 +08:00
|
|
|
|
ct_state_inc(RCU_DYNTICKS_IDX);
|
2016-11-03 08:25:06 +08:00
|
|
|
|
}
|
|
|
|
|
|
2016-11-03 05:12:05 +08:00
|
|
|
|
/*
|
|
|
|
|
* Snapshot the ->dynticks counter with full ordering so as to allow
|
|
|
|
|
* stable comparison of this counter with past and future snapshots.
|
|
|
|
|
*/
|
2022-06-08 22:40:29 +08:00
|
|
|
|
static int rcu_dynticks_snap(int cpu)
|
2016-11-03 05:12:05 +08:00
|
|
|
|
{
|
rcu: Weaken ->dynticks accesses and updates
Accesses to the rcu_data structure's ->dynticks field have always been
fully ordered because it was not possible to prove that weaker ordering
was safe. However, with the removal of the rcu_eqs_special_set() function
and the advent of the Linux-kernel memory model, it is now easy to show
that two of the four original full memory barriers can be weakened to
acquire and release operations. The remaining pair must remain full
memory barriers. This change makes the memory ordering requirements
more evident, and it might well also speed up the to-idle and from-idle
fastpaths on some architectures.
The following litmus test, adapted from one supplied off-list by Frederic
Weisbecker, models the RCU grace-period kthread detecting an idle CPU
that is concurrently transitioning to non-idle:
C dynticks-from-idle
{
DYNTICKS=0; (* Initially idle. *)
}
P0(int *X, int *DYNTICKS)
{
int dynticks;
int x;
// Idle.
dynticks = READ_ONCE(*DYNTICKS);
smp_store_release(DYNTICKS, dynticks + 1);
smp_mb();
// Now non-idle
x = READ_ONCE(*X);
}
P1(int *X, int *DYNTICKS)
{
int dynticks;
WRITE_ONCE(*X, 1);
smp_mb();
dynticks = smp_load_acquire(DYNTICKS);
}
exists (1:dynticks=0 /\ 0:x=1)
Running "herd7 -conf linux-kernel.cfg dynticks-from-idle.litmus" verifies
this transition, namely, showing that if the RCU grace-period kthread (P1)
sees another CPU as idle (P0), then any memory access prior to the start
of the grace period (P1's write to X) will be seen by any RCU read-side
critical section following the to-non-idle transition (P0's read from X).
This is a straightforward use of full memory barriers to force ordering
in a store-buffering (SB) litmus test.
The following litmus test, also adapted from the one supplied off-list
by Frederic Weisbecker, models the RCU grace-period kthread detecting
a non-idle CPU that is concurrently transitioning to idle:
C dynticks-into-idle
{
DYNTICKS=1; (* Initially non-idle. *)
}
P0(int *X, int *DYNTICKS)
{
int dynticks;
// Non-idle.
WRITE_ONCE(*X, 1);
dynticks = READ_ONCE(*DYNTICKS);
smp_store_release(DYNTICKS, dynticks + 1);
smp_mb();
// Now idle.
}
P1(int *X, int *DYNTICKS)
{
int x;
int dynticks;
smp_mb();
dynticks = smp_load_acquire(DYNTICKS);
x = READ_ONCE(*X);
}
exists (1:dynticks=2 /\ 1:x=0)
Running "herd7 -conf linux-kernel.cfg dynticks-into-idle.litmus" verifies
this transition, namely, showing that if the RCU grace-period kthread
(P1) sees another CPU as newly idle (P0), then any pre-idle memory access
(P0's write to X) will be seen by any code following the grace period
(P1's read from X). This is a simple release-acquire pair forcing
ordering in a message-passing (MP) litmus test.
Of course, if the grace-period kthread detects the CPU as non-idle,
it will refrain from reporting a quiescent state on behalf of that CPU,
so there are no ordering requirements from the grace-period kthread in
that case. However, other subsystems call rcu_is_idle_cpu() to check
for CPUs being non-idle from an RCU perspective. That case is also
verified by the above litmus tests with the proviso that the sense of
the low-order bit of the DYNTICKS counter be inverted.
Unfortunately, on x86 smp_mb() is as expensive as a cache-local atomic
increment. This commit therefore weakens only the read from ->dynticks.
However, the updates are abstracted into a rcu_dynticks_inc() function
to ease any future changes that might be needed.
[ paulmck: Apply Linus Torvalds feedback. ]
Link: https://lore.kernel.org/lkml/20210721202127.2129660-4-paulmck@kernel.org/
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-05-20 08:25:42 +08:00
|
|
|
|
smp_mb(); // Fundamental RCU ordering guarantee.
|
2022-06-08 22:40:29 +08:00
|
|
|
|
return ct_dynticks_cpu_acquire(cpu);
|
2016-11-03 05:12:05 +08:00
|
|
|
|
}
|
|
|
|
|
|
2016-11-03 08:25:06 +08:00
|
|
|
|
/*
|
|
|
|
|
* Return true if the snapshot returned from rcu_dynticks_snap()
|
|
|
|
|
* indicates that RCU is in an extended quiescent state.
|
|
|
|
|
*/
|
|
|
|
|
static bool rcu_dynticks_in_eqs(int snap)
|
|
|
|
|
{
|
2022-06-08 22:40:35 +08:00
|
|
|
|
return !(snap & RCU_DYNTICKS_IDX);
|
2016-11-03 08:25:06 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
2018-08-04 12:00:38 +08:00
|
|
|
|
* Return true if the CPU corresponding to the specified rcu_data
|
2016-11-03 08:25:06 +08:00
|
|
|
|
* structure has spent some time in an extended quiescent state since
|
|
|
|
|
* rcu_dynticks_snap() returned the specified snapshot.
|
|
|
|
|
*/
|
2018-08-04 12:00:38 +08:00
|
|
|
|
static bool rcu_dynticks_in_eqs_since(struct rcu_data *rdp, int snap)
|
2016-11-03 08:25:06 +08:00
|
|
|
|
{
|
2022-06-08 22:40:29 +08:00
|
|
|
|
return snap != rcu_dynticks_snap(rdp->cpu);
|
2016-11-03 08:25:06 +08:00
|
|
|
|
}
|
|
|
|
|
|
rcu-tasks: Avoid IPIing userspace/idle tasks if kernel is so built
Systems running CPU-bound real-time task do not want IPIs sent to CPUs
executing nohz_full userspace tasks. Battery-powered systems don't
want IPIs sent to idle CPUs in low-power mode. Unfortunately, RCU tasks
trace can and will send such IPIs in some cases.
Both of these situations occur only when the target CPU is in RCU
dyntick-idle mode, in other words, when RCU is not watching the
target CPU. This suggests that CPUs in dyntick-idle mode should use
memory barriers in outermost invocations of rcu_read_lock_trace()
and rcu_read_unlock_trace(), which would allow the RCU tasks trace
grace period to directly read out the target CPU's read-side state.
One challenge is that RCU tasks trace is not targeting a specific
CPU, but rather a task. And that task could switch from one CPU to
another at any time.
This commit therefore uses try_invoke_on_locked_down_task()
and checks for task_curr() in trc_inspect_reader_notrunning().
When this condition holds, the target task is running and cannot move.
If CONFIG_TASKS_TRACE_RCU_READ_MB=y, the new rcu_dynticks_zero_in_eqs()
function can be used to check if the specified integer (in this case,
t->trc_reader_nesting) is zero while the target CPU remains in that same
dyntick-idle sojourn. If so, the target task is in a quiescent state.
If not, trc_read_check_handler() must indicate failure so that the
grace-period kthread can take appropriate action or retry after an
appropriate delay, as the case may be.
With this change, given CONFIG_TASKS_TRACE_RCU_READ_MB=y, if a given
CPU remains idle or a given task continues executing in nohz_full mode,
the RCU tasks trace grace-period kthread will detect this without the
need to send an IPI.
Suggested-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-03-20 06:33:12 +08:00
|
|
|
|
/*
|
|
|
|
|
* Return true if the referenced integer is zero while the specified
|
|
|
|
|
* CPU remains within a single extended quiescent state.
|
|
|
|
|
*/
|
|
|
|
|
bool rcu_dynticks_zero_in_eqs(int cpu, int *vp)
|
|
|
|
|
{
|
|
|
|
|
int snap;
|
|
|
|
|
|
|
|
|
|
// If not quiescent, force back to earlier extended quiescent state.
|
2022-06-08 22:40:35 +08:00
|
|
|
|
snap = ct_dynticks_cpu(cpu) & ~RCU_DYNTICKS_IDX;
|
rcu-tasks: Avoid IPIing userspace/idle tasks if kernel is so built
Systems running CPU-bound real-time task do not want IPIs sent to CPUs
executing nohz_full userspace tasks. Battery-powered systems don't
want IPIs sent to idle CPUs in low-power mode. Unfortunately, RCU tasks
trace can and will send such IPIs in some cases.
Both of these situations occur only when the target CPU is in RCU
dyntick-idle mode, in other words, when RCU is not watching the
target CPU. This suggests that CPUs in dyntick-idle mode should use
memory barriers in outermost invocations of rcu_read_lock_trace()
and rcu_read_unlock_trace(), which would allow the RCU tasks trace
grace period to directly read out the target CPU's read-side state.
One challenge is that RCU tasks trace is not targeting a specific
CPU, but rather a task. And that task could switch from one CPU to
another at any time.
This commit therefore uses try_invoke_on_locked_down_task()
and checks for task_curr() in trc_inspect_reader_notrunning().
When this condition holds, the target task is running and cannot move.
If CONFIG_TASKS_TRACE_RCU_READ_MB=y, the new rcu_dynticks_zero_in_eqs()
function can be used to check if the specified integer (in this case,
t->trc_reader_nesting) is zero while the target CPU remains in that same
dyntick-idle sojourn. If so, the target task is in a quiescent state.
If not, trc_read_check_handler() must indicate failure so that the
grace-period kthread can take appropriate action or retry after an
appropriate delay, as the case may be.
With this change, given CONFIG_TASKS_TRACE_RCU_READ_MB=y, if a given
CPU remains idle or a given task continues executing in nohz_full mode,
the RCU tasks trace grace-period kthread will detect this without the
need to send an IPI.
Suggested-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-03-20 06:33:12 +08:00
|
|
|
|
smp_rmb(); // Order ->dynticks and *vp reads.
|
|
|
|
|
if (READ_ONCE(*vp))
|
|
|
|
|
return false; // Non-zero, so report failure;
|
|
|
|
|
smp_rmb(); // Order *vp read and ->dynticks re-read.
|
|
|
|
|
|
|
|
|
|
// If still in the same extended quiescent state, we are good!
|
2022-06-08 22:40:29 +08:00
|
|
|
|
return snap == ct_dynticks_cpu(cpu);
|
2016-11-03 04:33:57 +08:00
|
|
|
|
}
|
rcu: Make cond_resched_rcu_qs() apply to normal RCU flavors
Although cond_resched_rcu_qs() only applies to TASKS_RCU, it is used
in places where it would be useful for it to apply to the normal RCU
flavors, rcu_preempt, rcu_sched, and rcu_bh. This is especially the
case for workloads that aggressively overload the system, particularly
those that generate large numbers of RCU updates on systems running
NO_HZ_FULL CPUs. This commit therefore communicates quiescent states
from cond_resched_rcu_qs() to the normal RCU flavors.
Note that it is unfortunately necessary to leave the old ->passed_quiesce
mechanism in place to allow quiescent states that apply to only one
flavor to be recorded. (Yes, we could decrement ->rcu_qs_ctr_snap in
that case, but that is not so good for debugging of RCU internals.)
In addition, if one of the RCU flavor's grace period has stalled, this
will invoke rcu_momentary_dyntick_idle(), resulting in a heavy-weight
quiescent state visible from other CPUs.
Reported-by: Sasha Levin <sasha.levin@oracle.com>
Reported-by: Dave Jones <davej@redhat.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
[ paulmck: Merge commit from Sasha Levin fixing a bug where __this_cpu()
was used in preemptible code. ]
2014-12-14 12:32:04 +08:00
|
|
|
|
|
2014-06-21 07:49:01 +08:00
|
|
|
|
/*
|
|
|
|
|
* Let the RCU core know that this CPU has gone through the scheduler,
|
|
|
|
|
* which is a quiescent state. This is called when the need for a
|
|
|
|
|
* quiescent state is urgent, so we burn an atomic operation and full
|
|
|
|
|
* memory barriers to let the RCU core know about it, regardless of what
|
|
|
|
|
* this CPU might (or might not) do in the near future.
|
|
|
|
|
*
|
2017-01-28 05:17:02 +08:00
|
|
|
|
* We inform the RCU core by emulating a zero-duration dyntick-idle period.
|
2015-10-08 00:10:48 +08:00
|
|
|
|
*
|
2018-05-17 07:01:56 +08:00
|
|
|
|
* The caller must have disabled interrupts and must not be idle.
|
2014-06-21 07:49:01 +08:00
|
|
|
|
*/
|
2020-10-21 15:38:39 +08:00
|
|
|
|
notrace void rcu_momentary_dyntick_idle(void)
|
2014-06-21 07:49:01 +08:00
|
|
|
|
{
|
rcu: Weaken ->dynticks accesses and updates
Accesses to the rcu_data structure's ->dynticks field have always been
fully ordered because it was not possible to prove that weaker ordering
was safe. However, with the removal of the rcu_eqs_special_set() function
and the advent of the Linux-kernel memory model, it is now easy to show
that two of the four original full memory barriers can be weakened to
acquire and release operations. The remaining pair must remain full
memory barriers. This change makes the memory ordering requirements
more evident, and it might well also speed up the to-idle and from-idle
fastpaths on some architectures.
The following litmus test, adapted from one supplied off-list by Frederic
Weisbecker, models the RCU grace-period kthread detecting an idle CPU
that is concurrently transitioning to non-idle:
C dynticks-from-idle
{
DYNTICKS=0; (* Initially idle. *)
}
P0(int *X, int *DYNTICKS)
{
int dynticks;
int x;
// Idle.
dynticks = READ_ONCE(*DYNTICKS);
smp_store_release(DYNTICKS, dynticks + 1);
smp_mb();
// Now non-idle
x = READ_ONCE(*X);
}
P1(int *X, int *DYNTICKS)
{
int dynticks;
WRITE_ONCE(*X, 1);
smp_mb();
dynticks = smp_load_acquire(DYNTICKS);
}
exists (1:dynticks=0 /\ 0:x=1)
Running "herd7 -conf linux-kernel.cfg dynticks-from-idle.litmus" verifies
this transition, namely, showing that if the RCU grace-period kthread (P1)
sees another CPU as idle (P0), then any memory access prior to the start
of the grace period (P1's write to X) will be seen by any RCU read-side
critical section following the to-non-idle transition (P0's read from X).
This is a straightforward use of full memory barriers to force ordering
in a store-buffering (SB) litmus test.
The following litmus test, also adapted from the one supplied off-list
by Frederic Weisbecker, models the RCU grace-period kthread detecting
a non-idle CPU that is concurrently transitioning to idle:
C dynticks-into-idle
{
DYNTICKS=1; (* Initially non-idle. *)
}
P0(int *X, int *DYNTICKS)
{
int dynticks;
// Non-idle.
WRITE_ONCE(*X, 1);
dynticks = READ_ONCE(*DYNTICKS);
smp_store_release(DYNTICKS, dynticks + 1);
smp_mb();
// Now idle.
}
P1(int *X, int *DYNTICKS)
{
int x;
int dynticks;
smp_mb();
dynticks = smp_load_acquire(DYNTICKS);
x = READ_ONCE(*X);
}
exists (1:dynticks=2 /\ 1:x=0)
Running "herd7 -conf linux-kernel.cfg dynticks-into-idle.litmus" verifies
this transition, namely, showing that if the RCU grace-period kthread
(P1) sees another CPU as newly idle (P0), then any pre-idle memory access
(P0's write to X) will be seen by any code following the grace period
(P1's read from X). This is a simple release-acquire pair forcing
ordering in a message-passing (MP) litmus test.
Of course, if the grace-period kthread detects the CPU as non-idle,
it will refrain from reporting a quiescent state on behalf of that CPU,
so there are no ordering requirements from the grace-period kthread in
that case. However, other subsystems call rcu_is_idle_cpu() to check
for CPUs being non-idle from an RCU perspective. That case is also
verified by the above litmus tests with the proviso that the sense of
the low-order bit of the DYNTICKS counter be inverted.
Unfortunately, on x86 smp_mb() is as expensive as a cache-local atomic
increment. This commit therefore weakens only the read from ->dynticks.
However, the updates are abstracted into a rcu_dynticks_inc() function
to ease any future changes that might be needed.
[ paulmck: Apply Linus Torvalds feedback. ]
Link: https://lore.kernel.org/lkml/20210721202127.2129660-4-paulmck@kernel.org/
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-05-20 08:25:42 +08:00
|
|
|
|
int seq;
|
2018-05-17 07:01:56 +08:00
|
|
|
|
|
2018-08-04 12:00:38 +08:00
|
|
|
|
raw_cpu_write(rcu_data.rcu_need_heavy_qs, false);
|
2022-06-08 22:40:35 +08:00
|
|
|
|
seq = ct_state_inc(2 * RCU_DYNTICKS_IDX);
|
2018-05-17 07:01:56 +08:00
|
|
|
|
/* It is illegal to call this from idle state. */
|
2022-06-08 22:40:35 +08:00
|
|
|
|
WARN_ON_ONCE(!(seq & RCU_DYNTICKS_IDX));
|
rcu: Defer reporting RCU-preempt quiescent states when disabled
This commit defers reporting of RCU-preempt quiescent states at
rcu_read_unlock_special() time when any of interrupts, softirq, or
preemption are disabled. These deferred quiescent states are reported
at a later RCU_SOFTIRQ, context switch, idle entry, or CPU-hotplug
offline operation. Of course, if another RCU read-side critical
section has started in the meantime, the reporting of the quiescent
state will be further deferred.
This also means that disabling preemption, interrupts, and/or
softirqs will act as an RCU-preempt read-side critical section.
This is enforced by checking preempt_count() as needed.
Some special cases must be handled on an ad-hoc basis, for example,
context switch is a quiescent state even though both the scheduler and
do_exit() disable preemption. In these cases, additional calls to
rcu_preempt_deferred_qs() override the preemption disabling. Similar
logic overrides disabled interrupts in rcu_preempt_check_callbacks()
because in this case the quiescent state happened just before the
corresponding scheduling-clock interrupt.
In theory, this change lifts a long-standing restriction that required
that if interrupts were disabled across a call to rcu_read_unlock()
that the matching rcu_read_lock() also be contained within that
interrupts-disabled region of code. Because the reporting of the
corresponding RCU-preempt quiescent state is now deferred until
after interrupts have been enabled, it is no longer possible for this
situation to result in deadlocks involving the scheduler's runqueue and
priority-inheritance locks. This may allow some code simplification that
might reduce interrupt latency a bit. Unfortunately, in practice this
would also defer deboosting a low-priority task that had been subjected
to RCU priority boosting, so real-time-response considerations might
well force this restriction to remain in place.
Because RCU-preempt grace periods are now blocked not only by RCU
read-side critical sections, but also by disabling of interrupts,
preemption, and softirqs, it will be possible to eliminate RCU-bh and
RCU-sched in favor of RCU-preempt in CONFIG_PREEMPT=y kernels. This may
require some additional plumbing to provide the network denial-of-service
guarantees that have been traditionally provided by RCU-bh. Once these
are in place, CONFIG_PREEMPT=n kernels will be able to fold RCU-bh
into RCU-sched. This would mean that all kernels would have but
one flavor of RCU, which would open the door to significant code
cleanup.
Moving to a single flavor of RCU would also have the beneficial effect
of reducing the NOCB kthreads by at least a factor of two.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
[ paulmck: Apply rcu_read_unlock_special() preempt_count() feedback
from Joel Fernandes. ]
[ paulmck: Adjust rcu_eqs_enter() call to rcu_preempt_deferred_qs() in
response to bug reports from kbuild test robot. ]
[ paulmck: Fix bug located by kbuild test robot involving recursion
via rcu_preempt_deferred_qs(). ]
2018-06-22 03:50:01 +08:00
|
|
|
|
rcu_preempt_deferred_qs(current);
|
2014-06-21 07:49:01 +08:00
|
|
|
|
}
|
2019-08-05 04:17:35 +08:00
|
|
|
|
EXPORT_SYMBOL_GPL(rcu_momentary_dyntick_idle);
|
2014-06-21 07:49:01 +08:00
|
|
|
|
|
2018-07-03 05:30:37 +08:00
|
|
|
|
/**
|
2020-05-28 01:12:36 +08:00
|
|
|
|
* rcu_is_cpu_rrupt_from_idle - see if 'interrupted' from idle
|
rcu: Don't disable preemption for Tiny and Tree RCU readers
Because preempt_disable() maps to barrier() for non-debug builds,
it forces the compiler to spill and reload registers. Because Tree
RCU and Tiny RCU now only appear in CONFIG_PREEMPT=n builds, these
barrier() instances generate needless extra code for each instance of
rcu_read_lock() and rcu_read_unlock(). This extra code slows down Tree
RCU and bloats Tiny RCU.
This commit therefore removes the preempt_disable() and preempt_enable()
from the non-preemptible implementations of __rcu_read_lock() and
__rcu_read_unlock(), respectively. However, for debug purposes,
preempt_disable() and preempt_enable() are still invoked if
CONFIG_PREEMPT_COUNT=y, because this allows detection of sleeping inside
atomic sections in non-preemptible kernels.
However, Tiny and Tree RCU operates by coalescing all RCU read-side
critical sections on a given CPU that lie between successive quiescent
states. It is therefore necessary to compensate for removing barriers
from __rcu_read_lock() and __rcu_read_unlock() by adding them to a
couple of the RCU functions invoked during quiescent states, namely to
rcu_all_qs() and rcu_note_context_switch(). However, note that the latter
is more paranoia than necessity, at least until link-time optimizations
become more aggressive.
This is based on an earlier patch by Paul E. McKenney, fixing
a bug encountered in kernels built with CONFIG_PREEMPT=n and
CONFIG_PREEMPT_COUNT=y.
Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2015-07-31 07:55:38 +08:00
|
|
|
|
*
|
2019-03-27 03:24:09 +08:00
|
|
|
|
* If the current CPU is idle and running at a first-level (not nested)
|
2020-05-28 01:12:36 +08:00
|
|
|
|
* interrupt, or directly, from idle, return true.
|
|
|
|
|
*
|
|
|
|
|
* The caller must have at least disabled IRQs.
|
rcu: Make cond_resched_rcu_qs() apply to normal RCU flavors
Although cond_resched_rcu_qs() only applies to TASKS_RCU, it is used
in places where it would be useful for it to apply to the normal RCU
flavors, rcu_preempt, rcu_sched, and rcu_bh. This is especially the
case for workloads that aggressively overload the system, particularly
those that generate large numbers of RCU updates on systems running
NO_HZ_FULL CPUs. This commit therefore communicates quiescent states
from cond_resched_rcu_qs() to the normal RCU flavors.
Note that it is unfortunately necessary to leave the old ->passed_quiesce
mechanism in place to allow quiescent states that apply to only one
flavor to be recorded. (Yes, we could decrement ->rcu_qs_ctr_snap in
that case, but that is not so good for debugging of RCU internals.)
In addition, if one of the RCU flavor's grace period has stalled, this
will invoke rcu_momentary_dyntick_idle(), resulting in a heavy-weight
quiescent state visible from other CPUs.
Reported-by: Sasha Levin <sasha.levin@oracle.com>
Reported-by: Dave Jones <davej@redhat.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
[ paulmck: Merge commit from Sasha Levin fixing a bug where __this_cpu()
was used in preemptible code. ]
2014-12-14 12:32:04 +08:00
|
|
|
|
*/
|
2018-07-03 05:30:37 +08:00
|
|
|
|
static int rcu_is_cpu_rrupt_from_idle(void)
|
rcu: Make cond_resched_rcu_qs() apply to normal RCU flavors
Although cond_resched_rcu_qs() only applies to TASKS_RCU, it is used
in places where it would be useful for it to apply to the normal RCU
flavors, rcu_preempt, rcu_sched, and rcu_bh. This is especially the
case for workloads that aggressively overload the system, particularly
those that generate large numbers of RCU updates on systems running
NO_HZ_FULL CPUs. This commit therefore communicates quiescent states
from cond_resched_rcu_qs() to the normal RCU flavors.
Note that it is unfortunately necessary to leave the old ->passed_quiesce
mechanism in place to allow quiescent states that apply to only one
flavor to be recorded. (Yes, we could decrement ->rcu_qs_ctr_snap in
that case, but that is not so good for debugging of RCU internals.)
In addition, if one of the RCU flavor's grace period has stalled, this
will invoke rcu_momentary_dyntick_idle(), resulting in a heavy-weight
quiescent state visible from other CPUs.
Reported-by: Sasha Levin <sasha.levin@oracle.com>
Reported-by: Dave Jones <davej@redhat.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
[ paulmck: Merge commit from Sasha Levin fixing a bug where __this_cpu()
was used in preemptible code. ]
2014-12-14 12:32:04 +08:00
|
|
|
|
{
|
2020-05-28 01:12:36 +08:00
|
|
|
|
long nesting;
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Usually called from the tick; but also used from smp_function_call()
|
|
|
|
|
* for expedited grace periods. This latter can result in running from
|
|
|
|
|
* the idle task, instead of an actual IPI.
|
|
|
|
|
*/
|
|
|
|
|
lockdep_assert_irqs_disabled();
|
2019-03-27 03:24:09 +08:00
|
|
|
|
|
|
|
|
|
/* Check for counter underflows */
|
2022-06-08 22:40:30 +08:00
|
|
|
|
RCU_LOCKDEP_WARN(ct_dynticks_nesting() < 0,
|
2019-03-27 03:24:09 +08:00
|
|
|
|
"RCU dynticks_nesting counter underflow!");
|
2022-06-08 22:40:31 +08:00
|
|
|
|
RCU_LOCKDEP_WARN(ct_dynticks_nmi_nesting() <= 0,
|
2019-03-27 03:24:09 +08:00
|
|
|
|
"RCU dynticks_nmi_nesting counter underflow/zero!");
|
|
|
|
|
|
|
|
|
|
/* Are we at first interrupt nesting level? */
|
2022-06-08 22:40:31 +08:00
|
|
|
|
nesting = ct_dynticks_nmi_nesting();
|
2020-05-28 01:12:36 +08:00
|
|
|
|
if (nesting > 1)
|
2019-03-27 03:24:09 +08:00
|
|
|
|
return false;
|
|
|
|
|
|
2020-05-28 01:12:36 +08:00
|
|
|
|
/*
|
|
|
|
|
* If we're not in an interrupt, we must be in the idle task!
|
|
|
|
|
*/
|
|
|
|
|
WARN_ON_ONCE(!nesting && !is_idle_task(current));
|
|
|
|
|
|
2019-03-27 03:24:09 +08:00
|
|
|
|
/* Does CPU appear to be idle from an RCU standpoint? */
|
2022-06-08 22:40:30 +08:00
|
|
|
|
return ct_dynticks_nesting() == 0;
|
rcu: Make cond_resched_rcu_qs() apply to normal RCU flavors
Although cond_resched_rcu_qs() only applies to TASKS_RCU, it is used
in places where it would be useful for it to apply to the normal RCU
flavors, rcu_preempt, rcu_sched, and rcu_bh. This is especially the
case for workloads that aggressively overload the system, particularly
those that generate large numbers of RCU updates on systems running
NO_HZ_FULL CPUs. This commit therefore communicates quiescent states
from cond_resched_rcu_qs() to the normal RCU flavors.
Note that it is unfortunately necessary to leave the old ->passed_quiesce
mechanism in place to allow quiescent states that apply to only one
flavor to be recorded. (Yes, we could decrement ->rcu_qs_ctr_snap in
that case, but that is not so good for debugging of RCU internals.)
In addition, if one of the RCU flavor's grace period has stalled, this
will invoke rcu_momentary_dyntick_idle(), resulting in a heavy-weight
quiescent state visible from other CPUs.
Reported-by: Sasha Levin <sasha.levin@oracle.com>
Reported-by: Dave Jones <davej@redhat.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
[ paulmck: Merge commit from Sasha Levin fixing a bug where __this_cpu()
was used in preemptible code. ]
2014-12-14 12:32:04 +08:00
|
|
|
|
}
|
|
|
|
|
|
2020-08-06 21:39:30 +08:00
|
|
|
|
#define DEFAULT_RCU_BLIMIT (IS_ENABLED(CONFIG_RCU_STRICT_GRACE_PERIOD) ? 1000 : 10)
|
|
|
|
|
// Maximum callbacks per rcu_do_batch ...
|
|
|
|
|
#define DEFAULT_MAX_RCU_BLIMIT 10000 // ... even during callback flood.
|
2017-04-29 02:12:34 +08:00
|
|
|
|
static long blimit = DEFAULT_RCU_BLIMIT;
|
2020-08-06 21:39:30 +08:00
|
|
|
|
#define DEFAULT_RCU_QHIMARK 10000 // If this many pending, ignore blimit.
|
2017-04-29 02:12:34 +08:00
|
|
|
|
static long qhimark = DEFAULT_RCU_QHIMARK;
|
2020-08-06 21:39:30 +08:00
|
|
|
|
#define DEFAULT_RCU_QLOMARK 100 // Once only this many pending, use blimit.
|
2017-04-29 02:12:34 +08:00
|
|
|
|
static long qlowmark = DEFAULT_RCU_QLOMARK;
|
2019-10-31 02:56:10 +08:00
|
|
|
|
#define DEFAULT_RCU_QOVLD_MULT 2
|
|
|
|
|
#define DEFAULT_RCU_QOVLD (DEFAULT_RCU_QOVLD_MULT * DEFAULT_RCU_QHIMARK)
|
2020-08-06 21:39:30 +08:00
|
|
|
|
static long qovld = DEFAULT_RCU_QOVLD; // If this many pending, hammer QS.
|
|
|
|
|
static long qovld_calc = -1; // No pre-initialization lock acquisitions!
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
2012-10-18 19:55:36 +08:00
|
|
|
|
module_param(blimit, long, 0444);
|
|
|
|
|
module_param(qhimark, long, 0444);
|
|
|
|
|
module_param(qlowmark, long, 0444);
|
2019-10-31 02:56:10 +08:00
|
|
|
|
module_param(qovld, long, 0444);
|
2009-09-28 22:46:32 +08:00
|
|
|
|
|
2020-08-06 08:25:23 +08:00
|
|
|
|
static ulong jiffies_till_first_fqs = IS_ENABLED(CONFIG_RCU_STRICT_GRACE_PERIOD) ? 0 : ULONG_MAX;
|
2013-04-04 13:14:11 +08:00
|
|
|
|
static ulong jiffies_till_next_fqs = ULONG_MAX;
|
2016-01-04 12:29:57 +08:00
|
|
|
|
static bool rcu_kick_kthreads;
|
2019-07-25 09:07:52 +08:00
|
|
|
|
static int rcu_divisor = 7;
|
|
|
|
|
module_param(rcu_divisor, int, 0644);
|
|
|
|
|
|
|
|
|
|
/* Force an exit from rcu_do_batch() after 3 milliseconds. */
|
|
|
|
|
static long rcu_resched_ns = 3 * NSEC_PER_MSEC;
|
|
|
|
|
module_param(rcu_resched_ns, long, 0644);
|
rcu: Control grace-period duration from sysfs
Although almost everyone is well-served by the defaults, some uses of RCU
benefit from shorter grace periods, while others benefit more from the
greater efficiency provided by longer grace periods. Situations requiring
a large number of grace periods to elapse (and wireshark startup has
been called out as an example of this) are helped by lower-latency
grace periods. Furthermore, in some embedded applications, people are
willing to accept a small degradation in update efficiency (due to there
being more of the shorter grace-period operations) in order to gain the
lower latency.
In contrast, those few systems with thousands of CPUs need longer grace
periods because the CPU overhead of a grace period rises roughly
linearly with the number of CPUs. Such systems normally do not make
much use of facilities that require large numbers of grace periods to
elapse, so this is a good tradeoff.
Therefore, this commit allows the durations to be controlled from sysfs.
There are two sysfs parameters, one named "jiffies_till_first_fqs" that
specifies the delay in jiffies from the end of grace-period initialization
until the first attempt to force quiescent states, and the other named
"jiffies_till_next_fqs" that specifies the delay (again in jiffies)
between subsequent attempts to force quiescent states. They both default
to three jiffies, which is compatible with the old hard-coded behavior.
At some future time, it may be possible to automatically increase the
grace-period length with the number of CPUs, but we do not yet have
sufficient data to do a good job. Preliminary data indicates that we
should add an addiitonal jiffy to each of the delays for every 200 CPUs
in the system, but more experimentation is needed. For now, the number
of systems with more than 1,000 CPUs is small enough that this can be
relegated to boot-time hand tuning.
Signed-off-by: Paul E. McKenney <paul.mckenney@linaro.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2012-06-27 11:45:57 +08:00
|
|
|
|
|
rcu: Compute jiffies_till_sched_qs from other kernel parameters
The jiffies_till_sched_qs value used to determine how old a grace period
must be before RCU enlists the help of the scheduler to force a quiescent
state on the holdout CPU. Currently, this defaults to HZ/10 regardless of
system size and may be set only at boot time. This can be a problem for
very large systems, because if the values of the jiffies_till_first_fqs
and jiffies_till_next_fqs kernel parameters are left at their defaults,
they are calculated to increase as the number of CPUs actually configured
on the system increases. Thus, on a sufficiently large system, RCU would
enlist the help of the scheduler before the grace-period kthread had a
chance to scan for idle CPUs, which wastes CPU time.
This commit therefore allows jiffies_till_sched_qs to be set, if desired,
but if left as default, computes is as jiffies_till_first_fqs plus twice
jiffies_till_next_fqs, thus allowing three force-quiescent-state scans
for idle CPUs. This scales with the number of CPUs, providing sensible
default values.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-07-26 02:25:23 +08:00
|
|
|
|
/*
|
|
|
|
|
* How long the grace period must be before we start recruiting
|
|
|
|
|
* quiescent-state help from rcu_note_context_switch().
|
|
|
|
|
*/
|
|
|
|
|
static ulong jiffies_till_sched_qs = ULONG_MAX;
|
|
|
|
|
module_param(jiffies_till_sched_qs, ulong, 0444);
|
2019-03-12 06:45:13 +08:00
|
|
|
|
static ulong jiffies_to_sched_qs; /* See adjust_jiffies_till_sched_qs(). */
|
rcu: Compute jiffies_till_sched_qs from other kernel parameters
The jiffies_till_sched_qs value used to determine how old a grace period
must be before RCU enlists the help of the scheduler to force a quiescent
state on the holdout CPU. Currently, this defaults to HZ/10 regardless of
system size and may be set only at boot time. This can be a problem for
very large systems, because if the values of the jiffies_till_first_fqs
and jiffies_till_next_fqs kernel parameters are left at their defaults,
they are calculated to increase as the number of CPUs actually configured
on the system increases. Thus, on a sufficiently large system, RCU would
enlist the help of the scheduler before the grace-period kthread had a
chance to scan for idle CPUs, which wastes CPU time.
This commit therefore allows jiffies_till_sched_qs to be set, if desired,
but if left as default, computes is as jiffies_till_first_fqs plus twice
jiffies_till_next_fqs, thus allowing three force-quiescent-state scans
for idle CPUs. This scales with the number of CPUs, providing sensible
default values.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-07-26 02:25:23 +08:00
|
|
|
|
module_param(jiffies_to_sched_qs, ulong, 0444); /* Display only! */
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Make sure that we give the grace-period kthread time to detect any
|
|
|
|
|
* idle CPUs before taking active measures to force quiescent states.
|
|
|
|
|
* However, don't go below 100 milliseconds, adjusted upwards for really
|
|
|
|
|
* large systems.
|
|
|
|
|
*/
|
|
|
|
|
static void adjust_jiffies_till_sched_qs(void)
|
|
|
|
|
{
|
|
|
|
|
unsigned long j;
|
|
|
|
|
|
|
|
|
|
/* If jiffies_till_sched_qs was specified, respect the request. */
|
|
|
|
|
if (jiffies_till_sched_qs != ULONG_MAX) {
|
|
|
|
|
WRITE_ONCE(jiffies_to_sched_qs, jiffies_till_sched_qs);
|
|
|
|
|
return;
|
|
|
|
|
}
|
2019-03-12 06:45:13 +08:00
|
|
|
|
/* Otherwise, set to third fqs scan, but bound below on large system. */
|
rcu: Compute jiffies_till_sched_qs from other kernel parameters
The jiffies_till_sched_qs value used to determine how old a grace period
must be before RCU enlists the help of the scheduler to force a quiescent
state on the holdout CPU. Currently, this defaults to HZ/10 regardless of
system size and may be set only at boot time. This can be a problem for
very large systems, because if the values of the jiffies_till_first_fqs
and jiffies_till_next_fqs kernel parameters are left at their defaults,
they are calculated to increase as the number of CPUs actually configured
on the system increases. Thus, on a sufficiently large system, RCU would
enlist the help of the scheduler before the grace-period kthread had a
chance to scan for idle CPUs, which wastes CPU time.
This commit therefore allows jiffies_till_sched_qs to be set, if desired,
but if left as default, computes is as jiffies_till_first_fqs plus twice
jiffies_till_next_fqs, thus allowing three force-quiescent-state scans
for idle CPUs. This scales with the number of CPUs, providing sensible
default values.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-07-26 02:25:23 +08:00
|
|
|
|
j = READ_ONCE(jiffies_till_first_fqs) +
|
|
|
|
|
2 * READ_ONCE(jiffies_till_next_fqs);
|
|
|
|
|
if (j < HZ / 10 + nr_cpu_ids / RCU_JIFFIES_FQS_DIV)
|
|
|
|
|
j = HZ / 10 + nr_cpu_ids / RCU_JIFFIES_FQS_DIV;
|
|
|
|
|
pr_info("RCU calculated value of scheduler-enlistment delay is %ld jiffies.\n", j);
|
|
|
|
|
WRITE_ONCE(jiffies_to_sched_qs, j);
|
|
|
|
|
}
|
|
|
|
|
|
2018-06-01 10:03:09 +08:00
|
|
|
|
static int param_set_first_fqs_jiffies(const char *val, const struct kernel_param *kp)
|
|
|
|
|
{
|
|
|
|
|
ulong j;
|
|
|
|
|
int ret = kstrtoul(val, 0, &j);
|
|
|
|
|
|
rcu: Compute jiffies_till_sched_qs from other kernel parameters
The jiffies_till_sched_qs value used to determine how old a grace period
must be before RCU enlists the help of the scheduler to force a quiescent
state on the holdout CPU. Currently, this defaults to HZ/10 regardless of
system size and may be set only at boot time. This can be a problem for
very large systems, because if the values of the jiffies_till_first_fqs
and jiffies_till_next_fqs kernel parameters are left at their defaults,
they are calculated to increase as the number of CPUs actually configured
on the system increases. Thus, on a sufficiently large system, RCU would
enlist the help of the scheduler before the grace-period kthread had a
chance to scan for idle CPUs, which wastes CPU time.
This commit therefore allows jiffies_till_sched_qs to be set, if desired,
but if left as default, computes is as jiffies_till_first_fqs plus twice
jiffies_till_next_fqs, thus allowing three force-quiescent-state scans
for idle CPUs. This scales with the number of CPUs, providing sensible
default values.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-07-26 02:25:23 +08:00
|
|
|
|
if (!ret) {
|
2018-06-01 10:03:09 +08:00
|
|
|
|
WRITE_ONCE(*(ulong *)kp->arg, (j > HZ) ? HZ : j);
|
rcu: Compute jiffies_till_sched_qs from other kernel parameters
The jiffies_till_sched_qs value used to determine how old a grace period
must be before RCU enlists the help of the scheduler to force a quiescent
state on the holdout CPU. Currently, this defaults to HZ/10 regardless of
system size and may be set only at boot time. This can be a problem for
very large systems, because if the values of the jiffies_till_first_fqs
and jiffies_till_next_fqs kernel parameters are left at their defaults,
they are calculated to increase as the number of CPUs actually configured
on the system increases. Thus, on a sufficiently large system, RCU would
enlist the help of the scheduler before the grace-period kthread had a
chance to scan for idle CPUs, which wastes CPU time.
This commit therefore allows jiffies_till_sched_qs to be set, if desired,
but if left as default, computes is as jiffies_till_first_fqs plus twice
jiffies_till_next_fqs, thus allowing three force-quiescent-state scans
for idle CPUs. This scales with the number of CPUs, providing sensible
default values.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-07-26 02:25:23 +08:00
|
|
|
|
adjust_jiffies_till_sched_qs();
|
|
|
|
|
}
|
2018-06-01 10:03:09 +08:00
|
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int param_set_next_fqs_jiffies(const char *val, const struct kernel_param *kp)
|
|
|
|
|
{
|
|
|
|
|
ulong j;
|
|
|
|
|
int ret = kstrtoul(val, 0, &j);
|
|
|
|
|
|
rcu: Compute jiffies_till_sched_qs from other kernel parameters
The jiffies_till_sched_qs value used to determine how old a grace period
must be before RCU enlists the help of the scheduler to force a quiescent
state on the holdout CPU. Currently, this defaults to HZ/10 regardless of
system size and may be set only at boot time. This can be a problem for
very large systems, because if the values of the jiffies_till_first_fqs
and jiffies_till_next_fqs kernel parameters are left at their defaults,
they are calculated to increase as the number of CPUs actually configured
on the system increases. Thus, on a sufficiently large system, RCU would
enlist the help of the scheduler before the grace-period kthread had a
chance to scan for idle CPUs, which wastes CPU time.
This commit therefore allows jiffies_till_sched_qs to be set, if desired,
but if left as default, computes is as jiffies_till_first_fqs plus twice
jiffies_till_next_fqs, thus allowing three force-quiescent-state scans
for idle CPUs. This scales with the number of CPUs, providing sensible
default values.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-07-26 02:25:23 +08:00
|
|
|
|
if (!ret) {
|
2018-06-01 10:03:09 +08:00
|
|
|
|
WRITE_ONCE(*(ulong *)kp->arg, (j > HZ) ? HZ : (j ?: 1));
|
rcu: Compute jiffies_till_sched_qs from other kernel parameters
The jiffies_till_sched_qs value used to determine how old a grace period
must be before RCU enlists the help of the scheduler to force a quiescent
state on the holdout CPU. Currently, this defaults to HZ/10 regardless of
system size and may be set only at boot time. This can be a problem for
very large systems, because if the values of the jiffies_till_first_fqs
and jiffies_till_next_fqs kernel parameters are left at their defaults,
they are calculated to increase as the number of CPUs actually configured
on the system increases. Thus, on a sufficiently large system, RCU would
enlist the help of the scheduler before the grace-period kthread had a
chance to scan for idle CPUs, which wastes CPU time.
This commit therefore allows jiffies_till_sched_qs to be set, if desired,
but if left as default, computes is as jiffies_till_first_fqs plus twice
jiffies_till_next_fqs, thus allowing three force-quiescent-state scans
for idle CPUs. This scales with the number of CPUs, providing sensible
default values.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-07-26 02:25:23 +08:00
|
|
|
|
adjust_jiffies_till_sched_qs();
|
|
|
|
|
}
|
2018-06-01 10:03:09 +08:00
|
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
2020-10-04 08:18:08 +08:00
|
|
|
|
static const struct kernel_param_ops first_fqs_jiffies_ops = {
|
2018-06-01 10:03:09 +08:00
|
|
|
|
.set = param_set_first_fqs_jiffies,
|
|
|
|
|
.get = param_get_ulong,
|
|
|
|
|
};
|
|
|
|
|
|
2020-10-04 08:18:08 +08:00
|
|
|
|
static const struct kernel_param_ops next_fqs_jiffies_ops = {
|
2018-06-01 10:03:09 +08:00
|
|
|
|
.set = param_set_next_fqs_jiffies,
|
|
|
|
|
.get = param_get_ulong,
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
module_param_cb(jiffies_till_first_fqs, &first_fqs_jiffies_ops, &jiffies_till_first_fqs, 0644);
|
|
|
|
|
module_param_cb(jiffies_till_next_fqs, &next_fqs_jiffies_ops, &jiffies_till_next_fqs, 0644);
|
2016-01-04 12:29:57 +08:00
|
|
|
|
module_param(rcu_kick_kthreads, bool, 0644);
|
rcu: Control grace-period duration from sysfs
Although almost everyone is well-served by the defaults, some uses of RCU
benefit from shorter grace periods, while others benefit more from the
greater efficiency provided by longer grace periods. Situations requiring
a large number of grace periods to elapse (and wireshark startup has
been called out as an example of this) are helped by lower-latency
grace periods. Furthermore, in some embedded applications, people are
willing to accept a small degradation in update efficiency (due to there
being more of the shorter grace-period operations) in order to gain the
lower latency.
In contrast, those few systems with thousands of CPUs need longer grace
periods because the CPU overhead of a grace period rises roughly
linearly with the number of CPUs. Such systems normally do not make
much use of facilities that require large numbers of grace periods to
elapse, so this is a good tradeoff.
Therefore, this commit allows the durations to be controlled from sysfs.
There are two sysfs parameters, one named "jiffies_till_first_fqs" that
specifies the delay in jiffies from the end of grace-period initialization
until the first attempt to force quiescent states, and the other named
"jiffies_till_next_fqs" that specifies the delay (again in jiffies)
between subsequent attempts to force quiescent states. They both default
to three jiffies, which is compatible with the old hard-coded behavior.
At some future time, it may be possible to automatically increase the
grace-period length with the number of CPUs, but we do not yet have
sufficient data to do a good job. Preliminary data indicates that we
should add an addiitonal jiffy to each of the delays for every 200 CPUs
in the system, but more experimentation is needed. For now, the number
of systems with more than 1,000 CPUs is small enough that this can be
relegated to boot-time hand tuning.
Signed-off-by: Paul E. McKenney <paul.mckenney@linaro.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2012-06-27 11:45:57 +08:00
|
|
|
|
|
2018-07-06 08:55:14 +08:00
|
|
|
|
static void force_qs_rnp(int (*f)(struct rcu_data *rdp));
|
2019-09-14 18:39:22 +08:00
|
|
|
|
static int rcu_pending(int user);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
|
|
|
|
/*
|
2018-04-28 02:39:34 +08:00
|
|
|
|
* Return the number of RCU GPs completed thus far for debug & stats.
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
*/
|
2018-04-28 02:39:34 +08:00
|
|
|
|
unsigned long rcu_get_gp_seq(void)
|
2014-11-22 09:10:16 +08:00
|
|
|
|
{
|
2018-07-04 06:54:39 +08:00
|
|
|
|
return READ_ONCE(rcu_state.gp_seq);
|
2014-11-22 09:10:16 +08:00
|
|
|
|
}
|
2018-04-28 02:39:34 +08:00
|
|
|
|
EXPORT_SYMBOL_GPL(rcu_get_gp_seq);
|
2014-11-22 09:10:16 +08:00
|
|
|
|
|
2016-01-13 05:43:30 +08:00
|
|
|
|
/*
|
|
|
|
|
* Return the number of RCU expedited batches completed thus far for
|
|
|
|
|
* debug & stats. Odd numbers mean that a batch is in progress, even
|
|
|
|
|
* numbers mean idle. The value returned will thus be roughly double
|
|
|
|
|
* the cumulative batches since boot.
|
|
|
|
|
*/
|
|
|
|
|
unsigned long rcu_exp_batches_completed(void)
|
|
|
|
|
{
|
2018-07-04 06:54:39 +08:00
|
|
|
|
return rcu_state.expedited_sequence;
|
2016-01-13 05:43:30 +08:00
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(rcu_exp_batches_completed);
|
|
|
|
|
|
2018-12-11 08:09:49 +08:00
|
|
|
|
/*
|
|
|
|
|
* Return the root node of the rcu_state structure.
|
|
|
|
|
*/
|
|
|
|
|
static struct rcu_node *rcu_get_root(void)
|
|
|
|
|
{
|
|
|
|
|
return &rcu_state.node[0];
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-20 02:51:42 +08:00
|
|
|
|
/*
|
|
|
|
|
* Send along grace-period-related data for rcutorture diagnostics.
|
|
|
|
|
*/
|
|
|
|
|
void rcutorture_get_gp_data(enum rcutorture_type test_type, int *flags,
|
2018-05-01 21:42:51 +08:00
|
|
|
|
unsigned long *gp_seq)
|
2014-02-20 02:51:42 +08:00
|
|
|
|
{
|
|
|
|
|
switch (test_type) {
|
|
|
|
|
case RCU_FLAVOR:
|
2018-07-05 06:39:40 +08:00
|
|
|
|
*flags = READ_ONCE(rcu_state.gp_flags);
|
|
|
|
|
*gp_seq = rcu_seq_current(&rcu_state.gp_seq);
|
2014-02-20 02:51:42 +08:00
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(rcutorture_get_gp_data);
|
|
|
|
|
|
2022-06-08 22:40:33 +08:00
|
|
|
|
#if defined(CONFIG_NO_HZ_FULL) && (!defined(CONFIG_GENERIC_ENTRY) || !defined(CONFIG_KVM_XFER_TO_GUEST_WORK))
|
2021-02-01 07:05:46 +08:00
|
|
|
|
/*
|
|
|
|
|
* An empty function that will trigger a reschedule on
|
2021-02-01 07:05:48 +08:00
|
|
|
|
* IRQ tail once IRQs get re-enabled on userspace/guest resume.
|
2021-02-01 07:05:46 +08:00
|
|
|
|
*/
|
|
|
|
|
static void late_wakeup_func(struct irq_work *work)
|
|
|
|
|
{
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static DEFINE_PER_CPU(struct irq_work, late_wakeup_work) =
|
|
|
|
|
IRQ_WORK_INIT(late_wakeup_func);
|
|
|
|
|
|
2021-02-01 07:05:48 +08:00
|
|
|
|
/*
|
|
|
|
|
* If either:
|
|
|
|
|
*
|
|
|
|
|
* 1) the task is about to enter in guest mode and $ARCH doesn't support KVM generic work
|
|
|
|
|
* 2) the task is about to enter in user mode and $ARCH doesn't support generic entry.
|
|
|
|
|
*
|
|
|
|
|
* In these cases the late RCU wake ups aren't supported in the resched loops and our
|
|
|
|
|
* last resort is to fire a local irq_work that will trigger a reschedule once IRQs
|
|
|
|
|
* get re-enabled again.
|
|
|
|
|
*/
|
2022-06-08 22:40:32 +08:00
|
|
|
|
noinstr void rcu_irq_work_resched(void)
|
2021-02-01 07:05:48 +08:00
|
|
|
|
{
|
|
|
|
|
struct rcu_data *rdp = this_cpu_ptr(&rcu_data);
|
|
|
|
|
|
|
|
|
|
if (IS_ENABLED(CONFIG_GENERIC_ENTRY) && !(current->flags & PF_VCPU))
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
if (IS_ENABLED(CONFIG_KVM_XFER_TO_GUEST_WORK) && (current->flags & PF_VCPU))
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
instrumentation_begin();
|
|
|
|
|
if (do_nocb_deferred_wakeup(rdp) && need_resched()) {
|
|
|
|
|
irq_work_queue(this_cpu_ptr(&late_wakeup_work));
|
|
|
|
|
}
|
|
|
|
|
instrumentation_end();
|
|
|
|
|
}
|
2022-06-08 22:40:33 +08:00
|
|
|
|
#endif /* #if defined(CONFIG_NO_HZ_FULL) && (!defined(CONFIG_GENERIC_ENTRY) || !defined(CONFIG_KVM_XFER_TO_GUEST_WORK)) */
|
2015-10-31 15:59:01 +08:00
|
|
|
|
|
2020-05-22 04:05:16 +08:00
|
|
|
|
#ifdef CONFIG_PROVE_RCU
|
|
|
|
|
/**
|
|
|
|
|
* rcu_irq_exit_check_preempt - Validate that scheduling is possible
|
|
|
|
|
*/
|
|
|
|
|
void rcu_irq_exit_check_preempt(void)
|
|
|
|
|
{
|
|
|
|
|
lockdep_assert_irqs_disabled();
|
|
|
|
|
|
2022-06-08 22:40:30 +08:00
|
|
|
|
RCU_LOCKDEP_WARN(ct_dynticks_nesting() <= 0,
|
2020-05-22 04:05:16 +08:00
|
|
|
|
"RCU dynticks_nesting counter underflow/zero!");
|
2022-06-08 22:40:31 +08:00
|
|
|
|
RCU_LOCKDEP_WARN(ct_dynticks_nmi_nesting() !=
|
2020-05-22 04:05:16 +08:00
|
|
|
|
DYNTICK_IRQ_NONIDLE,
|
|
|
|
|
"Bad RCU dynticks_nmi_nesting counter\n");
|
|
|
|
|
RCU_LOCKDEP_WARN(rcu_dynticks_curr_cpu_in_eqs(),
|
|
|
|
|
"RCU in extended quiescent state!");
|
|
|
|
|
}
|
|
|
|
|
#endif /* #ifdef CONFIG_PROVE_RCU */
|
|
|
|
|
|
2015-05-14 01:41:58 +08:00
|
|
|
|
#ifdef CONFIG_NO_HZ_FULL
|
2020-05-22 04:05:15 +08:00
|
|
|
|
/**
|
|
|
|
|
* __rcu_irq_enter_check_tick - Enable scheduler tick on CPU if RCU needs it.
|
|
|
|
|
*
|
|
|
|
|
* The scheduler tick is not normally enabled when CPUs enter the kernel
|
|
|
|
|
* from nohz_full userspace execution. After all, nohz_full userspace
|
|
|
|
|
* execution is an RCU quiescent state and the time executing in the kernel
|
|
|
|
|
* is quite short. Except of course when it isn't. And it is not hard to
|
|
|
|
|
* cause a large system to spend tens of seconds or even minutes looping
|
|
|
|
|
* in the kernel, which can cause a number of problems, include RCU CPU
|
|
|
|
|
* stall warnings.
|
|
|
|
|
*
|
|
|
|
|
* Therefore, if a nohz_full CPU fails to report a quiescent state
|
|
|
|
|
* in a timely manner, the RCU grace-period kthread sets that CPU's
|
|
|
|
|
* ->rcu_urgent_qs flag with the expectation that the next interrupt or
|
|
|
|
|
* exception will invoke this function, which will turn on the scheduler
|
|
|
|
|
* tick, which will enable RCU to detect that CPU's quiescent states,
|
|
|
|
|
* for example, due to cond_resched() calls in CONFIG_PREEMPT=n kernels.
|
|
|
|
|
* The tick will be disabled once a quiescent state is reported for
|
|
|
|
|
* this CPU.
|
|
|
|
|
*
|
|
|
|
|
* Of course, in carefully tuned systems, there might never be an
|
|
|
|
|
* interrupt or exception. In that case, the RCU grace-period kthread
|
|
|
|
|
* will eventually cause one to happen. However, in less carefully
|
|
|
|
|
* controlled environments, this function allows RCU to get what it
|
|
|
|
|
* needs without creating otherwise useless interruptions.
|
|
|
|
|
*/
|
|
|
|
|
void __rcu_irq_enter_check_tick(void)
|
|
|
|
|
{
|
|
|
|
|
struct rcu_data *rdp = this_cpu_ptr(&rcu_data);
|
|
|
|
|
|
2020-11-16 20:10:12 +08:00
|
|
|
|
// If we're here from NMI there's nothing to do.
|
|
|
|
|
if (in_nmi())
|
2020-05-22 04:05:15 +08:00
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
RCU_LOCKDEP_WARN(rcu_dynticks_curr_cpu_in_eqs(),
|
|
|
|
|
"Illegal rcu_irq_enter_check_tick() from extended quiescent state");
|
|
|
|
|
|
|
|
|
|
if (!tick_nohz_full_cpu(rdp->cpu) ||
|
|
|
|
|
!READ_ONCE(rdp->rcu_urgent_qs) ||
|
|
|
|
|
READ_ONCE(rdp->rcu_forced_tick)) {
|
|
|
|
|
// RCU doesn't need nohz_full help from this CPU, or it is
|
|
|
|
|
// already getting that help.
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// We get here only when not in an extended quiescent state and
|
|
|
|
|
// from interrupts (as opposed to NMIs). Therefore, (1) RCU is
|
|
|
|
|
// already watching and (2) The fact that we are in an interrupt
|
|
|
|
|
// handler and that the rcu_node lock is an irq-disabled lock
|
|
|
|
|
// prevents self-deadlock. So we can safely recheck under the lock.
|
|
|
|
|
// Note that the nohz_full state currently cannot change.
|
|
|
|
|
raw_spin_lock_rcu_node(rdp->mynode);
|
2023-06-27 09:37:05 +08:00
|
|
|
|
if (READ_ONCE(rdp->rcu_urgent_qs) && !rdp->rcu_forced_tick) {
|
2020-05-22 04:05:15 +08:00
|
|
|
|
// A nohz_full CPU is in the kernel and RCU needs a
|
|
|
|
|
// quiescent state. Turn on the tick!
|
|
|
|
|
WRITE_ONCE(rdp->rcu_forced_tick, true);
|
|
|
|
|
tick_dep_set_cpu(rdp->cpu, TICK_DEP_BIT_RCU);
|
|
|
|
|
}
|
|
|
|
|
raw_spin_unlock_rcu_node(rdp->mynode);
|
|
|
|
|
}
|
2023-01-06 15:09:34 +08:00
|
|
|
|
NOKPROBE_SYMBOL(__rcu_irq_enter_check_tick);
|
2015-05-14 01:41:58 +08:00
|
|
|
|
#endif /* CONFIG_NO_HZ_FULL */
|
rcu: New rcu_user_enter_after_irq() and rcu_user_exit_after_irq() APIs
In some cases, it is necessary to enter or exit userspace-RCU-idle mode
from an interrupt handler, for example, if some other CPU sends this
CPU a resched IPI. In this case, the current CPU would enter the IPI
handler in userspace-RCU-idle mode, but would need to exit the IPI handler
after having exited that mode.
To allow this to work, this commit adds two new APIs to TREE_RCU:
- rcu_user_enter_after_irq(). This must be called from an interrupt between
rcu_irq_enter() and rcu_irq_exit(). After the irq calls rcu_irq_exit(),
the irq handler will return into an RCU extended quiescent state.
In theory, this interrupt is never a nested interrupt, but in practice
it might interrupt softirq, which looks to RCU like a nested interrupt.
- rcu_user_exit_after_irq(). This must be called from a non-nesting
interrupt, interrupting an RCU extended quiescent state, also
between rcu_irq_enter() and rcu_irq_exit(). After the irq calls
rcu_irq_exit(), the irq handler will return in an RCU non-quiescent
state.
[ Combined with "Allow calls to rcu_exit_user_irq from nesting irqs." ]
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2012-06-05 07:42:35 +08:00
|
|
|
|
|
2021-09-28 05:30:20 +08:00
|
|
|
|
/*
|
|
|
|
|
* Check to see if any future non-offloaded RCU-related work will need
|
|
|
|
|
* to be done by the current CPU, even if none need be done immediately,
|
|
|
|
|
* returning 1 if so. This function is part of the RCU implementation;
|
|
|
|
|
* it is -not- an exported member of the RCU API. This is used by
|
|
|
|
|
* the idle-entry code to figure out whether it is safe to disable the
|
|
|
|
|
* scheduler-clock interrupt.
|
|
|
|
|
*
|
|
|
|
|
* Just check whether or not this CPU has non-offloaded RCU callbacks
|
|
|
|
|
* queued.
|
|
|
|
|
*/
|
2022-02-09 00:16:33 +08:00
|
|
|
|
int rcu_needs_cpu(void)
|
2021-09-28 05:30:20 +08:00
|
|
|
|
{
|
|
|
|
|
return !rcu_segcblist_empty(&this_cpu_ptr(&rcu_data)->cblist) &&
|
|
|
|
|
!rcu_rdp_is_offloaded(this_cpu_ptr(&rcu_data));
|
|
|
|
|
}
|
|
|
|
|
|
2019-08-13 07:14:00 +08:00
|
|
|
|
/*
|
2019-09-06 01:26:41 +08:00
|
|
|
|
* If any sort of urgency was applied to the current CPU (for example,
|
|
|
|
|
* the scheduler-clock interrupt was enabled on a nohz_full CPU) in order
|
|
|
|
|
* to get to a quiescent state, disable it.
|
2019-08-13 07:14:00 +08:00
|
|
|
|
*/
|
2019-09-06 01:26:41 +08:00
|
|
|
|
static void rcu_disable_urgency_upon_qs(struct rcu_data *rdp)
|
2019-08-13 07:14:00 +08:00
|
|
|
|
{
|
2019-11-27 10:05:45 +08:00
|
|
|
|
raw_lockdep_assert_held_rcu_node(rdp->mynode);
|
2019-09-06 01:26:41 +08:00
|
|
|
|
WRITE_ONCE(rdp->rcu_urgent_qs, false);
|
|
|
|
|
WRITE_ONCE(rdp->rcu_need_heavy_qs, false);
|
2019-08-13 07:14:00 +08:00
|
|
|
|
if (tick_nohz_full_cpu(rdp->cpu) && rdp->rcu_forced_tick) {
|
|
|
|
|
tick_dep_clear_cpu(rdp->cpu, TICK_DEP_BIT_RCU);
|
2020-01-09 12:06:25 +08:00
|
|
|
|
WRITE_ONCE(rdp->rcu_forced_tick, false);
|
2019-08-13 07:14:00 +08:00
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2013-09-14 08:20:11 +08:00
|
|
|
|
/**
|
2023-05-23 17:22:10 +08:00
|
|
|
|
* rcu_is_watching - RCU read-side critical sections permitted on current CPU?
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
*
|
2023-05-23 17:22:10 +08:00
|
|
|
|
* Return @true if RCU is watching the running CPU and @false otherwise.
|
|
|
|
|
* An @true return means that this CPU can safely enter RCU read-side
|
|
|
|
|
* critical sections.
|
|
|
|
|
*
|
|
|
|
|
* Although calls to rcu_is_watching() from most parts of the kernel
|
|
|
|
|
* will return @true, there are important exceptions. For example, if the
|
|
|
|
|
* current CPU is deep within its idle loop, in kernel entry/exit code,
|
|
|
|
|
* or offline, rcu_is_watching() will return @false.
|
2020-09-29 19:33:40 +08:00
|
|
|
|
*
|
|
|
|
|
* Make notrace because it can be called by the internal functions of
|
|
|
|
|
* ftrace, and making this notrace removes unnecessary recursion calls.
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
*/
|
2020-09-29 19:33:40 +08:00
|
|
|
|
notrace bool rcu_is_watching(void)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
{
|
2014-07-09 06:26:11 +08:00
|
|
|
|
bool ret;
|
2011-10-04 02:38:52 +08:00
|
|
|
|
|
2015-06-17 01:35:18 +08:00
|
|
|
|
preempt_disable_notrace();
|
2017-05-04 02:06:05 +08:00
|
|
|
|
ret = !rcu_dynticks_curr_cpu_in_eqs();
|
2015-06-17 01:35:18 +08:00
|
|
|
|
preempt_enable_notrace();
|
2011-10-04 02:38:52 +08:00
|
|
|
|
return ret;
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
2013-09-14 08:20:11 +08:00
|
|
|
|
EXPORT_SYMBOL_GPL(rcu_is_watching);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
2017-04-12 06:50:41 +08:00
|
|
|
|
/*
|
|
|
|
|
* If a holdout task is actually running, request an urgent quiescent
|
|
|
|
|
* state from its CPU. This is unsynchronized, so migrations can cause
|
|
|
|
|
* the request to go to the wrong CPU. Which is OK, all that will happen
|
|
|
|
|
* is that the CPU's next context switch will be a bit slower and next
|
|
|
|
|
* time around this task will generate another request.
|
|
|
|
|
*/
|
|
|
|
|
void rcu_request_urgent_qs_task(struct task_struct *t)
|
|
|
|
|
{
|
|
|
|
|
int cpu;
|
|
|
|
|
|
|
|
|
|
barrier();
|
|
|
|
|
cpu = task_cpu(t);
|
|
|
|
|
if (!task_curr(t))
|
|
|
|
|
return; /* This task is not running on that CPU. */
|
2018-08-04 12:00:38 +08:00
|
|
|
|
smp_store_release(per_cpu_ptr(&rcu_data.rcu_urgent_qs, cpu), true);
|
2017-04-12 06:50:41 +08:00
|
|
|
|
}
|
|
|
|
|
|
2017-08-18 08:05:59 +08:00
|
|
|
|
/*
|
2021-03-31 04:47:42 +08:00
|
|
|
|
* When trying to report a quiescent state on behalf of some other CPU,
|
2017-08-18 08:05:59 +08:00
|
|
|
|
* it is our responsibility to check for and handle potential overflow
|
2018-04-28 09:06:08 +08:00
|
|
|
|
* of the rcu_node ->gp_seq counter with respect to the rcu_data counters.
|
2017-08-18 08:05:59 +08:00
|
|
|
|
* After all, the CPU might be in deep idle state, and thus executing no
|
|
|
|
|
* code whatsoever.
|
|
|
|
|
*/
|
|
|
|
|
static void rcu_gpnum_ovf(struct rcu_node *rnp, struct rcu_data *rdp)
|
|
|
|
|
{
|
2018-01-17 22:24:30 +08:00
|
|
|
|
raw_lockdep_assert_held_rcu_node(rnp);
|
2018-04-28 09:06:08 +08:00
|
|
|
|
if (ULONG_CMP_LT(rcu_seq_current(&rdp->gp_seq) + ULONG_MAX / 4,
|
|
|
|
|
rnp->gp_seq))
|
2017-08-18 08:05:59 +08:00
|
|
|
|
WRITE_ONCE(rdp->gpwrap, true);
|
2018-04-29 05:15:40 +08:00
|
|
|
|
if (ULONG_CMP_LT(rdp->rcu_iw_gp_seq + ULONG_MAX / 4, rnp->gp_seq))
|
|
|
|
|
rdp->rcu_iw_gp_seq = rnp->gp_seq + ULONG_MAX / 4;
|
2017-08-18 08:05:59 +08:00
|
|
|
|
}
|
|
|
|
|
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
/*
|
|
|
|
|
* Snapshot the specified CPU's dynticks counter so that we can later
|
|
|
|
|
* credit them with an implicit quiescent state. Return 1 if this CPU
|
2009-09-24 00:50:42 +08:00
|
|
|
|
* is in dynticks idle mode, which is an extended quiescent state.
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
*/
|
2017-05-12 02:26:22 +08:00
|
|
|
|
static int dyntick_save_progress_counter(struct rcu_data *rdp)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
{
|
2022-06-08 22:40:29 +08:00
|
|
|
|
rdp->dynticks_snap = rcu_dynticks_snap(rdp->cpu);
|
2016-11-03 08:25:06 +08:00
|
|
|
|
if (rcu_dynticks_in_eqs(rdp->dynticks_snap)) {
|
2018-07-05 05:45:00 +08:00
|
|
|
|
trace_rcu_fqs(rcu_state.name, rdp->gp_seq, rdp->cpu, TPS("dti"));
|
2017-08-18 08:05:59 +08:00
|
|
|
|
rcu_gpnum_ovf(rdp->mynode, rdp);
|
2015-12-14 00:57:10 +08:00
|
|
|
|
return 1;
|
2014-03-18 00:33:28 +08:00
|
|
|
|
}
|
2015-12-14 00:57:10 +08:00
|
|
|
|
return 0;
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
2023-10-31 16:53:08 +08:00
|
|
|
|
* Returns positive if the specified CPU has passed through a quiescent state
|
|
|
|
|
* by virtue of being in or having passed through an dynticks idle state since
|
|
|
|
|
* the last call to dyntick_save_progress_counter() for this same CPU, or by
|
|
|
|
|
* virtue of having been offline.
|
|
|
|
|
*
|
|
|
|
|
* Returns negative if the specified CPU needs a force resched.
|
|
|
|
|
*
|
|
|
|
|
* Returns zero otherwise.
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
*/
|
2017-05-12 02:26:22 +08:00
|
|
|
|
static int rcu_implicit_dynticks_qs(struct rcu_data *rdp)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
{
|
2016-12-01 03:21:21 +08:00
|
|
|
|
unsigned long jtsq;
|
2023-10-31 16:53:08 +08:00
|
|
|
|
int ret = 0;
|
2017-08-18 08:05:59 +08:00
|
|
|
|
struct rcu_node *rnp = rdp->mynode;
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* If the CPU passed through or entered a dynticks idle phase with
|
|
|
|
|
* no active irq/NMI handlers, then we can safely pretend that the CPU
|
|
|
|
|
* already acknowledged the request to pass through a quiescent
|
|
|
|
|
* state. Either way, that CPU cannot possibly be in an RCU
|
|
|
|
|
* read-side critical section that started before the beginning
|
|
|
|
|
* of the current RCU grace period.
|
|
|
|
|
*/
|
2018-08-04 12:00:38 +08:00
|
|
|
|
if (rcu_dynticks_in_eqs_since(rdp, rdp->dynticks_snap)) {
|
2018-07-05 05:45:00 +08:00
|
|
|
|
trace_rcu_fqs(rcu_state.name, rdp->gp_seq, rdp->cpu, TPS("dti"));
|
2017-08-18 08:05:59 +08:00
|
|
|
|
rcu_gpnum_ovf(rnp, rdp);
|
2016-12-01 03:21:21 +08:00
|
|
|
|
return 1;
|
|
|
|
|
}
|
|
|
|
|
|
2020-08-08 01:07:20 +08:00
|
|
|
|
/*
|
|
|
|
|
* Complain if a CPU that is considered to be offline from RCU's
|
|
|
|
|
* perspective has not yet reported a quiescent state. After all,
|
|
|
|
|
* the offline CPU should have reported a quiescent state during
|
|
|
|
|
* the CPU-offline process, or, failing that, by rcu_gp_init()
|
|
|
|
|
* if it ran concurrently with either the CPU going offline or the
|
|
|
|
|
* last task on a leaf rcu_node structure exiting its RCU read-side
|
|
|
|
|
* critical section while all CPUs corresponding to that structure
|
|
|
|
|
* are offline. This added warning detects bugs in any of these
|
|
|
|
|
* code paths.
|
|
|
|
|
*
|
|
|
|
|
* The rcu_node structure's ->lock is held here, which excludes
|
|
|
|
|
* the relevant portions the CPU-hotplug code, the grace-period
|
|
|
|
|
* initialization code, and the rcu_read_unlock() code paths.
|
|
|
|
|
*
|
|
|
|
|
* For more detail, please refer to the "Hotplug CPU" section
|
|
|
|
|
* of RCU's Requirements documentation.
|
|
|
|
|
*/
|
2021-12-11 05:44:17 +08:00
|
|
|
|
if (WARN_ON_ONCE(!rcu_rdp_cpu_online(rdp))) {
|
2018-05-16 07:23:23 +08:00
|
|
|
|
struct rcu_node *rnp1;
|
|
|
|
|
|
|
|
|
|
pr_info("%s: grp: %d-%d level: %d ->gp_seq %ld ->completedqs %ld\n",
|
|
|
|
|
__func__, rnp->grplo, rnp->grphi, rnp->level,
|
|
|
|
|
(long)rnp->gp_seq, (long)rnp->completedqs);
|
|
|
|
|
for (rnp1 = rnp; rnp1; rnp1 = rnp1->parent)
|
|
|
|
|
pr_info("%s: %d:%d ->qsmask %#lx ->qsmaskinit %#lx ->qsmaskinitnext %#lx ->rcu_gp_init_mask %#lx\n",
|
|
|
|
|
__func__, rnp1->grplo, rnp1->grphi, rnp1->qsmask, rnp1->qsmaskinit, rnp1->qsmaskinitnext, rnp1->rcu_gp_init_mask);
|
|
|
|
|
pr_info("%s %d: %c online: %ld(%d) offline: %ld(%d)\n",
|
2021-12-11 05:44:17 +08:00
|
|
|
|
__func__, rdp->cpu, ".o"[rcu_rdp_cpu_online(rdp)],
|
2018-05-16 07:23:23 +08:00
|
|
|
|
(long)rdp->rcu_onl_gp_seq, rdp->rcu_onl_gp_flags,
|
|
|
|
|
(long)rdp->rcu_ofl_gp_seq, rdp->rcu_ofl_gp_flags);
|
|
|
|
|
return 1; /* Break things loose after complaining. */
|
|
|
|
|
}
|
|
|
|
|
|
2013-04-13 07:19:10 +08:00
|
|
|
|
/*
|
2014-06-21 07:49:01 +08:00
|
|
|
|
* A CPU running for an extended time within the kernel can
|
rcu: Compute jiffies_till_sched_qs from other kernel parameters
The jiffies_till_sched_qs value used to determine how old a grace period
must be before RCU enlists the help of the scheduler to force a quiescent
state on the holdout CPU. Currently, this defaults to HZ/10 regardless of
system size and may be set only at boot time. This can be a problem for
very large systems, because if the values of the jiffies_till_first_fqs
and jiffies_till_next_fqs kernel parameters are left at their defaults,
they are calculated to increase as the number of CPUs actually configured
on the system increases. Thus, on a sufficiently large system, RCU would
enlist the help of the scheduler before the grace-period kthread had a
chance to scan for idle CPUs, which wastes CPU time.
This commit therefore allows jiffies_till_sched_qs to be set, if desired,
but if left as default, computes is as jiffies_till_first_fqs plus twice
jiffies_till_next_fqs, thus allowing three force-quiescent-state scans
for idle CPUs. This scales with the number of CPUs, providing sensible
default values.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-07-26 02:25:23 +08:00
|
|
|
|
* delay RCU grace periods: (1) At age jiffies_to_sched_qs,
|
|
|
|
|
* set .rcu_urgent_qs, (2) At age 2*jiffies_to_sched_qs, set
|
2018-07-11 23:09:28 +08:00
|
|
|
|
* both .rcu_need_heavy_qs and .rcu_urgent_qs. Note that the
|
|
|
|
|
* unsynchronized assignments to the per-CPU rcu_need_heavy_qs
|
|
|
|
|
* variable are safe because the assignments are repeated if this
|
|
|
|
|
* CPU failed to pass through a quiescent state. This code
|
rcu: Compute jiffies_till_sched_qs from other kernel parameters
The jiffies_till_sched_qs value used to determine how old a grace period
must be before RCU enlists the help of the scheduler to force a quiescent
state on the holdout CPU. Currently, this defaults to HZ/10 regardless of
system size and may be set only at boot time. This can be a problem for
very large systems, because if the values of the jiffies_till_first_fqs
and jiffies_till_next_fqs kernel parameters are left at their defaults,
they are calculated to increase as the number of CPUs actually configured
on the system increases. Thus, on a sufficiently large system, RCU would
enlist the help of the scheduler before the grace-period kthread had a
chance to scan for idle CPUs, which wastes CPU time.
This commit therefore allows jiffies_till_sched_qs to be set, if desired,
but if left as default, computes is as jiffies_till_first_fqs plus twice
jiffies_till_next_fqs, thus allowing three force-quiescent-state scans
for idle CPUs. This scales with the number of CPUs, providing sensible
default values.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-07-26 02:25:23 +08:00
|
|
|
|
* also checks .jiffies_resched in case jiffies_to_sched_qs
|
2018-07-11 23:09:28 +08:00
|
|
|
|
* is set way high.
|
2013-09-24 04:57:18 +08:00
|
|
|
|
*/
|
rcu: Compute jiffies_till_sched_qs from other kernel parameters
The jiffies_till_sched_qs value used to determine how old a grace period
must be before RCU enlists the help of the scheduler to force a quiescent
state on the holdout CPU. Currently, this defaults to HZ/10 regardless of
system size and may be set only at boot time. This can be a problem for
very large systems, because if the values of the jiffies_till_first_fqs
and jiffies_till_next_fqs kernel parameters are left at their defaults,
they are calculated to increase as the number of CPUs actually configured
on the system increases. Thus, on a sufficiently large system, RCU would
enlist the help of the scheduler before the grace-period kthread had a
chance to scan for idle CPUs, which wastes CPU time.
This commit therefore allows jiffies_till_sched_qs to be set, if desired,
but if left as default, computes is as jiffies_till_first_fqs plus twice
jiffies_till_next_fqs, thus allowing three force-quiescent-state scans
for idle CPUs. This scales with the number of CPUs, providing sensible
default values.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-07-26 02:25:23 +08:00
|
|
|
|
jtsq = READ_ONCE(jiffies_to_sched_qs);
|
2021-07-23 06:49:05 +08:00
|
|
|
|
if (!READ_ONCE(rdp->rcu_need_heavy_qs) &&
|
2018-07-11 23:09:28 +08:00
|
|
|
|
(time_after(jiffies, rcu_state.gp_start + jtsq * 2) ||
|
2019-10-31 02:56:10 +08:00
|
|
|
|
time_after(jiffies, rcu_state.jiffies_resched) ||
|
|
|
|
|
rcu_state.cbovld)) {
|
2021-07-23 06:49:05 +08:00
|
|
|
|
WRITE_ONCE(rdp->rcu_need_heavy_qs, true);
|
2017-01-28 06:17:50 +08:00
|
|
|
|
/* Store rcu_need_heavy_qs before rcu_urgent_qs. */
|
2021-07-23 07:47:42 +08:00
|
|
|
|
smp_store_release(&rdp->rcu_urgent_qs, true);
|
2018-07-11 23:09:28 +08:00
|
|
|
|
} else if (time_after(jiffies, rcu_state.gp_start + jtsq)) {
|
2021-07-23 07:47:42 +08:00
|
|
|
|
WRITE_ONCE(rdp->rcu_urgent_qs, true);
|
2013-09-24 04:57:18 +08:00
|
|
|
|
}
|
|
|
|
|
|
2016-12-02 03:31:31 +08:00
|
|
|
|
/*
|
2018-11-22 03:35:03 +08:00
|
|
|
|
* NO_HZ_FULL CPUs can run in-kernel without rcu_sched_clock_irq!
|
2018-07-26 02:49:47 +08:00
|
|
|
|
* The above code handles this, but only for straight cond_resched().
|
|
|
|
|
* And some in-kernel loops check need_resched() before calling
|
|
|
|
|
* cond_resched(), which defeats the above code for CPUs that are
|
|
|
|
|
* running in-kernel with scheduling-clock interrupts disabled.
|
|
|
|
|
* So hit them over the head with the resched_cpu() hammer!
|
2016-12-02 03:31:31 +08:00
|
|
|
|
*/
|
2018-07-26 02:49:47 +08:00
|
|
|
|
if (tick_nohz_full_cpu(rdp->cpu) &&
|
2019-10-31 02:56:10 +08:00
|
|
|
|
(time_after(jiffies, READ_ONCE(rdp->last_fqs_resched) + jtsq * 3) ||
|
|
|
|
|
rcu_state.cbovld)) {
|
2021-07-23 07:47:42 +08:00
|
|
|
|
WRITE_ONCE(rdp->rcu_urgent_qs, true);
|
2018-07-26 02:49:47 +08:00
|
|
|
|
WRITE_ONCE(rdp->last_fqs_resched, jiffies);
|
2023-10-31 16:53:08 +08:00
|
|
|
|
ret = -1;
|
2018-07-26 02:49:47 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* If more than halfway to RCU CPU stall-warning time, invoke
|
|
|
|
|
* resched_cpu() more frequently to try to loosen things up a bit.
|
|
|
|
|
* Also check to see if the CPU is getting hammered with interrupts,
|
|
|
|
|
* but only once per grace period, just to keep the IPIs down to
|
|
|
|
|
* a dull roar.
|
|
|
|
|
*/
|
|
|
|
|
if (time_after(jiffies, rcu_state.jiffies_resched)) {
|
|
|
|
|
if (time_after(jiffies,
|
|
|
|
|
READ_ONCE(rdp->last_fqs_resched) + jtsq)) {
|
|
|
|
|
WRITE_ONCE(rdp->last_fqs_resched, jiffies);
|
2023-10-31 16:53:08 +08:00
|
|
|
|
ret = -1;
|
2018-07-26 02:49:47 +08:00
|
|
|
|
}
|
2017-08-18 08:05:59 +08:00
|
|
|
|
if (IS_ENABLED(CONFIG_IRQ_WORK) &&
|
2018-04-29 05:15:40 +08:00
|
|
|
|
!rdp->rcu_iw_pending && rdp->rcu_iw_gp_seq != rnp->gp_seq &&
|
2017-08-18 08:05:59 +08:00
|
|
|
|
(rnp->ffmask & rdp->grpmask)) {
|
|
|
|
|
rdp->rcu_iw_pending = true;
|
2018-04-29 05:15:40 +08:00
|
|
|
|
rdp->rcu_iw_gp_seq = rnp->gp_seq;
|
2017-08-18 08:05:59 +08:00
|
|
|
|
irq_work_queue_on(&rdp->rcu_iw, rdp->cpu);
|
|
|
|
|
}
|
2022-11-19 17:25:06 +08:00
|
|
|
|
|
|
|
|
|
if (rcu_cpu_stall_cputime && rdp->snap_record.gp_seq != rdp->gp_seq) {
|
|
|
|
|
int cpu = rdp->cpu;
|
|
|
|
|
struct rcu_snap_record *rsrp;
|
|
|
|
|
struct kernel_cpustat *kcsp;
|
|
|
|
|
|
|
|
|
|
kcsp = &kcpustat_cpu(cpu);
|
|
|
|
|
|
|
|
|
|
rsrp = &rdp->snap_record;
|
|
|
|
|
rsrp->cputime_irq = kcpustat_field(kcsp, CPUTIME_IRQ, cpu);
|
|
|
|
|
rsrp->cputime_softirq = kcpustat_field(kcsp, CPUTIME_SOFTIRQ, cpu);
|
|
|
|
|
rsrp->cputime_system = kcpustat_field(kcsp, CPUTIME_SYSTEM, cpu);
|
|
|
|
|
rsrp->nr_hardirqs = kstat_cpu_irqs_sum(rdp->cpu);
|
|
|
|
|
rsrp->nr_softirqs = kstat_cpu_softirqs_sum(rdp->cpu);
|
|
|
|
|
rsrp->nr_csw = nr_context_switches_cpu(rdp->cpu);
|
|
|
|
|
rsrp->jiffies = jiffies;
|
|
|
|
|
rsrp->gp_seq = rdp->gp_seq;
|
|
|
|
|
}
|
2017-08-18 08:05:59 +08:00
|
|
|
|
}
|
2015-12-12 05:48:43 +08:00
|
|
|
|
|
2023-10-31 16:53:08 +08:00
|
|
|
|
return ret;
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
|
|
|
|
|
2018-04-13 02:24:09 +08:00
|
|
|
|
/* Trace-event wrapper function for trace_rcu_future_grace_period. */
|
|
|
|
|
static void trace_rcu_this_gp(struct rcu_node *rnp, struct rcu_data *rdp,
|
2018-05-21 12:42:18 +08:00
|
|
|
|
unsigned long gp_seq_req, const char *s)
|
2012-12-31 07:21:01 +08:00
|
|
|
|
{
|
2020-01-04 06:53:31 +08:00
|
|
|
|
trace_rcu_future_grace_period(rcu_state.name, READ_ONCE(rnp->gp_seq),
|
|
|
|
|
gp_seq_req, rnp->level,
|
|
|
|
|
rnp->grplo, rnp->grphi, s);
|
2012-12-31 07:21:01 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
2018-05-21 12:42:18 +08:00
|
|
|
|
* rcu_start_this_gp - Request the start of a particular grace period
|
2018-05-23 14:38:14 +08:00
|
|
|
|
* @rnp_start: The leaf node of the CPU from which to start.
|
2018-05-21 12:42:18 +08:00
|
|
|
|
* @rdp: The rcu_data corresponding to the CPU from which to start.
|
|
|
|
|
* @gp_seq_req: The gp_seq of the grace period to start.
|
|
|
|
|
*
|
2018-04-13 02:24:09 +08:00
|
|
|
|
* Start the specified grace period, as needed to handle newly arrived
|
2012-12-31 07:21:01 +08:00
|
|
|
|
* callbacks. The required future grace periods are recorded in each
|
2018-05-02 01:26:57 +08:00
|
|
|
|
* rcu_node structure's ->gp_seq_needed field. Returns true if there
|
rcu: Make callers awaken grace-period kthread
The rcu_start_gp_advanced() function currently uses irq_work_queue()
to defer wakeups of the RCU grace-period kthread. This deferring
is necessary to avoid RCU-scheduler deadlocks involving the rcu_node
structure's lock, meaning that RCU cannot call any of the scheduler's
wake-up functions while holding one of these locks.
Unfortunately, the second and subsequent calls to irq_work_queue() are
ignored, and the first call will be ignored (aside from queuing the work
item) if the scheduler-clock tick is turned off. This is OK for many
uses, especially those where irq_work_queue() is called from an interrupt
or softirq handler, because in those cases the scheduler-clock-tick state
will be re-evaluated, which will turn the scheduler-clock tick back on.
On the next tick, any deferred work will then be processed.
However, this strategy does not always work for RCU, which can be invoked
at process level from idle CPUs. In this case, the tick might never
be turned back on, indefinitely defering a grace-period start request.
Note that the RCU CPU stall detector cannot see this condition, because
there is no RCU grace period in progress. Therefore, we can (and do!)
see long tens-of-seconds stalls in grace-period handling. In theory,
we could see a full grace-period hang, but rcutorture testing to date
has seen only the tens-of-seconds stalls. Event tracing demonstrates
that irq_work_queue() is being called repeatedly to no effect during
these stalls: The "newreq" event appears repeatedly from a task that is
not one of the grace-period kthreads.
In theory, irq_work_queue() might be fixed to avoid this sort of issue,
but RCU's requirements are unusual and it is quite straightforward to pass
wake-up responsibility up through RCU's call chain, so that the wakeup
happens when the offending locks are released.
This commit therefore makes this change. The rcu_start_gp_advanced(),
rcu_start_future_gp(), rcu_accelerate_cbs(), rcu_advance_cbs(),
__note_gp_changes(), and rcu_start_gp() functions now return a boolean
which indicates when a wake-up is needed. A new rcu_gp_kthread_wake()
does the wakeup when it is necessary and safe to do so: No self-wakes,
no wake-ups if the ->gp_flags field indicates there is no need (as in
someone else did the wake-up before we got around to it), and no wake-ups
before the grace-period kthread has been created.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2014-03-12 04:02:16 +08:00
|
|
|
|
* is reason to awaken the grace-period kthread.
|
2012-12-31 07:21:01 +08:00
|
|
|
|
*
|
2018-04-13 01:45:06 +08:00
|
|
|
|
* The caller must hold the specified rcu_node structure's ->lock, which
|
|
|
|
|
* is why the caller is responsible for waking the grace-period kthread.
|
2018-05-21 12:42:18 +08:00
|
|
|
|
*
|
|
|
|
|
* Returns true if the GP thread needs to be awakened else false.
|
2012-12-31 07:21:01 +08:00
|
|
|
|
*/
|
2018-05-23 14:38:14 +08:00
|
|
|
|
static bool rcu_start_this_gp(struct rcu_node *rnp_start, struct rcu_data *rdp,
|
2018-05-21 12:42:18 +08:00
|
|
|
|
unsigned long gp_seq_req)
|
2012-12-31 07:21:01 +08:00
|
|
|
|
{
|
rcu: Make callers awaken grace-period kthread
The rcu_start_gp_advanced() function currently uses irq_work_queue()
to defer wakeups of the RCU grace-period kthread. This deferring
is necessary to avoid RCU-scheduler deadlocks involving the rcu_node
structure's lock, meaning that RCU cannot call any of the scheduler's
wake-up functions while holding one of these locks.
Unfortunately, the second and subsequent calls to irq_work_queue() are
ignored, and the first call will be ignored (aside from queuing the work
item) if the scheduler-clock tick is turned off. This is OK for many
uses, especially those where irq_work_queue() is called from an interrupt
or softirq handler, because in those cases the scheduler-clock-tick state
will be re-evaluated, which will turn the scheduler-clock tick back on.
On the next tick, any deferred work will then be processed.
However, this strategy does not always work for RCU, which can be invoked
at process level from idle CPUs. In this case, the tick might never
be turned back on, indefinitely defering a grace-period start request.
Note that the RCU CPU stall detector cannot see this condition, because
there is no RCU grace period in progress. Therefore, we can (and do!)
see long tens-of-seconds stalls in grace-period handling. In theory,
we could see a full grace-period hang, but rcutorture testing to date
has seen only the tens-of-seconds stalls. Event tracing demonstrates
that irq_work_queue() is being called repeatedly to no effect during
these stalls: The "newreq" event appears repeatedly from a task that is
not one of the grace-period kthreads.
In theory, irq_work_queue() might be fixed to avoid this sort of issue,
but RCU's requirements are unusual and it is quite straightforward to pass
wake-up responsibility up through RCU's call chain, so that the wakeup
happens when the offending locks are released.
This commit therefore makes this change. The rcu_start_gp_advanced(),
rcu_start_future_gp(), rcu_accelerate_cbs(), rcu_advance_cbs(),
__note_gp_changes(), and rcu_start_gp() functions now return a boolean
which indicates when a wake-up is needed. A new rcu_gp_kthread_wake()
does the wakeup when it is necessary and safe to do so: No self-wakes,
no wake-ups if the ->gp_flags field indicates there is no need (as in
someone else did the wake-up before we got around to it), and no wake-ups
before the grace-period kthread has been created.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2014-03-12 04:02:16 +08:00
|
|
|
|
bool ret = false;
|
2018-05-23 14:38:14 +08:00
|
|
|
|
struct rcu_node *rnp;
|
2012-12-31 07:21:01 +08:00
|
|
|
|
|
|
|
|
|
/*
|
2018-04-13 02:50:41 +08:00
|
|
|
|
* Use funnel locking to either acquire the root rcu_node
|
|
|
|
|
* structure's lock or bail out if the need for this grace period
|
2018-05-23 14:38:14 +08:00
|
|
|
|
* has already been recorded -- or if that grace period has in
|
|
|
|
|
* fact already started. If there is already a grace period in
|
|
|
|
|
* progress in a non-leaf node, no recording is needed because the
|
|
|
|
|
* end of the grace period will scan the leaf rcu_node structures.
|
|
|
|
|
* Note that rnp_start->lock must not be released.
|
2012-12-31 07:21:01 +08:00
|
|
|
|
*/
|
2018-05-23 14:38:14 +08:00
|
|
|
|
raw_lockdep_assert_held_rcu_node(rnp_start);
|
|
|
|
|
trace_rcu_this_gp(rnp_start, rdp, gp_seq_req, TPS("Startleaf"));
|
|
|
|
|
for (rnp = rnp_start; 1; rnp = rnp->parent) {
|
|
|
|
|
if (rnp != rnp_start)
|
|
|
|
|
raw_spin_lock_rcu_node(rnp);
|
|
|
|
|
if (ULONG_CMP_GE(rnp->gp_seq_needed, gp_seq_req) ||
|
|
|
|
|
rcu_seq_started(&rnp->gp_seq, gp_seq_req) ||
|
|
|
|
|
(rnp != rnp_start &&
|
|
|
|
|
rcu_seq_state(rcu_seq_current(&rnp->gp_seq)))) {
|
|
|
|
|
trace_rcu_this_gp(rnp, rdp, gp_seq_req,
|
2018-05-21 12:42:18 +08:00
|
|
|
|
TPS("Prestarted"));
|
2018-04-13 02:50:41 +08:00
|
|
|
|
goto unlock_out;
|
|
|
|
|
}
|
2020-01-05 03:33:17 +08:00
|
|
|
|
WRITE_ONCE(rnp->gp_seq_needed, gp_seq_req);
|
2018-05-23 14:38:15 +08:00
|
|
|
|
if (rcu_seq_state(rcu_seq_current(&rnp->gp_seq))) {
|
2018-05-12 22:42:20 +08:00
|
|
|
|
/*
|
2018-05-23 14:38:15 +08:00
|
|
|
|
* We just marked the leaf or internal node, and a
|
|
|
|
|
* grace period is in progress, which means that
|
|
|
|
|
* rcu_gp_cleanup() will see the marking. Bail to
|
|
|
|
|
* reduce contention.
|
2018-05-12 22:42:20 +08:00
|
|
|
|
*/
|
2018-05-23 14:38:14 +08:00
|
|
|
|
trace_rcu_this_gp(rnp_start, rdp, gp_seq_req,
|
2018-05-21 12:42:18 +08:00
|
|
|
|
TPS("Startedleaf"));
|
2018-05-12 22:42:20 +08:00
|
|
|
|
goto unlock_out;
|
|
|
|
|
}
|
2018-05-23 14:38:14 +08:00
|
|
|
|
if (rnp != rnp_start && rnp->parent != NULL)
|
|
|
|
|
raw_spin_unlock_rcu_node(rnp);
|
|
|
|
|
if (!rnp->parent)
|
2018-04-13 02:50:41 +08:00
|
|
|
|
break; /* At root, and perhaps also leaf. */
|
2012-12-31 07:21:01 +08:00
|
|
|
|
}
|
|
|
|
|
|
2018-04-13 02:50:41 +08:00
|
|
|
|
/* If GP already in progress, just leave, otherwise start one. */
|
2018-07-04 08:22:34 +08:00
|
|
|
|
if (rcu_gp_in_progress()) {
|
2018-05-23 14:38:14 +08:00
|
|
|
|
trace_rcu_this_gp(rnp, rdp, gp_seq_req, TPS("Startedleafroot"));
|
2012-12-31 07:21:01 +08:00
|
|
|
|
goto unlock_out;
|
|
|
|
|
}
|
2018-05-23 14:38:14 +08:00
|
|
|
|
trace_rcu_this_gp(rnp, rdp, gp_seq_req, TPS("Startedroot"));
|
2018-07-06 06:47:01 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_flags, rcu_state.gp_flags | RCU_GP_FLAG_INIT);
|
2020-01-04 07:17:12 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_req_activity, jiffies);
|
2020-01-22 04:30:22 +08:00
|
|
|
|
if (!READ_ONCE(rcu_state.gp_kthread)) {
|
2018-05-23 14:38:14 +08:00
|
|
|
|
trace_rcu_this_gp(rnp, rdp, gp_seq_req, TPS("NoGPkthread"));
|
2018-04-13 02:50:41 +08:00
|
|
|
|
goto unlock_out;
|
2012-12-31 07:21:01 +08:00
|
|
|
|
}
|
2020-03-22 10:52:20 +08:00
|
|
|
|
trace_rcu_grace_period(rcu_state.name, data_race(rcu_state.gp_seq), TPS("newreq"));
|
2018-04-13 02:50:41 +08:00
|
|
|
|
ret = true; /* Caller must wake GP kthread. */
|
2012-12-31 07:21:01 +08:00
|
|
|
|
unlock_out:
|
2018-05-02 02:07:23 +08:00
|
|
|
|
/* Push furthest requested GP to leaf node and rcu_data structure. */
|
2018-05-23 14:38:14 +08:00
|
|
|
|
if (ULONG_CMP_LT(gp_seq_req, rnp->gp_seq_needed)) {
|
2020-01-05 03:33:17 +08:00
|
|
|
|
WRITE_ONCE(rnp_start->gp_seq_needed, rnp->gp_seq_needed);
|
|
|
|
|
WRITE_ONCE(rdp->gp_seq_needed, rnp->gp_seq_needed);
|
2018-05-02 02:07:23 +08:00
|
|
|
|
}
|
2018-05-23 14:38:14 +08:00
|
|
|
|
if (rnp != rnp_start)
|
|
|
|
|
raw_spin_unlock_rcu_node(rnp);
|
rcu: Make callers awaken grace-period kthread
The rcu_start_gp_advanced() function currently uses irq_work_queue()
to defer wakeups of the RCU grace-period kthread. This deferring
is necessary to avoid RCU-scheduler deadlocks involving the rcu_node
structure's lock, meaning that RCU cannot call any of the scheduler's
wake-up functions while holding one of these locks.
Unfortunately, the second and subsequent calls to irq_work_queue() are
ignored, and the first call will be ignored (aside from queuing the work
item) if the scheduler-clock tick is turned off. This is OK for many
uses, especially those where irq_work_queue() is called from an interrupt
or softirq handler, because in those cases the scheduler-clock-tick state
will be re-evaluated, which will turn the scheduler-clock tick back on.
On the next tick, any deferred work will then be processed.
However, this strategy does not always work for RCU, which can be invoked
at process level from idle CPUs. In this case, the tick might never
be turned back on, indefinitely defering a grace-period start request.
Note that the RCU CPU stall detector cannot see this condition, because
there is no RCU grace period in progress. Therefore, we can (and do!)
see long tens-of-seconds stalls in grace-period handling. In theory,
we could see a full grace-period hang, but rcutorture testing to date
has seen only the tens-of-seconds stalls. Event tracing demonstrates
that irq_work_queue() is being called repeatedly to no effect during
these stalls: The "newreq" event appears repeatedly from a task that is
not one of the grace-period kthreads.
In theory, irq_work_queue() might be fixed to avoid this sort of issue,
but RCU's requirements are unusual and it is quite straightforward to pass
wake-up responsibility up through RCU's call chain, so that the wakeup
happens when the offending locks are released.
This commit therefore makes this change. The rcu_start_gp_advanced(),
rcu_start_future_gp(), rcu_accelerate_cbs(), rcu_advance_cbs(),
__note_gp_changes(), and rcu_start_gp() functions now return a boolean
which indicates when a wake-up is needed. A new rcu_gp_kthread_wake()
does the wakeup when it is necessary and safe to do so: No self-wakes,
no wake-ups if the ->gp_flags field indicates there is no need (as in
someone else did the wake-up before we got around to it), and no wake-ups
before the grace-period kthread has been created.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2014-03-12 04:02:16 +08:00
|
|
|
|
return ret;
|
2012-12-31 07:21:01 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Clean up any old requests for the just-ended grace period. Also return
|
2017-02-09 06:58:41 +08:00
|
|
|
|
* whether any additional grace periods have been requested.
|
2012-12-31 07:21:01 +08:00
|
|
|
|
*/
|
2018-07-04 08:22:34 +08:00
|
|
|
|
static bool rcu_future_gp_cleanup(struct rcu_node *rnp)
|
2012-12-31 07:21:01 +08:00
|
|
|
|
{
|
2018-04-12 22:20:30 +08:00
|
|
|
|
bool needmore;
|
2018-07-04 06:37:16 +08:00
|
|
|
|
struct rcu_data *rdp = this_cpu_ptr(&rcu_data);
|
2012-12-31 07:21:01 +08:00
|
|
|
|
|
2018-05-02 01:26:57 +08:00
|
|
|
|
needmore = ULONG_CMP_LT(rnp->gp_seq, rnp->gp_seq_needed);
|
|
|
|
|
if (!needmore)
|
|
|
|
|
rnp->gp_seq_needed = rnp->gp_seq; /* Avoid counter wrap. */
|
2018-05-21 12:42:18 +08:00
|
|
|
|
trace_rcu_this_gp(rnp, rdp, rnp->gp_seq,
|
2018-04-13 02:24:09 +08:00
|
|
|
|
needmore ? TPS("CleanupMore") : TPS("Cleanup"));
|
2012-12-31 07:21:01 +08:00
|
|
|
|
return needmore;
|
|
|
|
|
}
|
|
|
|
|
|
2023-12-19 07:19:15 +08:00
|
|
|
|
static void swake_up_one_online_ipi(void *arg)
|
|
|
|
|
{
|
|
|
|
|
struct swait_queue_head *wqh = arg;
|
|
|
|
|
|
|
|
|
|
swake_up_one(wqh);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void swake_up_one_online(struct swait_queue_head *wqh)
|
|
|
|
|
{
|
|
|
|
|
int cpu = get_cpu();
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* If called from rcutree_report_cpu_starting(), wake up
|
|
|
|
|
* is dangerous that late in the CPU-down hotplug process. The
|
|
|
|
|
* scheduler might queue an ignored hrtimer. Defer the wake up
|
|
|
|
|
* to an online CPU instead.
|
|
|
|
|
*/
|
|
|
|
|
if (unlikely(cpu_is_offline(cpu))) {
|
|
|
|
|
int target;
|
|
|
|
|
|
|
|
|
|
target = cpumask_any_and(housekeeping_cpumask(HK_TYPE_RCU),
|
|
|
|
|
cpu_online_mask);
|
|
|
|
|
|
|
|
|
|
smp_call_function_single(target, swake_up_one_online_ipi,
|
|
|
|
|
wqh, 0);
|
|
|
|
|
put_cpu();
|
|
|
|
|
} else {
|
|
|
|
|
put_cpu();
|
|
|
|
|
swake_up_one(wqh);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
rcu: Make callers awaken grace-period kthread
The rcu_start_gp_advanced() function currently uses irq_work_queue()
to defer wakeups of the RCU grace-period kthread. This deferring
is necessary to avoid RCU-scheduler deadlocks involving the rcu_node
structure's lock, meaning that RCU cannot call any of the scheduler's
wake-up functions while holding one of these locks.
Unfortunately, the second and subsequent calls to irq_work_queue() are
ignored, and the first call will be ignored (aside from queuing the work
item) if the scheduler-clock tick is turned off. This is OK for many
uses, especially those where irq_work_queue() is called from an interrupt
or softirq handler, because in those cases the scheduler-clock-tick state
will be re-evaluated, which will turn the scheduler-clock tick back on.
On the next tick, any deferred work will then be processed.
However, this strategy does not always work for RCU, which can be invoked
at process level from idle CPUs. In this case, the tick might never
be turned back on, indefinitely defering a grace-period start request.
Note that the RCU CPU stall detector cannot see this condition, because
there is no RCU grace period in progress. Therefore, we can (and do!)
see long tens-of-seconds stalls in grace-period handling. In theory,
we could see a full grace-period hang, but rcutorture testing to date
has seen only the tens-of-seconds stalls. Event tracing demonstrates
that irq_work_queue() is being called repeatedly to no effect during
these stalls: The "newreq" event appears repeatedly from a task that is
not one of the grace-period kthreads.
In theory, irq_work_queue() might be fixed to avoid this sort of issue,
but RCU's requirements are unusual and it is quite straightforward to pass
wake-up responsibility up through RCU's call chain, so that the wakeup
happens when the offending locks are released.
This commit therefore makes this change. The rcu_start_gp_advanced(),
rcu_start_future_gp(), rcu_accelerate_cbs(), rcu_advance_cbs(),
__note_gp_changes(), and rcu_start_gp() functions now return a boolean
which indicates when a wake-up is needed. A new rcu_gp_kthread_wake()
does the wakeup when it is necessary and safe to do so: No self-wakes,
no wake-ups if the ->gp_flags field indicates there is no need (as in
someone else did the wake-up before we got around to it), and no wake-ups
before the grace-period kthread has been created.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2014-03-12 04:02:16 +08:00
|
|
|
|
/*
|
2020-01-22 04:30:22 +08:00
|
|
|
|
* Awaken the grace-period kthread. Don't do a self-awaken (unless in an
|
|
|
|
|
* interrupt or softirq handler, in which case we just might immediately
|
|
|
|
|
* sleep upon return, resulting in a grace-period hang), and don't bother
|
|
|
|
|
* awakening when there is nothing for the grace-period kthread to do
|
|
|
|
|
* (as in several CPUs raced to awaken, we lost), and finally don't try
|
|
|
|
|
* to awaken a kthread that has not yet been created. If all those checks
|
|
|
|
|
* are passed, track some debug information and awaken.
|
rcu: Do RCU GP kthread self-wakeup from softirq and interrupt
The rcu_gp_kthread_wake() function is invoked when it might be necessary
to wake the RCU grace-period kthread. Because self-wakeups are normally
a useless waste of CPU cycles, if rcu_gp_kthread_wake() is invoked from
this kthread, it naturally refuses to do the wakeup.
Unfortunately, natural though it might be, this heuristic fails when
rcu_gp_kthread_wake() is invoked from an interrupt or softirq handler
that interrupted the grace-period kthread just after the final check of
the wait-event condition but just before the schedule() call. In this
case, a wakeup is required, even though the call to rcu_gp_kthread_wake()
is within the RCU grace-period kthread's context. Failing to provide
this wakeup can result in grace periods failing to start, which in turn
results in out-of-memory conditions.
This race window is quite narrow, but it actually did happen during real
testing. It would of course need to be fixed even if it was strictly
theoretical in nature.
This patch does not Cc stable because it does not apply cleanly to
earlier kernel versions.
Fixes: 48a7639ce80c ("rcu: Make callers awaken grace-period kthread")
Reported-by: "He, Bo" <bo.he@intel.com>
Co-developed-by: "Zhang, Jun" <jun.zhang@intel.com>
Co-developed-by: "He, Bo" <bo.he@intel.com>
Co-developed-by: "xiao, jin" <jin.xiao@intel.com>
Co-developed-by: Bai, Jie A <jie.a.bai@intel.com>
Signed-off: "Zhang, Jun" <jun.zhang@intel.com>
Signed-off: "He, Bo" <bo.he@intel.com>
Signed-off: "xiao, jin" <jin.xiao@intel.com>
Signed-off: Bai, Jie A <jie.a.bai@intel.com>
Signed-off-by: "Zhang, Jun" <jun.zhang@intel.com>
[ paulmck: Switch from !in_softirq() to "!in_interrupt() &&
!in_serving_softirq() to avoid redundant wakeups and to also handle the
interrupt-handler scenario as well as the softirq-handler scenario that
actually occurred in testing. ]
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Link: https://lkml.kernel.org/r/CD6925E8781EFD4D8E11882D20FC406D52A11F61@SHSMSX104.ccr.corp.intel.com
2018-12-18 22:55:01 +08:00
|
|
|
|
*
|
|
|
|
|
* So why do the self-wakeup when in an interrupt or softirq handler
|
|
|
|
|
* in the grace-period kthread's context? Because the kthread might have
|
|
|
|
|
* been interrupted just as it was going to sleep, and just after the final
|
|
|
|
|
* pre-sleep check of the awaken condition. In this case, a wakeup really
|
|
|
|
|
* is required, and is therefore supplied.
|
rcu: Make callers awaken grace-period kthread
The rcu_start_gp_advanced() function currently uses irq_work_queue()
to defer wakeups of the RCU grace-period kthread. This deferring
is necessary to avoid RCU-scheduler deadlocks involving the rcu_node
structure's lock, meaning that RCU cannot call any of the scheduler's
wake-up functions while holding one of these locks.
Unfortunately, the second and subsequent calls to irq_work_queue() are
ignored, and the first call will be ignored (aside from queuing the work
item) if the scheduler-clock tick is turned off. This is OK for many
uses, especially those where irq_work_queue() is called from an interrupt
or softirq handler, because in those cases the scheduler-clock-tick state
will be re-evaluated, which will turn the scheduler-clock tick back on.
On the next tick, any deferred work will then be processed.
However, this strategy does not always work for RCU, which can be invoked
at process level from idle CPUs. In this case, the tick might never
be turned back on, indefinitely defering a grace-period start request.
Note that the RCU CPU stall detector cannot see this condition, because
there is no RCU grace period in progress. Therefore, we can (and do!)
see long tens-of-seconds stalls in grace-period handling. In theory,
we could see a full grace-period hang, but rcutorture testing to date
has seen only the tens-of-seconds stalls. Event tracing demonstrates
that irq_work_queue() is being called repeatedly to no effect during
these stalls: The "newreq" event appears repeatedly from a task that is
not one of the grace-period kthreads.
In theory, irq_work_queue() might be fixed to avoid this sort of issue,
but RCU's requirements are unusual and it is quite straightforward to pass
wake-up responsibility up through RCU's call chain, so that the wakeup
happens when the offending locks are released.
This commit therefore makes this change. The rcu_start_gp_advanced(),
rcu_start_future_gp(), rcu_accelerate_cbs(), rcu_advance_cbs(),
__note_gp_changes(), and rcu_start_gp() functions now return a boolean
which indicates when a wake-up is needed. A new rcu_gp_kthread_wake()
does the wakeup when it is necessary and safe to do so: No self-wakes,
no wake-ups if the ->gp_flags field indicates there is no need (as in
someone else did the wake-up before we got around to it), and no wake-ups
before the grace-period kthread has been created.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2014-03-12 04:02:16 +08:00
|
|
|
|
*/
|
2018-07-04 08:22:34 +08:00
|
|
|
|
static void rcu_gp_kthread_wake(void)
|
rcu: Make callers awaken grace-period kthread
The rcu_start_gp_advanced() function currently uses irq_work_queue()
to defer wakeups of the RCU grace-period kthread. This deferring
is necessary to avoid RCU-scheduler deadlocks involving the rcu_node
structure's lock, meaning that RCU cannot call any of the scheduler's
wake-up functions while holding one of these locks.
Unfortunately, the second and subsequent calls to irq_work_queue() are
ignored, and the first call will be ignored (aside from queuing the work
item) if the scheduler-clock tick is turned off. This is OK for many
uses, especially those where irq_work_queue() is called from an interrupt
or softirq handler, because in those cases the scheduler-clock-tick state
will be re-evaluated, which will turn the scheduler-clock tick back on.
On the next tick, any deferred work will then be processed.
However, this strategy does not always work for RCU, which can be invoked
at process level from idle CPUs. In this case, the tick might never
be turned back on, indefinitely defering a grace-period start request.
Note that the RCU CPU stall detector cannot see this condition, because
there is no RCU grace period in progress. Therefore, we can (and do!)
see long tens-of-seconds stalls in grace-period handling. In theory,
we could see a full grace-period hang, but rcutorture testing to date
has seen only the tens-of-seconds stalls. Event tracing demonstrates
that irq_work_queue() is being called repeatedly to no effect during
these stalls: The "newreq" event appears repeatedly from a task that is
not one of the grace-period kthreads.
In theory, irq_work_queue() might be fixed to avoid this sort of issue,
but RCU's requirements are unusual and it is quite straightforward to pass
wake-up responsibility up through RCU's call chain, so that the wakeup
happens when the offending locks are released.
This commit therefore makes this change. The rcu_start_gp_advanced(),
rcu_start_future_gp(), rcu_accelerate_cbs(), rcu_advance_cbs(),
__note_gp_changes(), and rcu_start_gp() functions now return a boolean
which indicates when a wake-up is needed. A new rcu_gp_kthread_wake()
does the wakeup when it is necessary and safe to do so: No self-wakes,
no wake-ups if the ->gp_flags field indicates there is no need (as in
someone else did the wake-up before we got around to it), and no wake-ups
before the grace-period kthread has been created.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2014-03-12 04:02:16 +08:00
|
|
|
|
{
|
2020-01-22 04:30:22 +08:00
|
|
|
|
struct task_struct *t = READ_ONCE(rcu_state.gp_kthread);
|
|
|
|
|
|
2021-09-28 08:21:28 +08:00
|
|
|
|
if ((current == t && !in_hardirq() && !in_serving_softirq()) ||
|
2020-01-22 04:30:22 +08:00
|
|
|
|
!READ_ONCE(rcu_state.gp_flags) || !t)
|
rcu: Make callers awaken grace-period kthread
The rcu_start_gp_advanced() function currently uses irq_work_queue()
to defer wakeups of the RCU grace-period kthread. This deferring
is necessary to avoid RCU-scheduler deadlocks involving the rcu_node
structure's lock, meaning that RCU cannot call any of the scheduler's
wake-up functions while holding one of these locks.
Unfortunately, the second and subsequent calls to irq_work_queue() are
ignored, and the first call will be ignored (aside from queuing the work
item) if the scheduler-clock tick is turned off. This is OK for many
uses, especially those where irq_work_queue() is called from an interrupt
or softirq handler, because in those cases the scheduler-clock-tick state
will be re-evaluated, which will turn the scheduler-clock tick back on.
On the next tick, any deferred work will then be processed.
However, this strategy does not always work for RCU, which can be invoked
at process level from idle CPUs. In this case, the tick might never
be turned back on, indefinitely defering a grace-period start request.
Note that the RCU CPU stall detector cannot see this condition, because
there is no RCU grace period in progress. Therefore, we can (and do!)
see long tens-of-seconds stalls in grace-period handling. In theory,
we could see a full grace-period hang, but rcutorture testing to date
has seen only the tens-of-seconds stalls. Event tracing demonstrates
that irq_work_queue() is being called repeatedly to no effect during
these stalls: The "newreq" event appears repeatedly from a task that is
not one of the grace-period kthreads.
In theory, irq_work_queue() might be fixed to avoid this sort of issue,
but RCU's requirements are unusual and it is quite straightforward to pass
wake-up responsibility up through RCU's call chain, so that the wakeup
happens when the offending locks are released.
This commit therefore makes this change. The rcu_start_gp_advanced(),
rcu_start_future_gp(), rcu_accelerate_cbs(), rcu_advance_cbs(),
__note_gp_changes(), and rcu_start_gp() functions now return a boolean
which indicates when a wake-up is needed. A new rcu_gp_kthread_wake()
does the wakeup when it is necessary and safe to do so: No self-wakes,
no wake-ups if the ->gp_flags field indicates there is no need (as in
someone else did the wake-up before we got around to it), and no wake-ups
before the grace-period kthread has been created.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2014-03-12 04:02:16 +08:00
|
|
|
|
return;
|
2018-12-11 08:09:49 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_wake_time, jiffies);
|
|
|
|
|
WRITE_ONCE(rcu_state.gp_wake_seq, READ_ONCE(rcu_state.gp_seq));
|
2023-12-19 07:19:15 +08:00
|
|
|
|
swake_up_one_online(&rcu_state.gp_wq);
|
rcu: Make callers awaken grace-period kthread
The rcu_start_gp_advanced() function currently uses irq_work_queue()
to defer wakeups of the RCU grace-period kthread. This deferring
is necessary to avoid RCU-scheduler deadlocks involving the rcu_node
structure's lock, meaning that RCU cannot call any of the scheduler's
wake-up functions while holding one of these locks.
Unfortunately, the second and subsequent calls to irq_work_queue() are
ignored, and the first call will be ignored (aside from queuing the work
item) if the scheduler-clock tick is turned off. This is OK for many
uses, especially those where irq_work_queue() is called from an interrupt
or softirq handler, because in those cases the scheduler-clock-tick state
will be re-evaluated, which will turn the scheduler-clock tick back on.
On the next tick, any deferred work will then be processed.
However, this strategy does not always work for RCU, which can be invoked
at process level from idle CPUs. In this case, the tick might never
be turned back on, indefinitely defering a grace-period start request.
Note that the RCU CPU stall detector cannot see this condition, because
there is no RCU grace period in progress. Therefore, we can (and do!)
see long tens-of-seconds stalls in grace-period handling. In theory,
we could see a full grace-period hang, but rcutorture testing to date
has seen only the tens-of-seconds stalls. Event tracing demonstrates
that irq_work_queue() is being called repeatedly to no effect during
these stalls: The "newreq" event appears repeatedly from a task that is
not one of the grace-period kthreads.
In theory, irq_work_queue() might be fixed to avoid this sort of issue,
but RCU's requirements are unusual and it is quite straightforward to pass
wake-up responsibility up through RCU's call chain, so that the wakeup
happens when the offending locks are released.
This commit therefore makes this change. The rcu_start_gp_advanced(),
rcu_start_future_gp(), rcu_accelerate_cbs(), rcu_advance_cbs(),
__note_gp_changes(), and rcu_start_gp() functions now return a boolean
which indicates when a wake-up is needed. A new rcu_gp_kthread_wake()
does the wakeup when it is necessary and safe to do so: No self-wakes,
no wake-ups if the ->gp_flags field indicates there is no need (as in
someone else did the wake-up before we got around to it), and no wake-ups
before the grace-period kthread has been created.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2014-03-12 04:02:16 +08:00
|
|
|
|
}
|
|
|
|
|
|
2012-12-04 05:52:00 +08:00
|
|
|
|
/*
|
2018-05-01 01:57:36 +08:00
|
|
|
|
* If there is room, assign a ->gp_seq number to any callbacks on this
|
|
|
|
|
* CPU that have not already been assigned. Also accelerate any callbacks
|
|
|
|
|
* that were previously assigned a ->gp_seq number that has since proven
|
|
|
|
|
* to be too conservative, which can happen if callbacks get assigned a
|
|
|
|
|
* ->gp_seq number while RCU is idle, but with reference to a non-root
|
|
|
|
|
* rcu_node structure. This function is idempotent, so it does not hurt
|
|
|
|
|
* to call it repeatedly. Returns an flag saying that we should awaken
|
|
|
|
|
* the RCU grace-period kthread.
|
2012-12-04 05:52:00 +08:00
|
|
|
|
*
|
|
|
|
|
* The caller must hold rnp->lock with interrupts disabled.
|
|
|
|
|
*/
|
2018-07-04 08:22:34 +08:00
|
|
|
|
static bool rcu_accelerate_cbs(struct rcu_node *rnp, struct rcu_data *rdp)
|
2012-12-04 05:52:00 +08:00
|
|
|
|
{
|
2018-05-21 12:42:18 +08:00
|
|
|
|
unsigned long gp_seq_req;
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
bool ret = false;
|
2012-12-04 05:52:00 +08:00
|
|
|
|
|
rcu/nocb: Add bypass callback queueing
Use of the rcu_data structure's segmented ->cblist for no-CBs CPUs
takes advantage of unrelated grace periods, thus reducing the memory
footprint in the face of floods of call_rcu() invocations. However,
the ->cblist field is a more-complex rcu_segcblist structure which must
be protected via locking. Even though there are only three entities
which can acquire this lock (the CPU invoking call_rcu(), the no-CBs
grace-period kthread, and the no-CBs callbacks kthread), the contention
on this lock is excessive under heavy stress.
This commit therefore greatly reduces contention by provisioning
an rcu_cblist structure field named ->nocb_bypass within the
rcu_data structure. Each no-CBs CPU is permitted only a limited
number of enqueues onto the ->cblist per jiffy, controlled by a new
nocb_nobypass_lim_per_jiffy kernel boot parameter that defaults to
about 16 enqueues per millisecond (16 * 1000 / HZ). When that limit is
exceeded, the CPU instead enqueues onto the new ->nocb_bypass.
The ->nocb_bypass is flushed into the ->cblist every jiffy or when
the number of callbacks on ->nocb_bypass exceeds qhimark, whichever
happens first. During call_rcu() floods, this flushing is carried out
by the CPU during the course of its call_rcu() invocations. However,
a CPU could simply stop invoking call_rcu() at any time. The no-CBs
grace-period kthread therefore carries out less-aggressive flushing
(every few jiffies or when the number of callbacks on ->nocb_bypass
exceeds (2 * qhimark), whichever comes first). This means that the
no-CBs grace-period kthread cannot be permitted to do unbounded waits
while there are callbacks on ->nocb_bypass. A ->nocb_bypass_timer is
used to provide the needed wakeups.
[ paulmck: Apply Coverity feedback reported by Colin Ian King. ]
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
2019-07-03 07:03:33 +08:00
|
|
|
|
rcu_lockdep_assert_cblist_protected(rdp);
|
2018-01-17 22:24:30 +08:00
|
|
|
|
raw_lockdep_assert_held_rcu_node(rnp);
|
2017-04-29 03:32:15 +08:00
|
|
|
|
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
/* If no pending (not yet ready to invoke) callbacks, nothing to do. */
|
|
|
|
|
if (!rcu_segcblist_pend_cbs(&rdp->cblist))
|
rcu: Make callers awaken grace-period kthread
The rcu_start_gp_advanced() function currently uses irq_work_queue()
to defer wakeups of the RCU grace-period kthread. This deferring
is necessary to avoid RCU-scheduler deadlocks involving the rcu_node
structure's lock, meaning that RCU cannot call any of the scheduler's
wake-up functions while holding one of these locks.
Unfortunately, the second and subsequent calls to irq_work_queue() are
ignored, and the first call will be ignored (aside from queuing the work
item) if the scheduler-clock tick is turned off. This is OK for many
uses, especially those where irq_work_queue() is called from an interrupt
or softirq handler, because in those cases the scheduler-clock-tick state
will be re-evaluated, which will turn the scheduler-clock tick back on.
On the next tick, any deferred work will then be processed.
However, this strategy does not always work for RCU, which can be invoked
at process level from idle CPUs. In this case, the tick might never
be turned back on, indefinitely defering a grace-period start request.
Note that the RCU CPU stall detector cannot see this condition, because
there is no RCU grace period in progress. Therefore, we can (and do!)
see long tens-of-seconds stalls in grace-period handling. In theory,
we could see a full grace-period hang, but rcutorture testing to date
has seen only the tens-of-seconds stalls. Event tracing demonstrates
that irq_work_queue() is being called repeatedly to no effect during
these stalls: The "newreq" event appears repeatedly from a task that is
not one of the grace-period kthreads.
In theory, irq_work_queue() might be fixed to avoid this sort of issue,
but RCU's requirements are unusual and it is quite straightforward to pass
wake-up responsibility up through RCU's call chain, so that the wakeup
happens when the offending locks are released.
This commit therefore makes this change. The rcu_start_gp_advanced(),
rcu_start_future_gp(), rcu_accelerate_cbs(), rcu_advance_cbs(),
__note_gp_changes(), and rcu_start_gp() functions now return a boolean
which indicates when a wake-up is needed. A new rcu_gp_kthread_wake()
does the wakeup when it is necessary and safe to do so: No self-wakes,
no wake-ups if the ->gp_flags field indicates there is no need (as in
someone else did the wake-up before we got around to it), and no wake-ups
before the grace-period kthread has been created.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2014-03-12 04:02:16 +08:00
|
|
|
|
return false;
|
2012-12-04 05:52:00 +08:00
|
|
|
|
|
2020-11-15 03:31:32 +08:00
|
|
|
|
trace_rcu_segcb_stats(&rdp->cblist, TPS("SegCbPreAcc"));
|
|
|
|
|
|
2012-12-04 05:52:00 +08:00
|
|
|
|
/*
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
* Callbacks are often registered with incomplete grace-period
|
|
|
|
|
* information. Something about the fact that getting exact
|
|
|
|
|
* information requires acquiring a global lock... RCU therefore
|
|
|
|
|
* makes a conservative estimate of the grace period number at which
|
|
|
|
|
* a given callback will become ready to invoke. The following
|
|
|
|
|
* code checks this estimate and improves it when possible, thus
|
|
|
|
|
* accelerating callback invocation to an earlier grace-period
|
|
|
|
|
* number.
|
2012-12-04 05:52:00 +08:00
|
|
|
|
*/
|
2018-07-06 06:47:01 +08:00
|
|
|
|
gp_seq_req = rcu_seq_snap(&rcu_state.gp_seq);
|
2018-05-21 12:42:18 +08:00
|
|
|
|
if (rcu_segcblist_accelerate(&rdp->cblist, gp_seq_req))
|
|
|
|
|
ret = rcu_start_this_gp(rnp, rdp, gp_seq_req);
|
2012-11-28 08:55:44 +08:00
|
|
|
|
|
|
|
|
|
/* Trace depending on how much we were able to accelerate. */
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
if (rcu_segcblist_restempty(&rdp->cblist, RCU_WAIT_TAIL))
|
2020-06-19 09:36:40 +08:00
|
|
|
|
trace_rcu_grace_period(rcu_state.name, gp_seq_req, TPS("AccWaitCB"));
|
2012-11-28 08:55:44 +08:00
|
|
|
|
else
|
2020-06-19 09:36:40 +08:00
|
|
|
|
trace_rcu_grace_period(rcu_state.name, gp_seq_req, TPS("AccReadyCB"));
|
|
|
|
|
|
2020-11-15 03:31:32 +08:00
|
|
|
|
trace_rcu_segcb_stats(&rdp->cblist, TPS("SegCbPostAcc"));
|
|
|
|
|
|
rcu: Make callers awaken grace-period kthread
The rcu_start_gp_advanced() function currently uses irq_work_queue()
to defer wakeups of the RCU grace-period kthread. This deferring
is necessary to avoid RCU-scheduler deadlocks involving the rcu_node
structure's lock, meaning that RCU cannot call any of the scheduler's
wake-up functions while holding one of these locks.
Unfortunately, the second and subsequent calls to irq_work_queue() are
ignored, and the first call will be ignored (aside from queuing the work
item) if the scheduler-clock tick is turned off. This is OK for many
uses, especially those where irq_work_queue() is called from an interrupt
or softirq handler, because in those cases the scheduler-clock-tick state
will be re-evaluated, which will turn the scheduler-clock tick back on.
On the next tick, any deferred work will then be processed.
However, this strategy does not always work for RCU, which can be invoked
at process level from idle CPUs. In this case, the tick might never
be turned back on, indefinitely defering a grace-period start request.
Note that the RCU CPU stall detector cannot see this condition, because
there is no RCU grace period in progress. Therefore, we can (and do!)
see long tens-of-seconds stalls in grace-period handling. In theory,
we could see a full grace-period hang, but rcutorture testing to date
has seen only the tens-of-seconds stalls. Event tracing demonstrates
that irq_work_queue() is being called repeatedly to no effect during
these stalls: The "newreq" event appears repeatedly from a task that is
not one of the grace-period kthreads.
In theory, irq_work_queue() might be fixed to avoid this sort of issue,
but RCU's requirements are unusual and it is quite straightforward to pass
wake-up responsibility up through RCU's call chain, so that the wakeup
happens when the offending locks are released.
This commit therefore makes this change. The rcu_start_gp_advanced(),
rcu_start_future_gp(), rcu_accelerate_cbs(), rcu_advance_cbs(),
__note_gp_changes(), and rcu_start_gp() functions now return a boolean
which indicates when a wake-up is needed. A new rcu_gp_kthread_wake()
does the wakeup when it is necessary and safe to do so: No self-wakes,
no wake-ups if the ->gp_flags field indicates there is no need (as in
someone else did the wake-up before we got around to it), and no wake-ups
before the grace-period kthread has been created.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2014-03-12 04:02:16 +08:00
|
|
|
|
return ret;
|
2012-12-04 05:52:00 +08:00
|
|
|
|
}
|
|
|
|
|
|
2018-05-02 07:29:47 +08:00
|
|
|
|
/*
|
|
|
|
|
* Similar to rcu_accelerate_cbs(), but does not require that the leaf
|
|
|
|
|
* rcu_node structure's ->lock be held. It consults the cached value
|
|
|
|
|
* of ->gp_seq_needed in the rcu_data structure, and if that indicates
|
|
|
|
|
* that a new grace-period request be made, invokes rcu_accelerate_cbs()
|
|
|
|
|
* while holding the leaf rcu_node structure's ->lock.
|
|
|
|
|
*/
|
2018-07-04 08:22:34 +08:00
|
|
|
|
static void rcu_accelerate_cbs_unlocked(struct rcu_node *rnp,
|
2018-05-02 07:29:47 +08:00
|
|
|
|
struct rcu_data *rdp)
|
|
|
|
|
{
|
|
|
|
|
unsigned long c;
|
|
|
|
|
bool needwake;
|
|
|
|
|
|
rcu/nocb: Add bypass callback queueing
Use of the rcu_data structure's segmented ->cblist for no-CBs CPUs
takes advantage of unrelated grace periods, thus reducing the memory
footprint in the face of floods of call_rcu() invocations. However,
the ->cblist field is a more-complex rcu_segcblist structure which must
be protected via locking. Even though there are only three entities
which can acquire this lock (the CPU invoking call_rcu(), the no-CBs
grace-period kthread, and the no-CBs callbacks kthread), the contention
on this lock is excessive under heavy stress.
This commit therefore greatly reduces contention by provisioning
an rcu_cblist structure field named ->nocb_bypass within the
rcu_data structure. Each no-CBs CPU is permitted only a limited
number of enqueues onto the ->cblist per jiffy, controlled by a new
nocb_nobypass_lim_per_jiffy kernel boot parameter that defaults to
about 16 enqueues per millisecond (16 * 1000 / HZ). When that limit is
exceeded, the CPU instead enqueues onto the new ->nocb_bypass.
The ->nocb_bypass is flushed into the ->cblist every jiffy or when
the number of callbacks on ->nocb_bypass exceeds qhimark, whichever
happens first. During call_rcu() floods, this flushing is carried out
by the CPU during the course of its call_rcu() invocations. However,
a CPU could simply stop invoking call_rcu() at any time. The no-CBs
grace-period kthread therefore carries out less-aggressive flushing
(every few jiffies or when the number of callbacks on ->nocb_bypass
exceeds (2 * qhimark), whichever comes first). This means that the
no-CBs grace-period kthread cannot be permitted to do unbounded waits
while there are callbacks on ->nocb_bypass. A ->nocb_bypass_timer is
used to provide the needed wakeups.
[ paulmck: Apply Coverity feedback reported by Colin Ian King. ]
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
2019-07-03 07:03:33 +08:00
|
|
|
|
rcu_lockdep_assert_cblist_protected(rdp);
|
2018-07-04 08:22:34 +08:00
|
|
|
|
c = rcu_seq_snap(&rcu_state.gp_seq);
|
2020-01-08 07:48:39 +08:00
|
|
|
|
if (!READ_ONCE(rdp->gpwrap) && ULONG_CMP_GE(rdp->gp_seq_needed, c)) {
|
2018-05-02 07:29:47 +08:00
|
|
|
|
/* Old request still live, so mark recent callbacks. */
|
|
|
|
|
(void)rcu_segcblist_accelerate(&rdp->cblist, c);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
raw_spin_lock_rcu_node(rnp); /* irqs already disabled. */
|
2018-07-04 08:22:34 +08:00
|
|
|
|
needwake = rcu_accelerate_cbs(rnp, rdp);
|
2018-05-02 07:29:47 +08:00
|
|
|
|
raw_spin_unlock_rcu_node(rnp); /* irqs remain disabled. */
|
|
|
|
|
if (needwake)
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_gp_kthread_wake();
|
2018-05-02 07:29:47 +08:00
|
|
|
|
}
|
|
|
|
|
|
2012-12-04 05:52:00 +08:00
|
|
|
|
/*
|
|
|
|
|
* Move any callbacks whose grace period has completed to the
|
|
|
|
|
* RCU_DONE_TAIL sublist, then compact the remaining sublists and
|
2018-05-01 01:57:36 +08:00
|
|
|
|
* assign ->gp_seq numbers to any callbacks in the RCU_NEXT_TAIL
|
2012-12-04 05:52:00 +08:00
|
|
|
|
* sublist. This function is idempotent, so it does not hurt to
|
|
|
|
|
* invoke it repeatedly. As long as it is not invoked -too- often...
|
rcu: Make callers awaken grace-period kthread
The rcu_start_gp_advanced() function currently uses irq_work_queue()
to defer wakeups of the RCU grace-period kthread. This deferring
is necessary to avoid RCU-scheduler deadlocks involving the rcu_node
structure's lock, meaning that RCU cannot call any of the scheduler's
wake-up functions while holding one of these locks.
Unfortunately, the second and subsequent calls to irq_work_queue() are
ignored, and the first call will be ignored (aside from queuing the work
item) if the scheduler-clock tick is turned off. This is OK for many
uses, especially those where irq_work_queue() is called from an interrupt
or softirq handler, because in those cases the scheduler-clock-tick state
will be re-evaluated, which will turn the scheduler-clock tick back on.
On the next tick, any deferred work will then be processed.
However, this strategy does not always work for RCU, which can be invoked
at process level from idle CPUs. In this case, the tick might never
be turned back on, indefinitely defering a grace-period start request.
Note that the RCU CPU stall detector cannot see this condition, because
there is no RCU grace period in progress. Therefore, we can (and do!)
see long tens-of-seconds stalls in grace-period handling. In theory,
we could see a full grace-period hang, but rcutorture testing to date
has seen only the tens-of-seconds stalls. Event tracing demonstrates
that irq_work_queue() is being called repeatedly to no effect during
these stalls: The "newreq" event appears repeatedly from a task that is
not one of the grace-period kthreads.
In theory, irq_work_queue() might be fixed to avoid this sort of issue,
but RCU's requirements are unusual and it is quite straightforward to pass
wake-up responsibility up through RCU's call chain, so that the wakeup
happens when the offending locks are released.
This commit therefore makes this change. The rcu_start_gp_advanced(),
rcu_start_future_gp(), rcu_accelerate_cbs(), rcu_advance_cbs(),
__note_gp_changes(), and rcu_start_gp() functions now return a boolean
which indicates when a wake-up is needed. A new rcu_gp_kthread_wake()
does the wakeup when it is necessary and safe to do so: No self-wakes,
no wake-ups if the ->gp_flags field indicates there is no need (as in
someone else did the wake-up before we got around to it), and no wake-ups
before the grace-period kthread has been created.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2014-03-12 04:02:16 +08:00
|
|
|
|
* Returns true if the RCU grace-period kthread needs to be awakened.
|
2012-12-04 05:52:00 +08:00
|
|
|
|
*
|
|
|
|
|
* The caller must hold rnp->lock with interrupts disabled.
|
|
|
|
|
*/
|
2018-07-04 08:22:34 +08:00
|
|
|
|
static bool rcu_advance_cbs(struct rcu_node *rnp, struct rcu_data *rdp)
|
2012-12-04 05:52:00 +08:00
|
|
|
|
{
|
rcu/nocb: Add bypass callback queueing
Use of the rcu_data structure's segmented ->cblist for no-CBs CPUs
takes advantage of unrelated grace periods, thus reducing the memory
footprint in the face of floods of call_rcu() invocations. However,
the ->cblist field is a more-complex rcu_segcblist structure which must
be protected via locking. Even though there are only three entities
which can acquire this lock (the CPU invoking call_rcu(), the no-CBs
grace-period kthread, and the no-CBs callbacks kthread), the contention
on this lock is excessive under heavy stress.
This commit therefore greatly reduces contention by provisioning
an rcu_cblist structure field named ->nocb_bypass within the
rcu_data structure. Each no-CBs CPU is permitted only a limited
number of enqueues onto the ->cblist per jiffy, controlled by a new
nocb_nobypass_lim_per_jiffy kernel boot parameter that defaults to
about 16 enqueues per millisecond (16 * 1000 / HZ). When that limit is
exceeded, the CPU instead enqueues onto the new ->nocb_bypass.
The ->nocb_bypass is flushed into the ->cblist every jiffy or when
the number of callbacks on ->nocb_bypass exceeds qhimark, whichever
happens first. During call_rcu() floods, this flushing is carried out
by the CPU during the course of its call_rcu() invocations. However,
a CPU could simply stop invoking call_rcu() at any time. The no-CBs
grace-period kthread therefore carries out less-aggressive flushing
(every few jiffies or when the number of callbacks on ->nocb_bypass
exceeds (2 * qhimark), whichever comes first). This means that the
no-CBs grace-period kthread cannot be permitted to do unbounded waits
while there are callbacks on ->nocb_bypass. A ->nocb_bypass_timer is
used to provide the needed wakeups.
[ paulmck: Apply Coverity feedback reported by Colin Ian King. ]
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
2019-07-03 07:03:33 +08:00
|
|
|
|
rcu_lockdep_assert_cblist_protected(rdp);
|
2018-01-17 22:24:30 +08:00
|
|
|
|
raw_lockdep_assert_held_rcu_node(rnp);
|
2017-04-29 03:32:15 +08:00
|
|
|
|
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
/* If no pending (not yet ready to invoke) callbacks, nothing to do. */
|
|
|
|
|
if (!rcu_segcblist_pend_cbs(&rdp->cblist))
|
rcu: Make callers awaken grace-period kthread
The rcu_start_gp_advanced() function currently uses irq_work_queue()
to defer wakeups of the RCU grace-period kthread. This deferring
is necessary to avoid RCU-scheduler deadlocks involving the rcu_node
structure's lock, meaning that RCU cannot call any of the scheduler's
wake-up functions while holding one of these locks.
Unfortunately, the second and subsequent calls to irq_work_queue() are
ignored, and the first call will be ignored (aside from queuing the work
item) if the scheduler-clock tick is turned off. This is OK for many
uses, especially those where irq_work_queue() is called from an interrupt
or softirq handler, because in those cases the scheduler-clock-tick state
will be re-evaluated, which will turn the scheduler-clock tick back on.
On the next tick, any deferred work will then be processed.
However, this strategy does not always work for RCU, which can be invoked
at process level from idle CPUs. In this case, the tick might never
be turned back on, indefinitely defering a grace-period start request.
Note that the RCU CPU stall detector cannot see this condition, because
there is no RCU grace period in progress. Therefore, we can (and do!)
see long tens-of-seconds stalls in grace-period handling. In theory,
we could see a full grace-period hang, but rcutorture testing to date
has seen only the tens-of-seconds stalls. Event tracing demonstrates
that irq_work_queue() is being called repeatedly to no effect during
these stalls: The "newreq" event appears repeatedly from a task that is
not one of the grace-period kthreads.
In theory, irq_work_queue() might be fixed to avoid this sort of issue,
but RCU's requirements are unusual and it is quite straightforward to pass
wake-up responsibility up through RCU's call chain, so that the wakeup
happens when the offending locks are released.
This commit therefore makes this change. The rcu_start_gp_advanced(),
rcu_start_future_gp(), rcu_accelerate_cbs(), rcu_advance_cbs(),
__note_gp_changes(), and rcu_start_gp() functions now return a boolean
which indicates when a wake-up is needed. A new rcu_gp_kthread_wake()
does the wakeup when it is necessary and safe to do so: No self-wakes,
no wake-ups if the ->gp_flags field indicates there is no need (as in
someone else did the wake-up before we got around to it), and no wake-ups
before the grace-period kthread has been created.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2014-03-12 04:02:16 +08:00
|
|
|
|
return false;
|
2012-12-04 05:52:00 +08:00
|
|
|
|
|
|
|
|
|
/*
|
2018-05-01 01:57:36 +08:00
|
|
|
|
* Find all callbacks whose ->gp_seq numbers indicate that they
|
2012-12-04 05:52:00 +08:00
|
|
|
|
* are ready to invoke, and put them into the RCU_DONE_TAIL sublist.
|
|
|
|
|
*/
|
2018-05-01 01:57:36 +08:00
|
|
|
|
rcu_segcblist_advance(&rdp->cblist, rnp->gp_seq);
|
2012-12-04 05:52:00 +08:00
|
|
|
|
|
|
|
|
|
/* Classify any remaining callbacks. */
|
2018-07-04 08:22:34 +08:00
|
|
|
|
return rcu_accelerate_cbs(rnp, rdp);
|
2012-12-04 05:52:00 +08:00
|
|
|
|
}
|
|
|
|
|
|
2019-05-28 20:54:26 +08:00
|
|
|
|
/*
|
|
|
|
|
* Move and classify callbacks, but only if doing so won't require
|
|
|
|
|
* that the RCU grace-period kthread be awakened.
|
|
|
|
|
*/
|
|
|
|
|
static void __maybe_unused rcu_advance_cbs_nowake(struct rcu_node *rnp,
|
|
|
|
|
struct rcu_data *rdp)
|
|
|
|
|
{
|
rcu/nocb: Add bypass callback queueing
Use of the rcu_data structure's segmented ->cblist for no-CBs CPUs
takes advantage of unrelated grace periods, thus reducing the memory
footprint in the face of floods of call_rcu() invocations. However,
the ->cblist field is a more-complex rcu_segcblist structure which must
be protected via locking. Even though there are only three entities
which can acquire this lock (the CPU invoking call_rcu(), the no-CBs
grace-period kthread, and the no-CBs callbacks kthread), the contention
on this lock is excessive under heavy stress.
This commit therefore greatly reduces contention by provisioning
an rcu_cblist structure field named ->nocb_bypass within the
rcu_data structure. Each no-CBs CPU is permitted only a limited
number of enqueues onto the ->cblist per jiffy, controlled by a new
nocb_nobypass_lim_per_jiffy kernel boot parameter that defaults to
about 16 enqueues per millisecond (16 * 1000 / HZ). When that limit is
exceeded, the CPU instead enqueues onto the new ->nocb_bypass.
The ->nocb_bypass is flushed into the ->cblist every jiffy or when
the number of callbacks on ->nocb_bypass exceeds qhimark, whichever
happens first. During call_rcu() floods, this flushing is carried out
by the CPU during the course of its call_rcu() invocations. However,
a CPU could simply stop invoking call_rcu() at any time. The no-CBs
grace-period kthread therefore carries out less-aggressive flushing
(every few jiffies or when the number of callbacks on ->nocb_bypass
exceeds (2 * qhimark), whichever comes first). This means that the
no-CBs grace-period kthread cannot be permitted to do unbounded waits
while there are callbacks on ->nocb_bypass. A ->nocb_bypass_timer is
used to provide the needed wakeups.
[ paulmck: Apply Coverity feedback reported by Colin Ian King. ]
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
2019-07-03 07:03:33 +08:00
|
|
|
|
rcu_lockdep_assert_cblist_protected(rdp);
|
2021-09-18 06:04:48 +08:00
|
|
|
|
if (!rcu_seq_state(rcu_seq_current(&rnp->gp_seq)) || !raw_spin_trylock_rcu_node(rnp))
|
2019-05-28 20:54:26 +08:00
|
|
|
|
return;
|
2021-09-18 06:04:48 +08:00
|
|
|
|
// The grace period cannot end while we hold the rcu_node lock.
|
|
|
|
|
if (rcu_seq_state(rcu_seq_current(&rnp->gp_seq)))
|
|
|
|
|
WARN_ON_ONCE(rcu_advance_cbs(rnp, rdp));
|
2019-06-01 21:16:38 +08:00
|
|
|
|
raw_spin_unlock_rcu_node(rnp);
|
2019-05-28 20:54:26 +08:00
|
|
|
|
}
|
|
|
|
|
|
2020-08-07 07:35:08 +08:00
|
|
|
|
/*
|
|
|
|
|
* In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels, attempt to generate a
|
|
|
|
|
* quiescent state. This is intended to be invoked when the CPU notices
|
|
|
|
|
* a new grace period.
|
|
|
|
|
*/
|
|
|
|
|
static void rcu_strict_gp_check_qs(void)
|
|
|
|
|
{
|
|
|
|
|
if (IS_ENABLED(CONFIG_RCU_STRICT_GRACE_PERIOD)) {
|
|
|
|
|
rcu_read_lock();
|
|
|
|
|
rcu_read_unlock();
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2009-11-03 05:52:28 +08:00
|
|
|
|
/*
|
2013-03-20 02:53:31 +08:00
|
|
|
|
* Update CPU-local rcu_data state to record the beginnings and ends of
|
|
|
|
|
* grace periods. The caller must hold the ->lock of the leaf rcu_node
|
|
|
|
|
* structure corresponding to the current CPU, and must have irqs disabled.
|
rcu: Make callers awaken grace-period kthread
The rcu_start_gp_advanced() function currently uses irq_work_queue()
to defer wakeups of the RCU grace-period kthread. This deferring
is necessary to avoid RCU-scheduler deadlocks involving the rcu_node
structure's lock, meaning that RCU cannot call any of the scheduler's
wake-up functions while holding one of these locks.
Unfortunately, the second and subsequent calls to irq_work_queue() are
ignored, and the first call will be ignored (aside from queuing the work
item) if the scheduler-clock tick is turned off. This is OK for many
uses, especially those where irq_work_queue() is called from an interrupt
or softirq handler, because in those cases the scheduler-clock-tick state
will be re-evaluated, which will turn the scheduler-clock tick back on.
On the next tick, any deferred work will then be processed.
However, this strategy does not always work for RCU, which can be invoked
at process level from idle CPUs. In this case, the tick might never
be turned back on, indefinitely defering a grace-period start request.
Note that the RCU CPU stall detector cannot see this condition, because
there is no RCU grace period in progress. Therefore, we can (and do!)
see long tens-of-seconds stalls in grace-period handling. In theory,
we could see a full grace-period hang, but rcutorture testing to date
has seen only the tens-of-seconds stalls. Event tracing demonstrates
that irq_work_queue() is being called repeatedly to no effect during
these stalls: The "newreq" event appears repeatedly from a task that is
not one of the grace-period kthreads.
In theory, irq_work_queue() might be fixed to avoid this sort of issue,
but RCU's requirements are unusual and it is quite straightforward to pass
wake-up responsibility up through RCU's call chain, so that the wakeup
happens when the offending locks are released.
This commit therefore makes this change. The rcu_start_gp_advanced(),
rcu_start_future_gp(), rcu_accelerate_cbs(), rcu_advance_cbs(),
__note_gp_changes(), and rcu_start_gp() functions now return a boolean
which indicates when a wake-up is needed. A new rcu_gp_kthread_wake()
does the wakeup when it is necessary and safe to do so: No self-wakes,
no wake-ups if the ->gp_flags field indicates there is no need (as in
someone else did the wake-up before we got around to it), and no wake-ups
before the grace-period kthread has been created.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2014-03-12 04:02:16 +08:00
|
|
|
|
* Returns true if the grace-period kthread needs to be awakened.
|
2009-11-03 05:52:28 +08:00
|
|
|
|
*/
|
2018-07-04 08:22:34 +08:00
|
|
|
|
static bool __note_gp_changes(struct rcu_node *rnp, struct rcu_data *rdp)
|
2009-11-03 05:52:28 +08:00
|
|
|
|
{
|
2019-05-16 00:56:40 +08:00
|
|
|
|
bool ret = false;
|
rcu: Clear ->core_needs_qs at GP end or self-reported QS
The rcu_data structure's ->core_needs_qs field does not necessarily get
cleared in a timely fashion after the corresponding CPUs' quiescent state
has been reported. From a functional viewpoint, no harm done, but this
can result in excessive invocation of RCU core processing, as witnessed
by the kernel test robot, which saw greatly increased softirq overhead.
This commit therefore restores the rcu_report_qs_rdp() function's
clearing of this field, but only when running on the corresponding CPU.
Cases where some other CPU reports the quiescent state (for example, on
behalf of an idle CPU) are handled by setting this field appropriately
within the __note_gp_changes() function's end-of-grace-period checks.
This handling is carried out regardless of whether the end of a grace
period actually happened, thus handling the case where a CPU goes non-idle
after a quiescent state is reported on its behalf, but before the grace
period ends. This fix also avoids cross-CPU updates to ->core_needs_qs,
While in the area, this commit changes the __note_gp_changes() need_gp
variable's name to need_qs because it is a quiescent state that is needed
from the CPU in question.
Fixes: ed93dfc6bc00 ("rcu: Confine ->core_needs_qs accesses to the corresponding CPU")
Reported-by: kernel test robot <rong.a.chen@intel.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-12-10 07:19:45 +08:00
|
|
|
|
bool need_qs;
|
2020-11-12 08:51:21 +08:00
|
|
|
|
const bool offloaded = rcu_rdp_is_offloaded(rdp);
|
rcu: Make callers awaken grace-period kthread
The rcu_start_gp_advanced() function currently uses irq_work_queue()
to defer wakeups of the RCU grace-period kthread. This deferring
is necessary to avoid RCU-scheduler deadlocks involving the rcu_node
structure's lock, meaning that RCU cannot call any of the scheduler's
wake-up functions while holding one of these locks.
Unfortunately, the second and subsequent calls to irq_work_queue() are
ignored, and the first call will be ignored (aside from queuing the work
item) if the scheduler-clock tick is turned off. This is OK for many
uses, especially those where irq_work_queue() is called from an interrupt
or softirq handler, because in those cases the scheduler-clock-tick state
will be re-evaluated, which will turn the scheduler-clock tick back on.
On the next tick, any deferred work will then be processed.
However, this strategy does not always work for RCU, which can be invoked
at process level from idle CPUs. In this case, the tick might never
be turned back on, indefinitely defering a grace-period start request.
Note that the RCU CPU stall detector cannot see this condition, because
there is no RCU grace period in progress. Therefore, we can (and do!)
see long tens-of-seconds stalls in grace-period handling. In theory,
we could see a full grace-period hang, but rcutorture testing to date
has seen only the tens-of-seconds stalls. Event tracing demonstrates
that irq_work_queue() is being called repeatedly to no effect during
these stalls: The "newreq" event appears repeatedly from a task that is
not one of the grace-period kthreads.
In theory, irq_work_queue() might be fixed to avoid this sort of issue,
but RCU's requirements are unusual and it is quite straightforward to pass
wake-up responsibility up through RCU's call chain, so that the wakeup
happens when the offending locks are released.
This commit therefore makes this change. The rcu_start_gp_advanced(),
rcu_start_future_gp(), rcu_accelerate_cbs(), rcu_advance_cbs(),
__note_gp_changes(), and rcu_start_gp() functions now return a boolean
which indicates when a wake-up is needed. A new rcu_gp_kthread_wake()
does the wakeup when it is necessary and safe to do so: No self-wakes,
no wake-ups if the ->gp_flags field indicates there is no need (as in
someone else did the wake-up before we got around to it), and no wake-ups
before the grace-period kthread has been created.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2014-03-12 04:02:16 +08:00
|
|
|
|
|
2018-01-17 22:24:30 +08:00
|
|
|
|
raw_lockdep_assert_held_rcu_node(rnp);
|
2017-04-29 03:32:15 +08:00
|
|
|
|
|
2018-04-28 07:01:46 +08:00
|
|
|
|
if (rdp->gp_seq == rnp->gp_seq)
|
|
|
|
|
return false; /* Nothing to do. */
|
2009-11-03 05:52:28 +08:00
|
|
|
|
|
2018-04-28 07:01:46 +08:00
|
|
|
|
/* Handle the ends of any preceding grace periods first. */
|
|
|
|
|
if (rcu_seq_completed_gp(rdp->gp_seq, rnp->gp_seq) ||
|
|
|
|
|
unlikely(READ_ONCE(rdp->gpwrap))) {
|
2019-05-16 00:56:40 +08:00
|
|
|
|
if (!offloaded)
|
|
|
|
|
ret = rcu_advance_cbs(rnp, rdp); /* Advance CBs. */
|
rcu: Clear ->core_needs_qs at GP end or self-reported QS
The rcu_data structure's ->core_needs_qs field does not necessarily get
cleared in a timely fashion after the corresponding CPUs' quiescent state
has been reported. From a functional viewpoint, no harm done, but this
can result in excessive invocation of RCU core processing, as witnessed
by the kernel test robot, which saw greatly increased softirq overhead.
This commit therefore restores the rcu_report_qs_rdp() function's
clearing of this field, but only when running on the corresponding CPU.
Cases where some other CPU reports the quiescent state (for example, on
behalf of an idle CPU) are handled by setting this field appropriately
within the __note_gp_changes() function's end-of-grace-period checks.
This handling is carried out regardless of whether the end of a grace
period actually happened, thus handling the case where a CPU goes non-idle
after a quiescent state is reported on its behalf, but before the grace
period ends. This fix also avoids cross-CPU updates to ->core_needs_qs,
While in the area, this commit changes the __note_gp_changes() need_gp
variable's name to need_qs because it is a quiescent state that is needed
from the CPU in question.
Fixes: ed93dfc6bc00 ("rcu: Confine ->core_needs_qs accesses to the corresponding CPU")
Reported-by: kernel test robot <rong.a.chen@intel.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-12-10 07:19:45 +08:00
|
|
|
|
rdp->core_needs_qs = false;
|
2018-07-06 06:47:01 +08:00
|
|
|
|
trace_rcu_grace_period(rcu_state.name, rdp->gp_seq, TPS("cpuend"));
|
2018-04-28 07:01:46 +08:00
|
|
|
|
} else {
|
2019-05-16 00:56:40 +08:00
|
|
|
|
if (!offloaded)
|
|
|
|
|
ret = rcu_accelerate_cbs(rnp, rdp); /* Recent CBs. */
|
rcu: Clear ->core_needs_qs at GP end or self-reported QS
The rcu_data structure's ->core_needs_qs field does not necessarily get
cleared in a timely fashion after the corresponding CPUs' quiescent state
has been reported. From a functional viewpoint, no harm done, but this
can result in excessive invocation of RCU core processing, as witnessed
by the kernel test robot, which saw greatly increased softirq overhead.
This commit therefore restores the rcu_report_qs_rdp() function's
clearing of this field, but only when running on the corresponding CPU.
Cases where some other CPU reports the quiescent state (for example, on
behalf of an idle CPU) are handled by setting this field appropriately
within the __note_gp_changes() function's end-of-grace-period checks.
This handling is carried out regardless of whether the end of a grace
period actually happened, thus handling the case where a CPU goes non-idle
after a quiescent state is reported on its behalf, but before the grace
period ends. This fix also avoids cross-CPU updates to ->core_needs_qs,
While in the area, this commit changes the __note_gp_changes() need_gp
variable's name to need_qs because it is a quiescent state that is needed
from the CPU in question.
Fixes: ed93dfc6bc00 ("rcu: Confine ->core_needs_qs accesses to the corresponding CPU")
Reported-by: kernel test robot <rong.a.chen@intel.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-12-10 07:19:45 +08:00
|
|
|
|
if (rdp->core_needs_qs)
|
|
|
|
|
rdp->core_needs_qs = !!(rnp->qsmask & rdp->grpmask);
|
2009-11-03 05:52:28 +08:00
|
|
|
|
}
|
2013-03-20 01:53:14 +08:00
|
|
|
|
|
2018-04-28 07:01:46 +08:00
|
|
|
|
/* Now handle the beginnings of any new-to-this-CPU grace periods. */
|
|
|
|
|
if (rcu_seq_new_gp(rdp->gp_seq, rnp->gp_seq) ||
|
|
|
|
|
unlikely(READ_ONCE(rdp->gpwrap))) {
|
2013-03-20 01:08:37 +08:00
|
|
|
|
/*
|
|
|
|
|
* If the current grace period is waiting for this CPU,
|
|
|
|
|
* set up to detect a quiescent state, otherwise don't
|
|
|
|
|
* go looking for one.
|
|
|
|
|
*/
|
2018-07-06 06:47:01 +08:00
|
|
|
|
trace_rcu_grace_period(rcu_state.name, rnp->gp_seq, TPS("cpustart"));
|
rcu: Clear ->core_needs_qs at GP end or self-reported QS
The rcu_data structure's ->core_needs_qs field does not necessarily get
cleared in a timely fashion after the corresponding CPUs' quiescent state
has been reported. From a functional viewpoint, no harm done, but this
can result in excessive invocation of RCU core processing, as witnessed
by the kernel test robot, which saw greatly increased softirq overhead.
This commit therefore restores the rcu_report_qs_rdp() function's
clearing of this field, but only when running on the corresponding CPU.
Cases where some other CPU reports the quiescent state (for example, on
behalf of an idle CPU) are handled by setting this field appropriately
within the __note_gp_changes() function's end-of-grace-period checks.
This handling is carried out regardless of whether the end of a grace
period actually happened, thus handling the case where a CPU goes non-idle
after a quiescent state is reported on its behalf, but before the grace
period ends. This fix also avoids cross-CPU updates to ->core_needs_qs,
While in the area, this commit changes the __note_gp_changes() need_gp
variable's name to need_qs because it is a quiescent state that is needed
from the CPU in question.
Fixes: ed93dfc6bc00 ("rcu: Confine ->core_needs_qs accesses to the corresponding CPU")
Reported-by: kernel test robot <rong.a.chen@intel.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-12-10 07:19:45 +08:00
|
|
|
|
need_qs = !!(rnp->qsmask & rdp->grpmask);
|
|
|
|
|
rdp->cpu_no_qs.b.norm = need_qs;
|
|
|
|
|
rdp->core_needs_qs = need_qs;
|
2013-03-20 01:08:37 +08:00
|
|
|
|
zero_cpu_stall_ticks(rdp);
|
|
|
|
|
}
|
2018-04-28 07:01:46 +08:00
|
|
|
|
rdp->gp_seq = rnp->gp_seq; /* Remember new grace-period state. */
|
2018-12-20 02:37:34 +08:00
|
|
|
|
if (ULONG_CMP_LT(rdp->gp_seq_needed, rnp->gp_seq_needed) || rdp->gpwrap)
|
2020-01-05 03:33:17 +08:00
|
|
|
|
WRITE_ONCE(rdp->gp_seq_needed, rnp->gp_seq_needed);
|
2022-02-24 09:29:37 +08:00
|
|
|
|
if (IS_ENABLED(CONFIG_PROVE_RCU) && READ_ONCE(rdp->gpwrap))
|
|
|
|
|
WRITE_ONCE(rdp->last_sched_clock, jiffies);
|
2018-05-16 07:47:30 +08:00
|
|
|
|
WRITE_ONCE(rdp->gpwrap, false);
|
|
|
|
|
rcu_gpnum_ovf(rnp, rdp);
|
rcu: Make callers awaken grace-period kthread
The rcu_start_gp_advanced() function currently uses irq_work_queue()
to defer wakeups of the RCU grace-period kthread. This deferring
is necessary to avoid RCU-scheduler deadlocks involving the rcu_node
structure's lock, meaning that RCU cannot call any of the scheduler's
wake-up functions while holding one of these locks.
Unfortunately, the second and subsequent calls to irq_work_queue() are
ignored, and the first call will be ignored (aside from queuing the work
item) if the scheduler-clock tick is turned off. This is OK for many
uses, especially those where irq_work_queue() is called from an interrupt
or softirq handler, because in those cases the scheduler-clock-tick state
will be re-evaluated, which will turn the scheduler-clock tick back on.
On the next tick, any deferred work will then be processed.
However, this strategy does not always work for RCU, which can be invoked
at process level from idle CPUs. In this case, the tick might never
be turned back on, indefinitely defering a grace-period start request.
Note that the RCU CPU stall detector cannot see this condition, because
there is no RCU grace period in progress. Therefore, we can (and do!)
see long tens-of-seconds stalls in grace-period handling. In theory,
we could see a full grace-period hang, but rcutorture testing to date
has seen only the tens-of-seconds stalls. Event tracing demonstrates
that irq_work_queue() is being called repeatedly to no effect during
these stalls: The "newreq" event appears repeatedly from a task that is
not one of the grace-period kthreads.
In theory, irq_work_queue() might be fixed to avoid this sort of issue,
but RCU's requirements are unusual and it is quite straightforward to pass
wake-up responsibility up through RCU's call chain, so that the wakeup
happens when the offending locks are released.
This commit therefore makes this change. The rcu_start_gp_advanced(),
rcu_start_future_gp(), rcu_accelerate_cbs(), rcu_advance_cbs(),
__note_gp_changes(), and rcu_start_gp() functions now return a boolean
which indicates when a wake-up is needed. A new rcu_gp_kthread_wake()
does the wakeup when it is necessary and safe to do so: No self-wakes,
no wake-ups if the ->gp_flags field indicates there is no need (as in
someone else did the wake-up before we got around to it), and no wake-ups
before the grace-period kthread has been created.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2014-03-12 04:02:16 +08:00
|
|
|
|
return ret;
|
2013-03-20 01:08:37 +08:00
|
|
|
|
}
|
|
|
|
|
|
2018-07-04 08:22:34 +08:00
|
|
|
|
static void note_gp_changes(struct rcu_data *rdp)
|
2013-03-20 01:08:37 +08:00
|
|
|
|
{
|
|
|
|
|
unsigned long flags;
|
rcu: Make callers awaken grace-period kthread
The rcu_start_gp_advanced() function currently uses irq_work_queue()
to defer wakeups of the RCU grace-period kthread. This deferring
is necessary to avoid RCU-scheduler deadlocks involving the rcu_node
structure's lock, meaning that RCU cannot call any of the scheduler's
wake-up functions while holding one of these locks.
Unfortunately, the second and subsequent calls to irq_work_queue() are
ignored, and the first call will be ignored (aside from queuing the work
item) if the scheduler-clock tick is turned off. This is OK for many
uses, especially those where irq_work_queue() is called from an interrupt
or softirq handler, because in those cases the scheduler-clock-tick state
will be re-evaluated, which will turn the scheduler-clock tick back on.
On the next tick, any deferred work will then be processed.
However, this strategy does not always work for RCU, which can be invoked
at process level from idle CPUs. In this case, the tick might never
be turned back on, indefinitely defering a grace-period start request.
Note that the RCU CPU stall detector cannot see this condition, because
there is no RCU grace period in progress. Therefore, we can (and do!)
see long tens-of-seconds stalls in grace-period handling. In theory,
we could see a full grace-period hang, but rcutorture testing to date
has seen only the tens-of-seconds stalls. Event tracing demonstrates
that irq_work_queue() is being called repeatedly to no effect during
these stalls: The "newreq" event appears repeatedly from a task that is
not one of the grace-period kthreads.
In theory, irq_work_queue() might be fixed to avoid this sort of issue,
but RCU's requirements are unusual and it is quite straightforward to pass
wake-up responsibility up through RCU's call chain, so that the wakeup
happens when the offending locks are released.
This commit therefore makes this change. The rcu_start_gp_advanced(),
rcu_start_future_gp(), rcu_accelerate_cbs(), rcu_advance_cbs(),
__note_gp_changes(), and rcu_start_gp() functions now return a boolean
which indicates when a wake-up is needed. A new rcu_gp_kthread_wake()
does the wakeup when it is necessary and safe to do so: No self-wakes,
no wake-ups if the ->gp_flags field indicates there is no need (as in
someone else did the wake-up before we got around to it), and no wake-ups
before the grace-period kthread has been created.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2014-03-12 04:02:16 +08:00
|
|
|
|
bool needwake;
|
2013-03-20 01:08:37 +08:00
|
|
|
|
struct rcu_node *rnp;
|
|
|
|
|
|
|
|
|
|
local_irq_save(flags);
|
|
|
|
|
rnp = rdp->mynode;
|
2018-04-28 07:01:46 +08:00
|
|
|
|
if ((rdp->gp_seq == rcu_seq_current(&rnp->gp_seq) &&
|
2015-03-04 06:57:58 +08:00
|
|
|
|
!unlikely(READ_ONCE(rdp->gpwrap))) || /* w/out lock. */
|
2015-10-08 18:24:23 +08:00
|
|
|
|
!raw_spin_trylock_rcu_node(rnp)) { /* irqs already off, so later. */
|
2013-03-20 01:08:37 +08:00
|
|
|
|
local_irq_restore(flags);
|
|
|
|
|
return;
|
|
|
|
|
}
|
2018-07-04 08:22:34 +08:00
|
|
|
|
needwake = __note_gp_changes(rnp, rdp);
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
|
2020-08-07 07:35:08 +08:00
|
|
|
|
rcu_strict_gp_check_qs();
|
rcu: Make callers awaken grace-period kthread
The rcu_start_gp_advanced() function currently uses irq_work_queue()
to defer wakeups of the RCU grace-period kthread. This deferring
is necessary to avoid RCU-scheduler deadlocks involving the rcu_node
structure's lock, meaning that RCU cannot call any of the scheduler's
wake-up functions while holding one of these locks.
Unfortunately, the second and subsequent calls to irq_work_queue() are
ignored, and the first call will be ignored (aside from queuing the work
item) if the scheduler-clock tick is turned off. This is OK for many
uses, especially those where irq_work_queue() is called from an interrupt
or softirq handler, because in those cases the scheduler-clock-tick state
will be re-evaluated, which will turn the scheduler-clock tick back on.
On the next tick, any deferred work will then be processed.
However, this strategy does not always work for RCU, which can be invoked
at process level from idle CPUs. In this case, the tick might never
be turned back on, indefinitely defering a grace-period start request.
Note that the RCU CPU stall detector cannot see this condition, because
there is no RCU grace period in progress. Therefore, we can (and do!)
see long tens-of-seconds stalls in grace-period handling. In theory,
we could see a full grace-period hang, but rcutorture testing to date
has seen only the tens-of-seconds stalls. Event tracing demonstrates
that irq_work_queue() is being called repeatedly to no effect during
these stalls: The "newreq" event appears repeatedly from a task that is
not one of the grace-period kthreads.
In theory, irq_work_queue() might be fixed to avoid this sort of issue,
but RCU's requirements are unusual and it is quite straightforward to pass
wake-up responsibility up through RCU's call chain, so that the wakeup
happens when the offending locks are released.
This commit therefore makes this change. The rcu_start_gp_advanced(),
rcu_start_future_gp(), rcu_accelerate_cbs(), rcu_advance_cbs(),
__note_gp_changes(), and rcu_start_gp() functions now return a boolean
which indicates when a wake-up is needed. A new rcu_gp_kthread_wake()
does the wakeup when it is necessary and safe to do so: No self-wakes,
no wake-ups if the ->gp_flags field indicates there is no need (as in
someone else did the wake-up before we got around to it), and no wake-ups
before the grace-period kthread has been created.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2014-03-12 04:02:16 +08:00
|
|
|
|
if (needwake)
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_gp_kthread_wake();
|
2013-03-20 01:08:37 +08:00
|
|
|
|
}
|
|
|
|
|
|
2022-02-05 04:45:18 +08:00
|
|
|
|
static atomic_t *rcu_gp_slow_suppress;
|
|
|
|
|
|
|
|
|
|
/* Register a counter to suppress debugging grace-period delays. */
|
|
|
|
|
void rcu_gp_slow_register(atomic_t *rgssp)
|
|
|
|
|
{
|
|
|
|
|
WARN_ON_ONCE(rcu_gp_slow_suppress);
|
|
|
|
|
|
|
|
|
|
WRITE_ONCE(rcu_gp_slow_suppress, rgssp);
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(rcu_gp_slow_register);
|
|
|
|
|
|
|
|
|
|
/* Unregister a counter, with NULL for not caring which. */
|
|
|
|
|
void rcu_gp_slow_unregister(atomic_t *rgssp)
|
|
|
|
|
{
|
|
|
|
|
WARN_ON_ONCE(rgssp && rgssp != rcu_gp_slow_suppress);
|
|
|
|
|
|
|
|
|
|
WRITE_ONCE(rcu_gp_slow_suppress, NULL);
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(rcu_gp_slow_unregister);
|
|
|
|
|
|
|
|
|
|
static bool rcu_gp_slow_is_suppressed(void)
|
|
|
|
|
{
|
|
|
|
|
atomic_t *rgssp = READ_ONCE(rcu_gp_slow_suppress);
|
|
|
|
|
|
|
|
|
|
return rgssp && atomic_read(rgssp);
|
|
|
|
|
}
|
|
|
|
|
|
2018-07-04 08:22:34 +08:00
|
|
|
|
static void rcu_gp_slow(int delay)
|
2015-03-11 09:33:20 +08:00
|
|
|
|
{
|
2022-02-05 04:45:18 +08:00
|
|
|
|
if (!rcu_gp_slow_is_suppressed() && delay > 0 &&
|
|
|
|
|
!(rcu_seq_ctr(rcu_state.gp_seq) % (rcu_num_nodes * PER_RCU_NODE_PERIOD * delay)))
|
2020-05-08 06:44:46 +08:00
|
|
|
|
schedule_timeout_idle(delay);
|
2015-03-11 09:33:20 +08:00
|
|
|
|
}
|
|
|
|
|
|
2020-04-02 10:57:52 +08:00
|
|
|
|
static unsigned long sleep_duration;
|
|
|
|
|
|
|
|
|
|
/* Allow rcutorture to stall the grace-period kthread. */
|
|
|
|
|
void rcu_gp_set_torture_wait(int duration)
|
|
|
|
|
{
|
|
|
|
|
if (IS_ENABLED(CONFIG_RCU_TORTURE_TEST) && duration > 0)
|
|
|
|
|
WRITE_ONCE(sleep_duration, duration);
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(rcu_gp_set_torture_wait);
|
|
|
|
|
|
|
|
|
|
/* Actually implement the aforementioned wait. */
|
|
|
|
|
static void rcu_gp_torture_wait(void)
|
|
|
|
|
{
|
|
|
|
|
unsigned long duration;
|
|
|
|
|
|
|
|
|
|
if (!IS_ENABLED(CONFIG_RCU_TORTURE_TEST))
|
|
|
|
|
return;
|
|
|
|
|
duration = xchg(&sleep_duration, 0UL);
|
|
|
|
|
if (duration > 0) {
|
|
|
|
|
pr_alert("%s: Waiting %lu jiffies\n", __func__, duration);
|
2020-05-08 06:44:46 +08:00
|
|
|
|
schedule_timeout_idle(duration);
|
2020-04-02 10:57:52 +08:00
|
|
|
|
pr_alert("%s: Wait complete\n", __func__);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2020-08-07 10:21:48 +08:00
|
|
|
|
/*
|
|
|
|
|
* Handler for on_each_cpu() to invoke the target CPU's RCU core
|
|
|
|
|
* processing.
|
|
|
|
|
*/
|
|
|
|
|
static void rcu_strict_gp_boundary(void *unused)
|
|
|
|
|
{
|
|
|
|
|
invoke_rcu_core();
|
|
|
|
|
}
|
|
|
|
|
|
2022-04-14 08:46:15 +08:00
|
|
|
|
// Make the polled API aware of the beginning of a grace period.
|
|
|
|
|
static void rcu_poll_gp_seq_start(unsigned long *snap)
|
|
|
|
|
{
|
|
|
|
|
struct rcu_node *rnp = rcu_get_root();
|
|
|
|
|
|
2022-12-16 01:26:09 +08:00
|
|
|
|
if (rcu_scheduler_active != RCU_SCHEDULER_INACTIVE)
|
2022-04-14 08:46:15 +08:00
|
|
|
|
raw_lockdep_assert_held_rcu_node(rnp);
|
|
|
|
|
|
|
|
|
|
// If RCU was idle, note beginning of GP.
|
|
|
|
|
if (!rcu_seq_state(rcu_state.gp_seq_polled))
|
|
|
|
|
rcu_seq_start(&rcu_state.gp_seq_polled);
|
|
|
|
|
|
|
|
|
|
// Either way, record current state.
|
|
|
|
|
*snap = rcu_state.gp_seq_polled;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// Make the polled API aware of the end of a grace period.
|
|
|
|
|
static void rcu_poll_gp_seq_end(unsigned long *snap)
|
|
|
|
|
{
|
|
|
|
|
struct rcu_node *rnp = rcu_get_root();
|
|
|
|
|
|
2022-12-16 01:26:09 +08:00
|
|
|
|
if (rcu_scheduler_active != RCU_SCHEDULER_INACTIVE)
|
2022-04-14 08:46:15 +08:00
|
|
|
|
raw_lockdep_assert_held_rcu_node(rnp);
|
|
|
|
|
|
|
|
|
|
// If the previously noted GP is still in effect, record the
|
|
|
|
|
// end of that GP. Either way, zero counter to avoid counter-wrap
|
|
|
|
|
// problems.
|
|
|
|
|
if (*snap && *snap == rcu_state.gp_seq_polled) {
|
|
|
|
|
rcu_seq_end(&rcu_state.gp_seq_polled);
|
|
|
|
|
rcu_state.gp_seq_polled_snap = 0;
|
2022-04-14 21:56:35 +08:00
|
|
|
|
rcu_state.gp_seq_polled_exp_snap = 0;
|
2022-04-14 08:46:15 +08:00
|
|
|
|
} else {
|
|
|
|
|
*snap = 0;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// Make the polled API aware of the beginning of a grace period, but
|
|
|
|
|
// where caller does not hold the root rcu_node structure's lock.
|
|
|
|
|
static void rcu_poll_gp_seq_start_unlocked(unsigned long *snap)
|
|
|
|
|
{
|
2022-10-21 01:58:14 +08:00
|
|
|
|
unsigned long flags;
|
2022-04-14 08:46:15 +08:00
|
|
|
|
struct rcu_node *rnp = rcu_get_root();
|
|
|
|
|
|
|
|
|
|
if (rcu_init_invoked()) {
|
2022-12-16 01:26:09 +08:00
|
|
|
|
if (rcu_scheduler_active != RCU_SCHEDULER_INACTIVE)
|
|
|
|
|
lockdep_assert_irqs_enabled();
|
2022-10-21 01:58:14 +08:00
|
|
|
|
raw_spin_lock_irqsave_rcu_node(rnp, flags);
|
2022-04-14 08:46:15 +08:00
|
|
|
|
}
|
|
|
|
|
rcu_poll_gp_seq_start(snap);
|
|
|
|
|
if (rcu_init_invoked())
|
2022-10-21 01:58:14 +08:00
|
|
|
|
raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
|
2022-04-14 08:46:15 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// Make the polled API aware of the end of a grace period, but where
|
|
|
|
|
// caller does not hold the root rcu_node structure's lock.
|
|
|
|
|
static void rcu_poll_gp_seq_end_unlocked(unsigned long *snap)
|
|
|
|
|
{
|
2022-10-21 01:58:14 +08:00
|
|
|
|
unsigned long flags;
|
2022-04-14 08:46:15 +08:00
|
|
|
|
struct rcu_node *rnp = rcu_get_root();
|
|
|
|
|
|
|
|
|
|
if (rcu_init_invoked()) {
|
2022-12-16 01:26:09 +08:00
|
|
|
|
if (rcu_scheduler_active != RCU_SCHEDULER_INACTIVE)
|
|
|
|
|
lockdep_assert_irqs_enabled();
|
2022-10-21 01:58:14 +08:00
|
|
|
|
raw_spin_lock_irqsave_rcu_node(rnp, flags);
|
2022-04-14 08:46:15 +08:00
|
|
|
|
}
|
|
|
|
|
rcu_poll_gp_seq_end(snap);
|
|
|
|
|
if (rcu_init_invoked())
|
2022-10-21 01:58:14 +08:00
|
|
|
|
raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
|
2022-04-14 08:46:15 +08:00
|
|
|
|
}
|
|
|
|
|
|
2012-06-19 09:36:08 +08:00
|
|
|
|
/*
|
2015-11-08 15:35:00 +08:00
|
|
|
|
* Initialize a new grace period. Return false if no grace period required.
|
2012-06-19 09:36:08 +08:00
|
|
|
|
*/
|
2021-06-08 12:57:02 +08:00
|
|
|
|
static noinline_for_stack bool rcu_gp_init(void)
|
2012-06-19 09:36:08 +08:00
|
|
|
|
{
|
rcu: Fix grace-period hangs from mid-init task resume
Without special fail-safe quiescent-state-propagation checks, grace-period
hangs can result from the following scenario:
1. A task running on a given CPU is preempted in its RCU read-side
critical section.
2. That CPU goes offline, and there are now no online CPUs
corresponding to that CPU's leaf rcu_node structure.
3. The rcu_gp_init() function does the first phase of grace-period
initialization, and sets the aforementioned leaf rcu_node
structure's ->qsmaskinit field to all zeroes. Because there
is a blocked task, it does not propagate the zeroing of either
->qsmaskinit or ->qsmaskinitnext up the rcu_node tree.
4. The task resumes on some other CPU and exits its critical section.
There is no grace period in progress, so the resulting quiescent
state is not reported up the tree.
5. The rcu_gp_init() function does the second phase of grace-period
initialization, which results in the leaf rcu_node structure
being initialized to expect no further quiescent states, but
with that structure's parent expecting a quiescent-state report.
The parent will never receive a quiescent state from this leaf
rcu_node structure, so the grace period will hang, resulting in
RCU CPU stall warnings.
It would be good to get rid of the special fail-safe quiescent-state
propagation checks. This commit therefore checks the leaf rcu_node
structure's ->wait_blkd_tasks field during grace-period initialization.
If this flag is set, the rcu_report_qs_rnp() is invoked to immediately
report the possible quiescent state. While in the neighborhood, this
commit also report quiescent states for any CPUs that went offline between
the two phases of grace-period initialization, thus reducing grace-period
delays and hopefully eventually allowing removal of offline-CPU checks
from the force-quiescent-state code path.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-05-08 00:34:17 +08:00
|
|
|
|
unsigned long flags;
|
rcu: Process offlining and onlining only at grace-period start
Races between CPU hotplug and grace periods can be difficult to resolve,
so the ->onoff_mutex is used to exclude the two events. Unfortunately,
this means that it is impossible for an outgoing CPU to perform the
last bits of its offlining from its last pass through the idle loop,
because sleeplocks cannot be acquired in that context.
This commit avoids these problems by buffering online and offline events
in a new ->qsmaskinitnext field in the leaf rcu_node structures. When a
grace period starts, the events accumulated in this mask are applied to
the ->qsmaskinit field, and, if needed, up the rcu_node tree. The special
case of all CPUs corresponding to a given leaf rcu_node structure being
offline while there are still elements in that structure's ->blkd_tasks
list is handled using a new ->wait_blkd_tasks field. In this case,
propagating the offline bits up the tree is deferred until the beginning
of the grace period after all of the tasks have exited their RCU read-side
critical sections and removed themselves from the list, at which point
the ->wait_blkd_tasks flag is cleared. If one of that leaf rcu_node
structure's CPUs comes back online before the list empties, then the
->wait_blkd_tasks flag is simply cleared.
This of course means that RCU's notion of which CPUs are offline can be
out of date. This is OK because RCU need only wait on CPUs that were
online at the time that the grace period started. In addition, RCU's
force-quiescent-state actions will handle the case where a CPU goes
offline after the grace period starts.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2015-01-24 13:52:37 +08:00
|
|
|
|
unsigned long oldmask;
|
rcu: Fix grace-period hangs from mid-init task resume
Without special fail-safe quiescent-state-propagation checks, grace-period
hangs can result from the following scenario:
1. A task running on a given CPU is preempted in its RCU read-side
critical section.
2. That CPU goes offline, and there are now no online CPUs
corresponding to that CPU's leaf rcu_node structure.
3. The rcu_gp_init() function does the first phase of grace-period
initialization, and sets the aforementioned leaf rcu_node
structure's ->qsmaskinit field to all zeroes. Because there
is a blocked task, it does not propagate the zeroing of either
->qsmaskinit or ->qsmaskinitnext up the rcu_node tree.
4. The task resumes on some other CPU and exits its critical section.
There is no grace period in progress, so the resulting quiescent
state is not reported up the tree.
5. The rcu_gp_init() function does the second phase of grace-period
initialization, which results in the leaf rcu_node structure
being initialized to expect no further quiescent states, but
with that structure's parent expecting a quiescent-state report.
The parent will never receive a quiescent state from this leaf
rcu_node structure, so the grace period will hang, resulting in
RCU CPU stall warnings.
It would be good to get rid of the special fail-safe quiescent-state
propagation checks. This commit therefore checks the leaf rcu_node
structure's ->wait_blkd_tasks field during grace-period initialization.
If this flag is set, the rcu_report_qs_rnp() is invoked to immediately
report the possible quiescent state. While in the neighborhood, this
commit also report quiescent states for any CPUs that went offline between
the two phases of grace-period initialization, thus reducing grace-period
delays and hopefully eventually allowing removal of offline-CPU checks
from the force-quiescent-state code path.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-05-08 00:34:17 +08:00
|
|
|
|
unsigned long mask;
|
2012-06-19 09:36:08 +08:00
|
|
|
|
struct rcu_data *rdp;
|
2018-07-04 08:22:34 +08:00
|
|
|
|
struct rcu_node *rnp = rcu_get_root();
|
2012-06-19 09:36:08 +08:00
|
|
|
|
|
2018-07-06 06:47:01 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_activity, jiffies);
|
2015-10-08 18:24:23 +08:00
|
|
|
|
raw_spin_lock_irq_rcu_node(rnp);
|
2018-07-06 06:47:01 +08:00
|
|
|
|
if (!READ_ONCE(rcu_state.gp_flags)) {
|
2013-08-09 09:27:52 +08:00
|
|
|
|
/* Spurious wakeup, tell caller to go back to sleep. */
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_unlock_irq_rcu_node(rnp);
|
2015-11-08 15:35:00 +08:00
|
|
|
|
return false;
|
2013-08-09 09:27:52 +08:00
|
|
|
|
}
|
2018-07-06 06:47:01 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_flags, 0); /* Clear all flags: New GP. */
|
2012-06-19 09:36:08 +08:00
|
|
|
|
|
2018-07-04 08:22:34 +08:00
|
|
|
|
if (WARN_ON_ONCE(rcu_gp_in_progress())) {
|
2013-08-09 09:27:52 +08:00
|
|
|
|
/*
|
|
|
|
|
* Grace period already in progress, don't start another.
|
|
|
|
|
* Not supposed to be able to happen.
|
|
|
|
|
*/
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_unlock_irq_rcu_node(rnp);
|
2015-11-08 15:35:00 +08:00
|
|
|
|
return false;
|
2012-06-23 02:08:41 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Advance to a new grace period and initialize state. */
|
2018-07-04 08:22:34 +08:00
|
|
|
|
record_gp_stall_check_time();
|
2018-05-02 05:34:08 +08:00
|
|
|
|
/* Record GP times before starting GP, hence rcu_seq_start(). */
|
2018-07-06 06:47:01 +08:00
|
|
|
|
rcu_seq_start(&rcu_state.gp_seq);
|
2020-03-22 10:52:20 +08:00
|
|
|
|
ASSERT_EXCLUSIVE_WRITER(rcu_state.gp_seq);
|
2018-07-06 06:47:01 +08:00
|
|
|
|
trace_rcu_grace_period(rcu_state.name, rcu_state.gp_seq, TPS("start"));
|
2022-04-14 08:46:15 +08:00
|
|
|
|
rcu_poll_gp_seq_start(&rcu_state.gp_seq_polled_snap);
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_unlock_irq_rcu_node(rnp);
|
2012-06-23 02:08:41 +08:00
|
|
|
|
|
rcu: Process offlining and onlining only at grace-period start
Races between CPU hotplug and grace periods can be difficult to resolve,
so the ->onoff_mutex is used to exclude the two events. Unfortunately,
this means that it is impossible for an outgoing CPU to perform the
last bits of its offlining from its last pass through the idle loop,
because sleeplocks cannot be acquired in that context.
This commit avoids these problems by buffering online and offline events
in a new ->qsmaskinitnext field in the leaf rcu_node structures. When a
grace period starts, the events accumulated in this mask are applied to
the ->qsmaskinit field, and, if needed, up the rcu_node tree. The special
case of all CPUs corresponding to a given leaf rcu_node structure being
offline while there are still elements in that structure's ->blkd_tasks
list is handled using a new ->wait_blkd_tasks field. In this case,
propagating the offline bits up the tree is deferred until the beginning
of the grace period after all of the tasks have exited their RCU read-side
critical sections and removed themselves from the list, at which point
the ->wait_blkd_tasks flag is cleared. If one of that leaf rcu_node
structure's CPUs comes back online before the list empties, then the
->wait_blkd_tasks flag is simply cleared.
This of course means that RCU's notion of which CPUs are offline can be
out of date. This is OK because RCU need only wait on CPUs that were
online at the time that the grace period started. In addition, RCU's
force-quiescent-state actions will handle the case where a CPU goes
offline after the grace period starts.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2015-01-24 13:52:37 +08:00
|
|
|
|
/*
|
2020-08-08 01:07:19 +08:00
|
|
|
|
* Apply per-leaf buffered online and offline operations to
|
|
|
|
|
* the rcu_node tree. Note that this new grace period need not
|
|
|
|
|
* wait for subsequent online CPUs, and that RCU hooks in the CPU
|
|
|
|
|
* offlining path, when combined with checks in this function,
|
|
|
|
|
* will handle CPUs that are currently going offline or that will
|
|
|
|
|
* go offline later. Please also refer to "Hotplug CPU" section
|
|
|
|
|
* of RCU's Requirements documentation.
|
rcu: Process offlining and onlining only at grace-period start
Races between CPU hotplug and grace periods can be difficult to resolve,
so the ->onoff_mutex is used to exclude the two events. Unfortunately,
this means that it is impossible for an outgoing CPU to perform the
last bits of its offlining from its last pass through the idle loop,
because sleeplocks cannot be acquired in that context.
This commit avoids these problems by buffering online and offline events
in a new ->qsmaskinitnext field in the leaf rcu_node structures. When a
grace period starts, the events accumulated in this mask are applied to
the ->qsmaskinit field, and, if needed, up the rcu_node tree. The special
case of all CPUs corresponding to a given leaf rcu_node structure being
offline while there are still elements in that structure's ->blkd_tasks
list is handled using a new ->wait_blkd_tasks field. In this case,
propagating the offline bits up the tree is deferred until the beginning
of the grace period after all of the tasks have exited their RCU read-side
critical sections and removed themselves from the list, at which point
the ->wait_blkd_tasks flag is cleared. If one of that leaf rcu_node
structure's CPUs comes back online before the list empties, then the
->wait_blkd_tasks flag is simply cleared.
This of course means that RCU's notion of which CPUs are offline can be
out of date. This is OK because RCU need only wait on CPUs that were
online at the time that the grace period started. In addition, RCU's
force-quiescent-state actions will handle the case where a CPU goes
offline after the grace period starts.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2015-01-24 13:52:37 +08:00
|
|
|
|
*/
|
rcu: Check and report missed fqs timer wakeup on RCU stall
For a new grace period request, the RCU GP kthread transitions through
following states:
a. [RCU_GP_WAIT_GPS] -> [RCU_GP_DONE_GPS]
The RCU_GP_WAIT_GPS state is where the GP kthread waits for a request
for a new GP. Once it receives a request (for example, when a new RCU
callback is queued), the GP kthread transitions to RCU_GP_DONE_GPS.
b. [RCU_GP_DONE_GPS] -> [RCU_GP_ONOFF]
Grace period initialization starts in rcu_gp_init(), which records the
start of new GP in rcu_state.gp_seq and transitions to RCU_GP_ONOFF.
c. [RCU_GP_ONOFF] -> [RCU_GP_INIT]
The purpose of the RCU_GP_ONOFF state is to apply the online/offline
information that was buffered for any CPUs that recently came online or
went offline. This state is maintained in per-leaf rcu_node bitmasks,
with the buffered state in ->qsmaskinitnext and the state for the upcoming
GP in ->qsmaskinit. At the end of this RCU_GP_ONOFF state, each bit in
->qsmaskinit will correspond to a CPU that must pass through a quiescent
state before the upcoming grace period is allowed to complete.
However, a leaf rcu_node structure with an all-zeroes ->qsmaskinit
cannot necessarily be ignored. In preemptible RCU, there might well be
tasks still in RCU read-side critical sections that were first preempted
while running on one of the CPUs managed by this structure. Such tasks
will be queued on this structure's ->blkd_tasks list. Only after this
list fully drains can this leaf rcu_node structure be ignored, and even
then only if none of its CPUs have come back online in the meantime.
Once that happens, the ->qsmaskinit masks further up the tree will be
updated to exclude this leaf rcu_node structure.
Once the ->qsmaskinitnext and ->qsmaskinit fields have been updated
as needed, the GP kthread transitions to RCU_GP_INIT.
d. [RCU_GP_INIT] -> [RCU_GP_WAIT_FQS]
The purpose of the RCU_GP_INIT state is to copy each ->qsmaskinit to
the ->qsmask field within each rcu_node structure. This copying is done
breadth-first from the root to the leaves. Why not just copy directly
from ->qsmaskinitnext to ->qsmask? Because the ->qsmaskinitnext masks
can change in the meantime as additional CPUs come online or go offline.
Such changes would result in inconsistencies in the ->qsmask fields up and
down the tree, which could in turn result in too-short grace periods or
grace-period hangs. These issues are avoided by snapshotting the leaf
rcu_node structures' ->qsmaskinitnext fields into their ->qsmaskinit
counterparts, generating a consistent set of ->qsmaskinit fields
throughout the tree, and only then copying these consistent ->qsmaskinit
fields to their ->qsmask counterparts.
Once this initialization step is complete, the GP kthread transitions
to RCU_GP_WAIT_FQS, where it waits to do a force-quiescent-state scan
on the one hand or for the end of the grace period on the other.
e. [RCU_GP_WAIT_FQS] -> [RCU_GP_DOING_FQS]
The RCU_GP_WAIT_FQS state waits for one of three things: (1) An
explicit request to do a force-quiescent-state scan, (2) The end of
the grace period, or (3) A short interval of time, after which it
will do a force-quiescent-state (FQS) scan. The explicit request can
come from rcutorture or from any CPU that has too many RCU callbacks
queued (see the qhimark kernel parameter and the RCU_GP_FLAG_OVLD
flag). The aforementioned "short period of time" is specified by the
jiffies_till_first_fqs boot parameter for a given grace period's first
FQS scan and by the jiffies_till_next_fqs for later FQS scans.
Either way, once the wait is over, the GP kthread transitions to
RCU_GP_DOING_FQS.
f. [RCU_GP_DOING_FQS] -> [RCU_GP_CLEANUP]
The RCU_GP_DOING_FQS state performs an FQS scan. Each such scan carries
out two functions for any CPU whose bit is still set in its leaf rcu_node
structure's ->qsmask field, that is, for any CPU that has not yet reported
a quiescent state for the current grace period:
i. Report quiescent states on behalf of CPUs that have been observed
to be idle (from an RCU perspective) since the beginning of the
grace period.
ii. If the current grace period is too old, take various actions to
encourage holdout CPUs to pass through quiescent states, including
enlisting the aid of any calls to cond_resched() and might_sleep(),
and even including IPIing the holdout CPUs.
These checks are skipped for any leaf rcu_node structure with a all-zero
->qsmask field, however such structures are subject to RCU priority
boosting if there are tasks on a given structure blocking the current
grace period. The end of the grace period is detected when the root
rcu_node structure's ->qsmask is zero and when there are no longer any
preempted tasks blocking the current grace period. (No, this last check
is not redundant. To see this, consider an rcu_node tree having exactly
one structure that serves as both root and leaf.)
Once the end of the grace period is detected, the GP kthread transitions
to RCU_GP_CLEANUP.
g. [RCU_GP_CLEANUP] -> [RCU_GP_CLEANED]
The RCU_GP_CLEANUP state marks the end of grace period by updating the
rcu_state structure's ->gp_seq field and also all rcu_node structures'
->gp_seq field. As before, the rcu_node tree is traversed in breadth
first order. Once this update is complete, the GP kthread transitions
to the RCU_GP_CLEANED state.
i. [RCU_GP_CLEANED] -> [RCU_GP_INIT]
Once in the RCU_GP_CLEANED state, the GP kthread immediately transitions
into the RCU_GP_INIT state.
j. The role of timers.
If there is at least one idle CPU, and if timers are not firing, the
transition from RCU_GP_DOING_FQS to RCU_GP_CLEANUP will never happen.
Timers can fail to fire for a number of reasons, including issues in
timer configuration, issues in the timer framework, and failure to handle
softirqs (for example, when there is a storm of interrupts). Whatever the
reason, if the timers fail to fire, the GP kthread will never be awakened,
resulting in RCU CPU stall warnings and eventually in OOM.
However, an RCU CPU stall warning has a large number of potential causes,
as documented in Documentation/RCU/stallwarn.rst. This commit therefore
adds analysis to the RCU CPU stall-warning code to emit an additional
message if the cause of the stall is likely to be timer failure.
Signed-off-by: Neeraj Upadhyay <neeraju@codeaurora.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-11-17 00:06:00 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_state, RCU_GP_ONOFF);
|
2021-02-16 23:04:34 +08:00
|
|
|
|
/* Exclude CPU hotplug operations. */
|
2018-07-05 05:33:59 +08:00
|
|
|
|
rcu_for_each_leaf_node(rnp) {
|
2021-02-16 23:04:34 +08:00
|
|
|
|
local_irq_save(flags);
|
|
|
|
|
arch_spin_lock(&rcu_state.ofl_lock);
|
|
|
|
|
raw_spin_lock_rcu_node(rnp);
|
rcu: Process offlining and onlining only at grace-period start
Races between CPU hotplug and grace periods can be difficult to resolve,
so the ->onoff_mutex is used to exclude the two events. Unfortunately,
this means that it is impossible for an outgoing CPU to perform the
last bits of its offlining from its last pass through the idle loop,
because sleeplocks cannot be acquired in that context.
This commit avoids these problems by buffering online and offline events
in a new ->qsmaskinitnext field in the leaf rcu_node structures. When a
grace period starts, the events accumulated in this mask are applied to
the ->qsmaskinit field, and, if needed, up the rcu_node tree. The special
case of all CPUs corresponding to a given leaf rcu_node structure being
offline while there are still elements in that structure's ->blkd_tasks
list is handled using a new ->wait_blkd_tasks field. In this case,
propagating the offline bits up the tree is deferred until the beginning
of the grace period after all of the tasks have exited their RCU read-side
critical sections and removed themselves from the list, at which point
the ->wait_blkd_tasks flag is cleared. If one of that leaf rcu_node
structure's CPUs comes back online before the list empties, then the
->wait_blkd_tasks flag is simply cleared.
This of course means that RCU's notion of which CPUs are offline can be
out of date. This is OK because RCU need only wait on CPUs that were
online at the time that the grace period started. In addition, RCU's
force-quiescent-state actions will handle the case where a CPU goes
offline after the grace period starts.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2015-01-24 13:52:37 +08:00
|
|
|
|
if (rnp->qsmaskinit == rnp->qsmaskinitnext &&
|
|
|
|
|
!rnp->wait_blkd_tasks) {
|
|
|
|
|
/* Nothing to do on this leaf rcu_node structure. */
|
2021-02-16 23:04:34 +08:00
|
|
|
|
raw_spin_unlock_rcu_node(rnp);
|
|
|
|
|
arch_spin_unlock(&rcu_state.ofl_lock);
|
|
|
|
|
local_irq_restore(flags);
|
rcu: Process offlining and onlining only at grace-period start
Races between CPU hotplug and grace periods can be difficult to resolve,
so the ->onoff_mutex is used to exclude the two events. Unfortunately,
this means that it is impossible for an outgoing CPU to perform the
last bits of its offlining from its last pass through the idle loop,
because sleeplocks cannot be acquired in that context.
This commit avoids these problems by buffering online and offline events
in a new ->qsmaskinitnext field in the leaf rcu_node structures. When a
grace period starts, the events accumulated in this mask are applied to
the ->qsmaskinit field, and, if needed, up the rcu_node tree. The special
case of all CPUs corresponding to a given leaf rcu_node structure being
offline while there are still elements in that structure's ->blkd_tasks
list is handled using a new ->wait_blkd_tasks field. In this case,
propagating the offline bits up the tree is deferred until the beginning
of the grace period after all of the tasks have exited their RCU read-side
critical sections and removed themselves from the list, at which point
the ->wait_blkd_tasks flag is cleared. If one of that leaf rcu_node
structure's CPUs comes back online before the list empties, then the
->wait_blkd_tasks flag is simply cleared.
This of course means that RCU's notion of which CPUs are offline can be
out of date. This is OK because RCU need only wait on CPUs that were
online at the time that the grace period started. In addition, RCU's
force-quiescent-state actions will handle the case where a CPU goes
offline after the grace period starts.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2015-01-24 13:52:37 +08:00
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Record old state, apply changes to ->qsmaskinit field. */
|
|
|
|
|
oldmask = rnp->qsmaskinit;
|
|
|
|
|
rnp->qsmaskinit = rnp->qsmaskinitnext;
|
|
|
|
|
|
|
|
|
|
/* If zero-ness of ->qsmaskinit changed, propagate up tree. */
|
|
|
|
|
if (!oldmask != !rnp->qsmaskinit) {
|
rcu: Clean up handling of tasks blocked across full-rcu_node offline
Commit 0aa04b055e71 ("rcu: Process offlining and onlining only at
grace-period start") deferred handling of CPU-hotplug events until the
start of the next grace period, but consider the following sequence
of events:
1. A task is preempted within an RCU-preempt read-side critical
section.
2. The CPU that this task was running on goes offline, along with all
other CPUs sharing the corresponding leaf rcu_node structure.
3. The task resumes execution.
4. One of those CPUs comes back online before a new grace period starts.
In step 2, the code in the next rcu_gp_init() invocation will (correctly)
defer removing the leaf rcu_node structure from the upper-level bitmasks,
and will (correctly) set that structure's ->wait_blkd_tasks field. During
the ensuing interval, RCU will (correctly) track the tasks preempted on
that structure because they must block any subsequent grace period.
In step 3, the code in rcu_read_unlock_special() will (correctly) remove
the task from the leaf rcu_node structure. From this point forward, RCU
need not pay attention to this structure, at least not until one of the
corresponding CPUs comes back online.
In step 4, the code in the next rcu_gp_init() invocation will
(incorrectly) invoke rcu_init_new_rnp(). This is incorrect because
the corresponding rcu_cleanup_dead_rnp() was never invoked. This is
nevertheless harmless because the upper-level bits are still set.
So, no harm, no foul, right?
At least, all is well until a little further into rcu_gp_init()
invocation, which will notice that there are no longer any tasks blocked
on the leaf rcu_node structure, conclude that there is no longer anything
left over from step 2's offline operation, and will therefore invoke
rcu_cleanup_dead_rnp(). But this invocation of rcu_cleanup_dead_rnp()
is for the beginning of the earlier offline interval, and the previous
invocation of rcu_init_new_rnp() is for the end of that same interval.
That is right, they are invoked out of order.
That cannot be good, can it?
It turns out that this is not a (correctness!) problem because
rcu_cleanup_dead_rnp() checks to see if any of the corresponding CPUs
are online, and refuses to do anything if so. In other words, in the
case where rcu_init_new_rnp() and rcu_cleanup_dead_rnp() execute out of
order, they both have no effect.
But this is at best an accident waiting to happen.
This commit therefore adds logic to rcu_gp_init() so that
rcu_init_new_rnp() and rcu_cleanup_dead_rnp() are always invoked in
order, and so that neither are invoked at all in cases where RCU had to
pay attention to the leaf rcu_node structure during the entire time that
all corresponding CPUs were offline.
And, while in the area, this commit reduces confusion by using formal
parameters rather than local variables that just happen to have the same
value at that particular point in the code.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-05-03 03:49:21 +08:00
|
|
|
|
if (!oldmask) { /* First online CPU for rcu_node. */
|
|
|
|
|
if (!rnp->wait_blkd_tasks) /* Ever offline? */
|
|
|
|
|
rcu_init_new_rnp(rnp);
|
|
|
|
|
} else if (rcu_preempt_has_tasks(rnp)) {
|
|
|
|
|
rnp->wait_blkd_tasks = true; /* blocked tasks */
|
|
|
|
|
} else { /* Last offline CPU and can propagate. */
|
rcu: Process offlining and onlining only at grace-period start
Races between CPU hotplug and grace periods can be difficult to resolve,
so the ->onoff_mutex is used to exclude the two events. Unfortunately,
this means that it is impossible for an outgoing CPU to perform the
last bits of its offlining from its last pass through the idle loop,
because sleeplocks cannot be acquired in that context.
This commit avoids these problems by buffering online and offline events
in a new ->qsmaskinitnext field in the leaf rcu_node structures. When a
grace period starts, the events accumulated in this mask are applied to
the ->qsmaskinit field, and, if needed, up the rcu_node tree. The special
case of all CPUs corresponding to a given leaf rcu_node structure being
offline while there are still elements in that structure's ->blkd_tasks
list is handled using a new ->wait_blkd_tasks field. In this case,
propagating the offline bits up the tree is deferred until the beginning
of the grace period after all of the tasks have exited their RCU read-side
critical sections and removed themselves from the list, at which point
the ->wait_blkd_tasks flag is cleared. If one of that leaf rcu_node
structure's CPUs comes back online before the list empties, then the
->wait_blkd_tasks flag is simply cleared.
This of course means that RCU's notion of which CPUs are offline can be
out of date. This is OK because RCU need only wait on CPUs that were
online at the time that the grace period started. In addition, RCU's
force-quiescent-state actions will handle the case where a CPU goes
offline after the grace period starts.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2015-01-24 13:52:37 +08:00
|
|
|
|
rcu_cleanup_dead_rnp(rnp);
|
rcu: Clean up handling of tasks blocked across full-rcu_node offline
Commit 0aa04b055e71 ("rcu: Process offlining and onlining only at
grace-period start") deferred handling of CPU-hotplug events until the
start of the next grace period, but consider the following sequence
of events:
1. A task is preempted within an RCU-preempt read-side critical
section.
2. The CPU that this task was running on goes offline, along with all
other CPUs sharing the corresponding leaf rcu_node structure.
3. The task resumes execution.
4. One of those CPUs comes back online before a new grace period starts.
In step 2, the code in the next rcu_gp_init() invocation will (correctly)
defer removing the leaf rcu_node structure from the upper-level bitmasks,
and will (correctly) set that structure's ->wait_blkd_tasks field. During
the ensuing interval, RCU will (correctly) track the tasks preempted on
that structure because they must block any subsequent grace period.
In step 3, the code in rcu_read_unlock_special() will (correctly) remove
the task from the leaf rcu_node structure. From this point forward, RCU
need not pay attention to this structure, at least not until one of the
corresponding CPUs comes back online.
In step 4, the code in the next rcu_gp_init() invocation will
(incorrectly) invoke rcu_init_new_rnp(). This is incorrect because
the corresponding rcu_cleanup_dead_rnp() was never invoked. This is
nevertheless harmless because the upper-level bits are still set.
So, no harm, no foul, right?
At least, all is well until a little further into rcu_gp_init()
invocation, which will notice that there are no longer any tasks blocked
on the leaf rcu_node structure, conclude that there is no longer anything
left over from step 2's offline operation, and will therefore invoke
rcu_cleanup_dead_rnp(). But this invocation of rcu_cleanup_dead_rnp()
is for the beginning of the earlier offline interval, and the previous
invocation of rcu_init_new_rnp() is for the end of that same interval.
That is right, they are invoked out of order.
That cannot be good, can it?
It turns out that this is not a (correctness!) problem because
rcu_cleanup_dead_rnp() checks to see if any of the corresponding CPUs
are online, and refuses to do anything if so. In other words, in the
case where rcu_init_new_rnp() and rcu_cleanup_dead_rnp() execute out of
order, they both have no effect.
But this is at best an accident waiting to happen.
This commit therefore adds logic to rcu_gp_init() so that
rcu_init_new_rnp() and rcu_cleanup_dead_rnp() are always invoked in
order, and so that neither are invoked at all in cases where RCU had to
pay attention to the leaf rcu_node structure during the entire time that
all corresponding CPUs were offline.
And, while in the area, this commit reduces confusion by using formal
parameters rather than local variables that just happen to have the same
value at that particular point in the code.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-05-03 03:49:21 +08:00
|
|
|
|
}
|
rcu: Process offlining and onlining only at grace-period start
Races between CPU hotplug and grace periods can be difficult to resolve,
so the ->onoff_mutex is used to exclude the two events. Unfortunately,
this means that it is impossible for an outgoing CPU to perform the
last bits of its offlining from its last pass through the idle loop,
because sleeplocks cannot be acquired in that context.
This commit avoids these problems by buffering online and offline events
in a new ->qsmaskinitnext field in the leaf rcu_node structures. When a
grace period starts, the events accumulated in this mask are applied to
the ->qsmaskinit field, and, if needed, up the rcu_node tree. The special
case of all CPUs corresponding to a given leaf rcu_node structure being
offline while there are still elements in that structure's ->blkd_tasks
list is handled using a new ->wait_blkd_tasks field. In this case,
propagating the offline bits up the tree is deferred until the beginning
of the grace period after all of the tasks have exited their RCU read-side
critical sections and removed themselves from the list, at which point
the ->wait_blkd_tasks flag is cleared. If one of that leaf rcu_node
structure's CPUs comes back online before the list empties, then the
->wait_blkd_tasks flag is simply cleared.
This of course means that RCU's notion of which CPUs are offline can be
out of date. This is OK because RCU need only wait on CPUs that were
online at the time that the grace period started. In addition, RCU's
force-quiescent-state actions will handle the case where a CPU goes
offline after the grace period starts.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2015-01-24 13:52:37 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* If all waited-on tasks from prior grace period are
|
|
|
|
|
* done, and if all this rcu_node structure's CPUs are
|
|
|
|
|
* still offline, propagate up the rcu_node tree and
|
|
|
|
|
* clear ->wait_blkd_tasks. Otherwise, if one of this
|
|
|
|
|
* rcu_node structure's CPUs has since come back online,
|
rcu: Clean up handling of tasks blocked across full-rcu_node offline
Commit 0aa04b055e71 ("rcu: Process offlining and onlining only at
grace-period start") deferred handling of CPU-hotplug events until the
start of the next grace period, but consider the following sequence
of events:
1. A task is preempted within an RCU-preempt read-side critical
section.
2. The CPU that this task was running on goes offline, along with all
other CPUs sharing the corresponding leaf rcu_node structure.
3. The task resumes execution.
4. One of those CPUs comes back online before a new grace period starts.
In step 2, the code in the next rcu_gp_init() invocation will (correctly)
defer removing the leaf rcu_node structure from the upper-level bitmasks,
and will (correctly) set that structure's ->wait_blkd_tasks field. During
the ensuing interval, RCU will (correctly) track the tasks preempted on
that structure because they must block any subsequent grace period.
In step 3, the code in rcu_read_unlock_special() will (correctly) remove
the task from the leaf rcu_node structure. From this point forward, RCU
need not pay attention to this structure, at least not until one of the
corresponding CPUs comes back online.
In step 4, the code in the next rcu_gp_init() invocation will
(incorrectly) invoke rcu_init_new_rnp(). This is incorrect because
the corresponding rcu_cleanup_dead_rnp() was never invoked. This is
nevertheless harmless because the upper-level bits are still set.
So, no harm, no foul, right?
At least, all is well until a little further into rcu_gp_init()
invocation, which will notice that there are no longer any tasks blocked
on the leaf rcu_node structure, conclude that there is no longer anything
left over from step 2's offline operation, and will therefore invoke
rcu_cleanup_dead_rnp(). But this invocation of rcu_cleanup_dead_rnp()
is for the beginning of the earlier offline interval, and the previous
invocation of rcu_init_new_rnp() is for the end of that same interval.
That is right, they are invoked out of order.
That cannot be good, can it?
It turns out that this is not a (correctness!) problem because
rcu_cleanup_dead_rnp() checks to see if any of the corresponding CPUs
are online, and refuses to do anything if so. In other words, in the
case where rcu_init_new_rnp() and rcu_cleanup_dead_rnp() execute out of
order, they both have no effect.
But this is at best an accident waiting to happen.
This commit therefore adds logic to rcu_gp_init() so that
rcu_init_new_rnp() and rcu_cleanup_dead_rnp() are always invoked in
order, and so that neither are invoked at all in cases where RCU had to
pay attention to the leaf rcu_node structure during the entire time that
all corresponding CPUs were offline.
And, while in the area, this commit reduces confusion by using formal
parameters rather than local variables that just happen to have the same
value at that particular point in the code.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-05-03 03:49:21 +08:00
|
|
|
|
* simply clear ->wait_blkd_tasks.
|
rcu: Process offlining and onlining only at grace-period start
Races between CPU hotplug and grace periods can be difficult to resolve,
so the ->onoff_mutex is used to exclude the two events. Unfortunately,
this means that it is impossible for an outgoing CPU to perform the
last bits of its offlining from its last pass through the idle loop,
because sleeplocks cannot be acquired in that context.
This commit avoids these problems by buffering online and offline events
in a new ->qsmaskinitnext field in the leaf rcu_node structures. When a
grace period starts, the events accumulated in this mask are applied to
the ->qsmaskinit field, and, if needed, up the rcu_node tree. The special
case of all CPUs corresponding to a given leaf rcu_node structure being
offline while there are still elements in that structure's ->blkd_tasks
list is handled using a new ->wait_blkd_tasks field. In this case,
propagating the offline bits up the tree is deferred until the beginning
of the grace period after all of the tasks have exited their RCU read-side
critical sections and removed themselves from the list, at which point
the ->wait_blkd_tasks flag is cleared. If one of that leaf rcu_node
structure's CPUs comes back online before the list empties, then the
->wait_blkd_tasks flag is simply cleared.
This of course means that RCU's notion of which CPUs are offline can be
out of date. This is OK because RCU need only wait on CPUs that were
online at the time that the grace period started. In addition, RCU's
force-quiescent-state actions will handle the case where a CPU goes
offline after the grace period starts.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2015-01-24 13:52:37 +08:00
|
|
|
|
*/
|
|
|
|
|
if (rnp->wait_blkd_tasks &&
|
rcu: Clean up handling of tasks blocked across full-rcu_node offline
Commit 0aa04b055e71 ("rcu: Process offlining and onlining only at
grace-period start") deferred handling of CPU-hotplug events until the
start of the next grace period, but consider the following sequence
of events:
1. A task is preempted within an RCU-preempt read-side critical
section.
2. The CPU that this task was running on goes offline, along with all
other CPUs sharing the corresponding leaf rcu_node structure.
3. The task resumes execution.
4. One of those CPUs comes back online before a new grace period starts.
In step 2, the code in the next rcu_gp_init() invocation will (correctly)
defer removing the leaf rcu_node structure from the upper-level bitmasks,
and will (correctly) set that structure's ->wait_blkd_tasks field. During
the ensuing interval, RCU will (correctly) track the tasks preempted on
that structure because they must block any subsequent grace period.
In step 3, the code in rcu_read_unlock_special() will (correctly) remove
the task from the leaf rcu_node structure. From this point forward, RCU
need not pay attention to this structure, at least not until one of the
corresponding CPUs comes back online.
In step 4, the code in the next rcu_gp_init() invocation will
(incorrectly) invoke rcu_init_new_rnp(). This is incorrect because
the corresponding rcu_cleanup_dead_rnp() was never invoked. This is
nevertheless harmless because the upper-level bits are still set.
So, no harm, no foul, right?
At least, all is well until a little further into rcu_gp_init()
invocation, which will notice that there are no longer any tasks blocked
on the leaf rcu_node structure, conclude that there is no longer anything
left over from step 2's offline operation, and will therefore invoke
rcu_cleanup_dead_rnp(). But this invocation of rcu_cleanup_dead_rnp()
is for the beginning of the earlier offline interval, and the previous
invocation of rcu_init_new_rnp() is for the end of that same interval.
That is right, they are invoked out of order.
That cannot be good, can it?
It turns out that this is not a (correctness!) problem because
rcu_cleanup_dead_rnp() checks to see if any of the corresponding CPUs
are online, and refuses to do anything if so. In other words, in the
case where rcu_init_new_rnp() and rcu_cleanup_dead_rnp() execute out of
order, they both have no effect.
But this is at best an accident waiting to happen.
This commit therefore adds logic to rcu_gp_init() so that
rcu_init_new_rnp() and rcu_cleanup_dead_rnp() are always invoked in
order, and so that neither are invoked at all in cases where RCU had to
pay attention to the leaf rcu_node structure during the entire time that
all corresponding CPUs were offline.
And, while in the area, this commit reduces confusion by using formal
parameters rather than local variables that just happen to have the same
value at that particular point in the code.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-05-03 03:49:21 +08:00
|
|
|
|
(!rcu_preempt_has_tasks(rnp) || rnp->qsmaskinit)) {
|
rcu: Process offlining and onlining only at grace-period start
Races between CPU hotplug and grace periods can be difficult to resolve,
so the ->onoff_mutex is used to exclude the two events. Unfortunately,
this means that it is impossible for an outgoing CPU to perform the
last bits of its offlining from its last pass through the idle loop,
because sleeplocks cannot be acquired in that context.
This commit avoids these problems by buffering online and offline events
in a new ->qsmaskinitnext field in the leaf rcu_node structures. When a
grace period starts, the events accumulated in this mask are applied to
the ->qsmaskinit field, and, if needed, up the rcu_node tree. The special
case of all CPUs corresponding to a given leaf rcu_node structure being
offline while there are still elements in that structure's ->blkd_tasks
list is handled using a new ->wait_blkd_tasks field. In this case,
propagating the offline bits up the tree is deferred until the beginning
of the grace period after all of the tasks have exited their RCU read-side
critical sections and removed themselves from the list, at which point
the ->wait_blkd_tasks flag is cleared. If one of that leaf rcu_node
structure's CPUs comes back online before the list empties, then the
->wait_blkd_tasks flag is simply cleared.
This of course means that RCU's notion of which CPUs are offline can be
out of date. This is OK because RCU need only wait on CPUs that were
online at the time that the grace period started. In addition, RCU's
force-quiescent-state actions will handle the case where a CPU goes
offline after the grace period starts.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2015-01-24 13:52:37 +08:00
|
|
|
|
rnp->wait_blkd_tasks = false;
|
rcu: Clean up handling of tasks blocked across full-rcu_node offline
Commit 0aa04b055e71 ("rcu: Process offlining and onlining only at
grace-period start") deferred handling of CPU-hotplug events until the
start of the next grace period, but consider the following sequence
of events:
1. A task is preempted within an RCU-preempt read-side critical
section.
2. The CPU that this task was running on goes offline, along with all
other CPUs sharing the corresponding leaf rcu_node structure.
3. The task resumes execution.
4. One of those CPUs comes back online before a new grace period starts.
In step 2, the code in the next rcu_gp_init() invocation will (correctly)
defer removing the leaf rcu_node structure from the upper-level bitmasks,
and will (correctly) set that structure's ->wait_blkd_tasks field. During
the ensuing interval, RCU will (correctly) track the tasks preempted on
that structure because they must block any subsequent grace period.
In step 3, the code in rcu_read_unlock_special() will (correctly) remove
the task from the leaf rcu_node structure. From this point forward, RCU
need not pay attention to this structure, at least not until one of the
corresponding CPUs comes back online.
In step 4, the code in the next rcu_gp_init() invocation will
(incorrectly) invoke rcu_init_new_rnp(). This is incorrect because
the corresponding rcu_cleanup_dead_rnp() was never invoked. This is
nevertheless harmless because the upper-level bits are still set.
So, no harm, no foul, right?
At least, all is well until a little further into rcu_gp_init()
invocation, which will notice that there are no longer any tasks blocked
on the leaf rcu_node structure, conclude that there is no longer anything
left over from step 2's offline operation, and will therefore invoke
rcu_cleanup_dead_rnp(). But this invocation of rcu_cleanup_dead_rnp()
is for the beginning of the earlier offline interval, and the previous
invocation of rcu_init_new_rnp() is for the end of that same interval.
That is right, they are invoked out of order.
That cannot be good, can it?
It turns out that this is not a (correctness!) problem because
rcu_cleanup_dead_rnp() checks to see if any of the corresponding CPUs
are online, and refuses to do anything if so. In other words, in the
case where rcu_init_new_rnp() and rcu_cleanup_dead_rnp() execute out of
order, they both have no effect.
But this is at best an accident waiting to happen.
This commit therefore adds logic to rcu_gp_init() so that
rcu_init_new_rnp() and rcu_cleanup_dead_rnp() are always invoked in
order, and so that neither are invoked at all in cases where RCU had to
pay attention to the leaf rcu_node structure during the entire time that
all corresponding CPUs were offline.
And, while in the area, this commit reduces confusion by using formal
parameters rather than local variables that just happen to have the same
value at that particular point in the code.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-05-03 03:49:21 +08:00
|
|
|
|
if (!rnp->qsmaskinit)
|
|
|
|
|
rcu_cleanup_dead_rnp(rnp);
|
rcu: Process offlining and onlining only at grace-period start
Races between CPU hotplug and grace periods can be difficult to resolve,
so the ->onoff_mutex is used to exclude the two events. Unfortunately,
this means that it is impossible for an outgoing CPU to perform the
last bits of its offlining from its last pass through the idle loop,
because sleeplocks cannot be acquired in that context.
This commit avoids these problems by buffering online and offline events
in a new ->qsmaskinitnext field in the leaf rcu_node structures. When a
grace period starts, the events accumulated in this mask are applied to
the ->qsmaskinit field, and, if needed, up the rcu_node tree. The special
case of all CPUs corresponding to a given leaf rcu_node structure being
offline while there are still elements in that structure's ->blkd_tasks
list is handled using a new ->wait_blkd_tasks field. In this case,
propagating the offline bits up the tree is deferred until the beginning
of the grace period after all of the tasks have exited their RCU read-side
critical sections and removed themselves from the list, at which point
the ->wait_blkd_tasks flag is cleared. If one of that leaf rcu_node
structure's CPUs comes back online before the list empties, then the
->wait_blkd_tasks flag is simply cleared.
This of course means that RCU's notion of which CPUs are offline can be
out of date. This is OK because RCU need only wait on CPUs that were
online at the time that the grace period started. In addition, RCU's
force-quiescent-state actions will handle the case where a CPU goes
offline after the grace period starts.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2015-01-24 13:52:37 +08:00
|
|
|
|
}
|
|
|
|
|
|
2021-02-16 23:04:34 +08:00
|
|
|
|
raw_spin_unlock_rcu_node(rnp);
|
|
|
|
|
arch_spin_unlock(&rcu_state.ofl_lock);
|
|
|
|
|
local_irq_restore(flags);
|
rcu: Process offlining and onlining only at grace-period start
Races between CPU hotplug and grace periods can be difficult to resolve,
so the ->onoff_mutex is used to exclude the two events. Unfortunately,
this means that it is impossible for an outgoing CPU to perform the
last bits of its offlining from its last pass through the idle loop,
because sleeplocks cannot be acquired in that context.
This commit avoids these problems by buffering online and offline events
in a new ->qsmaskinitnext field in the leaf rcu_node structures. When a
grace period starts, the events accumulated in this mask are applied to
the ->qsmaskinit field, and, if needed, up the rcu_node tree. The special
case of all CPUs corresponding to a given leaf rcu_node structure being
offline while there are still elements in that structure's ->blkd_tasks
list is handled using a new ->wait_blkd_tasks field. In this case,
propagating the offline bits up the tree is deferred until the beginning
of the grace period after all of the tasks have exited their RCU read-side
critical sections and removed themselves from the list, at which point
the ->wait_blkd_tasks flag is cleared. If one of that leaf rcu_node
structure's CPUs comes back online before the list empties, then the
->wait_blkd_tasks flag is simply cleared.
This of course means that RCU's notion of which CPUs are offline can be
out of date. This is OK because RCU need only wait on CPUs that were
online at the time that the grace period started. In addition, RCU's
force-quiescent-state actions will handle the case where a CPU goes
offline after the grace period starts.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2015-01-24 13:52:37 +08:00
|
|
|
|
}
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_gp_slow(gp_preinit_delay); /* Races with CPU hotplug. */
|
2012-06-23 02:08:41 +08:00
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Set the quiescent-state-needed bits in all the rcu_node
|
2018-07-06 06:47:01 +08:00
|
|
|
|
* structures for all currently online CPUs in breadth-first
|
|
|
|
|
* order, starting from the root rcu_node structure, relying on the
|
|
|
|
|
* layout of the tree within the rcu_state.node[] array. Note that
|
|
|
|
|
* other CPUs will access only the leaves of the hierarchy, thus
|
|
|
|
|
* seeing that no grace period is in progress, at least until the
|
|
|
|
|
* corresponding leaf node has been initialized.
|
2012-06-23 02:08:41 +08:00
|
|
|
|
*
|
|
|
|
|
* The grace period cannot complete until the initialization
|
|
|
|
|
* process finishes, because this kthread handles both.
|
|
|
|
|
*/
|
rcu: Check and report missed fqs timer wakeup on RCU stall
For a new grace period request, the RCU GP kthread transitions through
following states:
a. [RCU_GP_WAIT_GPS] -> [RCU_GP_DONE_GPS]
The RCU_GP_WAIT_GPS state is where the GP kthread waits for a request
for a new GP. Once it receives a request (for example, when a new RCU
callback is queued), the GP kthread transitions to RCU_GP_DONE_GPS.
b. [RCU_GP_DONE_GPS] -> [RCU_GP_ONOFF]
Grace period initialization starts in rcu_gp_init(), which records the
start of new GP in rcu_state.gp_seq and transitions to RCU_GP_ONOFF.
c. [RCU_GP_ONOFF] -> [RCU_GP_INIT]
The purpose of the RCU_GP_ONOFF state is to apply the online/offline
information that was buffered for any CPUs that recently came online or
went offline. This state is maintained in per-leaf rcu_node bitmasks,
with the buffered state in ->qsmaskinitnext and the state for the upcoming
GP in ->qsmaskinit. At the end of this RCU_GP_ONOFF state, each bit in
->qsmaskinit will correspond to a CPU that must pass through a quiescent
state before the upcoming grace period is allowed to complete.
However, a leaf rcu_node structure with an all-zeroes ->qsmaskinit
cannot necessarily be ignored. In preemptible RCU, there might well be
tasks still in RCU read-side critical sections that were first preempted
while running on one of the CPUs managed by this structure. Such tasks
will be queued on this structure's ->blkd_tasks list. Only after this
list fully drains can this leaf rcu_node structure be ignored, and even
then only if none of its CPUs have come back online in the meantime.
Once that happens, the ->qsmaskinit masks further up the tree will be
updated to exclude this leaf rcu_node structure.
Once the ->qsmaskinitnext and ->qsmaskinit fields have been updated
as needed, the GP kthread transitions to RCU_GP_INIT.
d. [RCU_GP_INIT] -> [RCU_GP_WAIT_FQS]
The purpose of the RCU_GP_INIT state is to copy each ->qsmaskinit to
the ->qsmask field within each rcu_node structure. This copying is done
breadth-first from the root to the leaves. Why not just copy directly
from ->qsmaskinitnext to ->qsmask? Because the ->qsmaskinitnext masks
can change in the meantime as additional CPUs come online or go offline.
Such changes would result in inconsistencies in the ->qsmask fields up and
down the tree, which could in turn result in too-short grace periods or
grace-period hangs. These issues are avoided by snapshotting the leaf
rcu_node structures' ->qsmaskinitnext fields into their ->qsmaskinit
counterparts, generating a consistent set of ->qsmaskinit fields
throughout the tree, and only then copying these consistent ->qsmaskinit
fields to their ->qsmask counterparts.
Once this initialization step is complete, the GP kthread transitions
to RCU_GP_WAIT_FQS, where it waits to do a force-quiescent-state scan
on the one hand or for the end of the grace period on the other.
e. [RCU_GP_WAIT_FQS] -> [RCU_GP_DOING_FQS]
The RCU_GP_WAIT_FQS state waits for one of three things: (1) An
explicit request to do a force-quiescent-state scan, (2) The end of
the grace period, or (3) A short interval of time, after which it
will do a force-quiescent-state (FQS) scan. The explicit request can
come from rcutorture or from any CPU that has too many RCU callbacks
queued (see the qhimark kernel parameter and the RCU_GP_FLAG_OVLD
flag). The aforementioned "short period of time" is specified by the
jiffies_till_first_fqs boot parameter for a given grace period's first
FQS scan and by the jiffies_till_next_fqs for later FQS scans.
Either way, once the wait is over, the GP kthread transitions to
RCU_GP_DOING_FQS.
f. [RCU_GP_DOING_FQS] -> [RCU_GP_CLEANUP]
The RCU_GP_DOING_FQS state performs an FQS scan. Each such scan carries
out two functions for any CPU whose bit is still set in its leaf rcu_node
structure's ->qsmask field, that is, for any CPU that has not yet reported
a quiescent state for the current grace period:
i. Report quiescent states on behalf of CPUs that have been observed
to be idle (from an RCU perspective) since the beginning of the
grace period.
ii. If the current grace period is too old, take various actions to
encourage holdout CPUs to pass through quiescent states, including
enlisting the aid of any calls to cond_resched() and might_sleep(),
and even including IPIing the holdout CPUs.
These checks are skipped for any leaf rcu_node structure with a all-zero
->qsmask field, however such structures are subject to RCU priority
boosting if there are tasks on a given structure blocking the current
grace period. The end of the grace period is detected when the root
rcu_node structure's ->qsmask is zero and when there are no longer any
preempted tasks blocking the current grace period. (No, this last check
is not redundant. To see this, consider an rcu_node tree having exactly
one structure that serves as both root and leaf.)
Once the end of the grace period is detected, the GP kthread transitions
to RCU_GP_CLEANUP.
g. [RCU_GP_CLEANUP] -> [RCU_GP_CLEANED]
The RCU_GP_CLEANUP state marks the end of grace period by updating the
rcu_state structure's ->gp_seq field and also all rcu_node structures'
->gp_seq field. As before, the rcu_node tree is traversed in breadth
first order. Once this update is complete, the GP kthread transitions
to the RCU_GP_CLEANED state.
i. [RCU_GP_CLEANED] -> [RCU_GP_INIT]
Once in the RCU_GP_CLEANED state, the GP kthread immediately transitions
into the RCU_GP_INIT state.
j. The role of timers.
If there is at least one idle CPU, and if timers are not firing, the
transition from RCU_GP_DOING_FQS to RCU_GP_CLEANUP will never happen.
Timers can fail to fire for a number of reasons, including issues in
timer configuration, issues in the timer framework, and failure to handle
softirqs (for example, when there is a storm of interrupts). Whatever the
reason, if the timers fail to fire, the GP kthread will never be awakened,
resulting in RCU CPU stall warnings and eventually in OOM.
However, an RCU CPU stall warning has a large number of potential causes,
as documented in Documentation/RCU/stallwarn.rst. This commit therefore
adds analysis to the RCU CPU stall-warning code to emit an additional
message if the cause of the stall is likely to be timer failure.
Signed-off-by: Neeraj Upadhyay <neeraju@codeaurora.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-11-17 00:06:00 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_state, RCU_GP_INIT);
|
2018-07-05 05:33:59 +08:00
|
|
|
|
rcu_for_each_node_breadth_first(rnp) {
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_gp_slow(gp_init_delay);
|
rcu: Fix grace-period hangs from mid-init task resume
Without special fail-safe quiescent-state-propagation checks, grace-period
hangs can result from the following scenario:
1. A task running on a given CPU is preempted in its RCU read-side
critical section.
2. That CPU goes offline, and there are now no online CPUs
corresponding to that CPU's leaf rcu_node structure.
3. The rcu_gp_init() function does the first phase of grace-period
initialization, and sets the aforementioned leaf rcu_node
structure's ->qsmaskinit field to all zeroes. Because there
is a blocked task, it does not propagate the zeroing of either
->qsmaskinit or ->qsmaskinitnext up the rcu_node tree.
4. The task resumes on some other CPU and exits its critical section.
There is no grace period in progress, so the resulting quiescent
state is not reported up the tree.
5. The rcu_gp_init() function does the second phase of grace-period
initialization, which results in the leaf rcu_node structure
being initialized to expect no further quiescent states, but
with that structure's parent expecting a quiescent-state report.
The parent will never receive a quiescent state from this leaf
rcu_node structure, so the grace period will hang, resulting in
RCU CPU stall warnings.
It would be good to get rid of the special fail-safe quiescent-state
propagation checks. This commit therefore checks the leaf rcu_node
structure's ->wait_blkd_tasks field during grace-period initialization.
If this flag is set, the rcu_report_qs_rnp() is invoked to immediately
report the possible quiescent state. While in the neighborhood, this
commit also report quiescent states for any CPUs that went offline between
the two phases of grace-period initialization, thus reducing grace-period
delays and hopefully eventually allowing removal of offline-CPU checks
from the force-quiescent-state code path.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-05-08 00:34:17 +08:00
|
|
|
|
raw_spin_lock_irqsave_rcu_node(rnp, flags);
|
2018-07-04 06:37:16 +08:00
|
|
|
|
rdp = this_cpu_ptr(&rcu_data);
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_preempt_check_blocked_tasks(rnp);
|
2012-06-23 02:08:41 +08:00
|
|
|
|
rnp->qsmask = rnp->qsmaskinit;
|
2018-07-06 06:47:01 +08:00
|
|
|
|
WRITE_ONCE(rnp->gp_seq, rcu_state.gp_seq);
|
2012-06-23 02:08:41 +08:00
|
|
|
|
if (rnp == rdp->mynode)
|
2018-07-04 08:22:34 +08:00
|
|
|
|
(void)__note_gp_changes(rnp, rdp);
|
2012-06-23 02:08:41 +08:00
|
|
|
|
rcu_preempt_boost_start_gp(rnp);
|
2018-07-06 06:47:01 +08:00
|
|
|
|
trace_rcu_grace_period_init(rcu_state.name, rnp->gp_seq,
|
2012-06-23 02:08:41 +08:00
|
|
|
|
rnp->level, rnp->grplo,
|
|
|
|
|
rnp->grphi, rnp->qsmask);
|
rcu: Fix grace-period hangs from mid-init task resume
Without special fail-safe quiescent-state-propagation checks, grace-period
hangs can result from the following scenario:
1. A task running on a given CPU is preempted in its RCU read-side
critical section.
2. That CPU goes offline, and there are now no online CPUs
corresponding to that CPU's leaf rcu_node structure.
3. The rcu_gp_init() function does the first phase of grace-period
initialization, and sets the aforementioned leaf rcu_node
structure's ->qsmaskinit field to all zeroes. Because there
is a blocked task, it does not propagate the zeroing of either
->qsmaskinit or ->qsmaskinitnext up the rcu_node tree.
4. The task resumes on some other CPU and exits its critical section.
There is no grace period in progress, so the resulting quiescent
state is not reported up the tree.
5. The rcu_gp_init() function does the second phase of grace-period
initialization, which results in the leaf rcu_node structure
being initialized to expect no further quiescent states, but
with that structure's parent expecting a quiescent-state report.
The parent will never receive a quiescent state from this leaf
rcu_node structure, so the grace period will hang, resulting in
RCU CPU stall warnings.
It would be good to get rid of the special fail-safe quiescent-state
propagation checks. This commit therefore checks the leaf rcu_node
structure's ->wait_blkd_tasks field during grace-period initialization.
If this flag is set, the rcu_report_qs_rnp() is invoked to immediately
report the possible quiescent state. While in the neighborhood, this
commit also report quiescent states for any CPUs that went offline between
the two phases of grace-period initialization, thus reducing grace-period
delays and hopefully eventually allowing removal of offline-CPU checks
from the force-quiescent-state code path.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-05-08 00:34:17 +08:00
|
|
|
|
/* Quiescent states for tasks on any now-offline CPUs. */
|
|
|
|
|
mask = rnp->qsmask & ~rnp->qsmaskinitnext;
|
2018-05-16 07:23:23 +08:00
|
|
|
|
rnp->rcu_gp_init_mask = mask;
|
rcu: Fix grace-period hangs from mid-init task resume
Without special fail-safe quiescent-state-propagation checks, grace-period
hangs can result from the following scenario:
1. A task running on a given CPU is preempted in its RCU read-side
critical section.
2. That CPU goes offline, and there are now no online CPUs
corresponding to that CPU's leaf rcu_node structure.
3. The rcu_gp_init() function does the first phase of grace-period
initialization, and sets the aforementioned leaf rcu_node
structure's ->qsmaskinit field to all zeroes. Because there
is a blocked task, it does not propagate the zeroing of either
->qsmaskinit or ->qsmaskinitnext up the rcu_node tree.
4. The task resumes on some other CPU and exits its critical section.
There is no grace period in progress, so the resulting quiescent
state is not reported up the tree.
5. The rcu_gp_init() function does the second phase of grace-period
initialization, which results in the leaf rcu_node structure
being initialized to expect no further quiescent states, but
with that structure's parent expecting a quiescent-state report.
The parent will never receive a quiescent state from this leaf
rcu_node structure, so the grace period will hang, resulting in
RCU CPU stall warnings.
It would be good to get rid of the special fail-safe quiescent-state
propagation checks. This commit therefore checks the leaf rcu_node
structure's ->wait_blkd_tasks field during grace-period initialization.
If this flag is set, the rcu_report_qs_rnp() is invoked to immediately
report the possible quiescent state. While in the neighborhood, this
commit also report quiescent states for any CPUs that went offline between
the two phases of grace-period initialization, thus reducing grace-period
delays and hopefully eventually allowing removal of offline-CPU checks
from the force-quiescent-state code path.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-05-08 00:34:17 +08:00
|
|
|
|
if ((mask || rnp->wait_blkd_tasks) && rcu_is_leaf_node(rnp))
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_report_qs_rnp(mask, rnp, rnp->gp_seq, flags);
|
rcu: Fix grace-period hangs from mid-init task resume
Without special fail-safe quiescent-state-propagation checks, grace-period
hangs can result from the following scenario:
1. A task running on a given CPU is preempted in its RCU read-side
critical section.
2. That CPU goes offline, and there are now no online CPUs
corresponding to that CPU's leaf rcu_node structure.
3. The rcu_gp_init() function does the first phase of grace-period
initialization, and sets the aforementioned leaf rcu_node
structure's ->qsmaskinit field to all zeroes. Because there
is a blocked task, it does not propagate the zeroing of either
->qsmaskinit or ->qsmaskinitnext up the rcu_node tree.
4. The task resumes on some other CPU and exits its critical section.
There is no grace period in progress, so the resulting quiescent
state is not reported up the tree.
5. The rcu_gp_init() function does the second phase of grace-period
initialization, which results in the leaf rcu_node structure
being initialized to expect no further quiescent states, but
with that structure's parent expecting a quiescent-state report.
The parent will never receive a quiescent state from this leaf
rcu_node structure, so the grace period will hang, resulting in
RCU CPU stall warnings.
It would be good to get rid of the special fail-safe quiescent-state
propagation checks. This commit therefore checks the leaf rcu_node
structure's ->wait_blkd_tasks field during grace-period initialization.
If this flag is set, the rcu_report_qs_rnp() is invoked to immediately
report the possible quiescent state. While in the neighborhood, this
commit also report quiescent states for any CPUs that went offline between
the two phases of grace-period initialization, thus reducing grace-period
delays and hopefully eventually allowing removal of offline-CPU checks
from the force-quiescent-state code path.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-05-08 00:34:17 +08:00
|
|
|
|
else
|
|
|
|
|
raw_spin_unlock_irq_rcu_node(rnp);
|
2018-03-03 08:35:27 +08:00
|
|
|
|
cond_resched_tasks_rcu_qs();
|
2018-07-06 06:47:01 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_activity, jiffies);
|
2012-06-23 02:08:41 +08:00
|
|
|
|
}
|
2012-06-19 09:36:08 +08:00
|
|
|
|
|
2020-08-07 10:21:48 +08:00
|
|
|
|
// If strict, make all CPUs aware of new grace period.
|
|
|
|
|
if (IS_ENABLED(CONFIG_RCU_STRICT_GRACE_PERIOD))
|
|
|
|
|
on_each_cpu(rcu_strict_gp_boundary, NULL, 0);
|
|
|
|
|
|
2015-11-08 15:35:00 +08:00
|
|
|
|
return true;
|
2012-06-23 02:08:41 +08:00
|
|
|
|
}
|
2012-06-19 09:36:08 +08:00
|
|
|
|
|
2015-07-02 04:50:28 +08:00
|
|
|
|
/*
|
2018-06-12 16:34:52 +08:00
|
|
|
|
* Helper function for swait_event_idle_exclusive() wakeup at force-quiescent-state
|
2017-06-21 05:45:47 +08:00
|
|
|
|
* time.
|
2015-07-02 04:50:28 +08:00
|
|
|
|
*/
|
2018-07-04 08:22:34 +08:00
|
|
|
|
static bool rcu_gp_fqs_check_wake(int *gfp)
|
2015-07-02 04:50:28 +08:00
|
|
|
|
{
|
2018-07-04 08:22:34 +08:00
|
|
|
|
struct rcu_node *rnp = rcu_get_root();
|
2015-07-02 04:50:28 +08:00
|
|
|
|
|
2020-02-23 12:07:09 +08:00
|
|
|
|
// If under overload conditions, force an immediate FQS scan.
|
|
|
|
|
if (*gfp & RCU_GP_FLAG_OVLD)
|
|
|
|
|
return true;
|
|
|
|
|
|
|
|
|
|
// Someone like call_rcu() requested a force-quiescent-state scan.
|
2018-07-04 08:22:34 +08:00
|
|
|
|
*gfp = READ_ONCE(rcu_state.gp_flags);
|
2015-07-02 04:50:28 +08:00
|
|
|
|
if (*gfp & RCU_GP_FLAG_FQS)
|
|
|
|
|
return true;
|
|
|
|
|
|
2020-02-23 12:07:09 +08:00
|
|
|
|
// The current grace period has completed.
|
2015-07-02 04:50:28 +08:00
|
|
|
|
if (!READ_ONCE(rnp->qsmask) && !rcu_preempt_blocked_readers_cgp(rnp))
|
|
|
|
|
return true;
|
|
|
|
|
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2012-06-23 08:06:26 +08:00
|
|
|
|
/*
|
|
|
|
|
* Do one round of quiescent-state forcing.
|
|
|
|
|
*/
|
2018-07-04 08:22:34 +08:00
|
|
|
|
static void rcu_gp_fqs(bool first_time)
|
2012-06-23 08:06:26 +08:00
|
|
|
|
{
|
2023-09-05 08:02:11 +08:00
|
|
|
|
int nr_fqs = READ_ONCE(rcu_state.nr_fqs_jiffies_stall);
|
2018-07-04 08:22:34 +08:00
|
|
|
|
struct rcu_node *rnp = rcu_get_root();
|
2012-06-23 08:06:26 +08:00
|
|
|
|
|
2018-07-06 06:47:01 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_activity, jiffies);
|
2021-07-20 21:16:27 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.n_force_qs, rcu_state.n_force_qs + 1);
|
2023-09-05 08:02:11 +08:00
|
|
|
|
|
|
|
|
|
WARN_ON_ONCE(nr_fqs > 3);
|
|
|
|
|
/* Only countdown nr_fqs for stall purposes if jiffies moves. */
|
|
|
|
|
if (nr_fqs) {
|
|
|
|
|
if (nr_fqs == 1) {
|
|
|
|
|
WRITE_ONCE(rcu_state.jiffies_stall,
|
|
|
|
|
jiffies + rcu_jiffies_till_stall_check());
|
|
|
|
|
}
|
|
|
|
|
WRITE_ONCE(rcu_state.nr_fqs_jiffies_stall, --nr_fqs);
|
|
|
|
|
}
|
|
|
|
|
|
2015-09-10 03:09:49 +08:00
|
|
|
|
if (first_time) {
|
2012-06-23 08:06:26 +08:00
|
|
|
|
/* Collect dyntick-idle snapshots. */
|
2018-07-04 08:22:34 +08:00
|
|
|
|
force_qs_rnp(dyntick_save_progress_counter);
|
2012-06-23 08:06:26 +08:00
|
|
|
|
} else {
|
|
|
|
|
/* Handle dyntick-idle and offline CPUs. */
|
2018-07-04 08:22:34 +08:00
|
|
|
|
force_qs_rnp(rcu_implicit_dynticks_qs);
|
2012-06-23 08:06:26 +08:00
|
|
|
|
}
|
|
|
|
|
/* Clear flag to prevent immediate re-entry. */
|
2018-07-06 06:47:01 +08:00
|
|
|
|
if (READ_ONCE(rcu_state.gp_flags) & RCU_GP_FLAG_FQS) {
|
2015-10-08 18:24:23 +08:00
|
|
|
|
raw_spin_lock_irq_rcu_node(rnp);
|
2018-07-06 06:47:01 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_flags,
|
|
|
|
|
READ_ONCE(rcu_state.gp_flags) & ~RCU_GP_FLAG_FQS);
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_unlock_irq_rcu_node(rnp);
|
2012-06-23 08:06:26 +08:00
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2018-07-06 09:23:23 +08:00
|
|
|
|
/*
|
|
|
|
|
* Loop doing repeated quiescent-state forcing until the grace period ends.
|
|
|
|
|
*/
|
2021-06-08 12:57:02 +08:00
|
|
|
|
static noinline_for_stack void rcu_gp_fqs_loop(void)
|
2018-07-06 09:23:23 +08:00
|
|
|
|
{
|
2022-06-09 00:34:10 +08:00
|
|
|
|
bool first_gp_fqs = true;
|
2020-02-23 12:07:09 +08:00
|
|
|
|
int gf = 0;
|
2018-07-06 09:23:23 +08:00
|
|
|
|
unsigned long j;
|
|
|
|
|
int ret;
|
|
|
|
|
struct rcu_node *rnp = rcu_get_root();
|
|
|
|
|
|
rcu: Compute jiffies_till_sched_qs from other kernel parameters
The jiffies_till_sched_qs value used to determine how old a grace period
must be before RCU enlists the help of the scheduler to force a quiescent
state on the holdout CPU. Currently, this defaults to HZ/10 regardless of
system size and may be set only at boot time. This can be a problem for
very large systems, because if the values of the jiffies_till_first_fqs
and jiffies_till_next_fqs kernel parameters are left at their defaults,
they are calculated to increase as the number of CPUs actually configured
on the system increases. Thus, on a sufficiently large system, RCU would
enlist the help of the scheduler before the grace-period kthread had a
chance to scan for idle CPUs, which wastes CPU time.
This commit therefore allows jiffies_till_sched_qs to be set, if desired,
but if left as default, computes is as jiffies_till_first_fqs plus twice
jiffies_till_next_fqs, thus allowing three force-quiescent-state scans
for idle CPUs. This scales with the number of CPUs, providing sensible
default values.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-07-26 02:25:23 +08:00
|
|
|
|
j = READ_ONCE(jiffies_till_first_fqs);
|
2020-02-23 12:07:09 +08:00
|
|
|
|
if (rcu_state.cbovld)
|
|
|
|
|
gf = RCU_GP_FLAG_OVLD;
|
2018-07-06 09:23:23 +08:00
|
|
|
|
ret = 0;
|
|
|
|
|
for (;;) {
|
2022-04-13 06:08:14 +08:00
|
|
|
|
if (rcu_state.cbovld) {
|
|
|
|
|
j = (j + 2) / 3;
|
|
|
|
|
if (j <= 0)
|
|
|
|
|
j = 1;
|
|
|
|
|
}
|
|
|
|
|
if (!ret || time_before(jiffies + j, rcu_state.jiffies_force_qs)) {
|
rcu: Check and report missed fqs timer wakeup on RCU stall
For a new grace period request, the RCU GP kthread transitions through
following states:
a. [RCU_GP_WAIT_GPS] -> [RCU_GP_DONE_GPS]
The RCU_GP_WAIT_GPS state is where the GP kthread waits for a request
for a new GP. Once it receives a request (for example, when a new RCU
callback is queued), the GP kthread transitions to RCU_GP_DONE_GPS.
b. [RCU_GP_DONE_GPS] -> [RCU_GP_ONOFF]
Grace period initialization starts in rcu_gp_init(), which records the
start of new GP in rcu_state.gp_seq and transitions to RCU_GP_ONOFF.
c. [RCU_GP_ONOFF] -> [RCU_GP_INIT]
The purpose of the RCU_GP_ONOFF state is to apply the online/offline
information that was buffered for any CPUs that recently came online or
went offline. This state is maintained in per-leaf rcu_node bitmasks,
with the buffered state in ->qsmaskinitnext and the state for the upcoming
GP in ->qsmaskinit. At the end of this RCU_GP_ONOFF state, each bit in
->qsmaskinit will correspond to a CPU that must pass through a quiescent
state before the upcoming grace period is allowed to complete.
However, a leaf rcu_node structure with an all-zeroes ->qsmaskinit
cannot necessarily be ignored. In preemptible RCU, there might well be
tasks still in RCU read-side critical sections that were first preempted
while running on one of the CPUs managed by this structure. Such tasks
will be queued on this structure's ->blkd_tasks list. Only after this
list fully drains can this leaf rcu_node structure be ignored, and even
then only if none of its CPUs have come back online in the meantime.
Once that happens, the ->qsmaskinit masks further up the tree will be
updated to exclude this leaf rcu_node structure.
Once the ->qsmaskinitnext and ->qsmaskinit fields have been updated
as needed, the GP kthread transitions to RCU_GP_INIT.
d. [RCU_GP_INIT] -> [RCU_GP_WAIT_FQS]
The purpose of the RCU_GP_INIT state is to copy each ->qsmaskinit to
the ->qsmask field within each rcu_node structure. This copying is done
breadth-first from the root to the leaves. Why not just copy directly
from ->qsmaskinitnext to ->qsmask? Because the ->qsmaskinitnext masks
can change in the meantime as additional CPUs come online or go offline.
Such changes would result in inconsistencies in the ->qsmask fields up and
down the tree, which could in turn result in too-short grace periods or
grace-period hangs. These issues are avoided by snapshotting the leaf
rcu_node structures' ->qsmaskinitnext fields into their ->qsmaskinit
counterparts, generating a consistent set of ->qsmaskinit fields
throughout the tree, and only then copying these consistent ->qsmaskinit
fields to their ->qsmask counterparts.
Once this initialization step is complete, the GP kthread transitions
to RCU_GP_WAIT_FQS, where it waits to do a force-quiescent-state scan
on the one hand or for the end of the grace period on the other.
e. [RCU_GP_WAIT_FQS] -> [RCU_GP_DOING_FQS]
The RCU_GP_WAIT_FQS state waits for one of three things: (1) An
explicit request to do a force-quiescent-state scan, (2) The end of
the grace period, or (3) A short interval of time, after which it
will do a force-quiescent-state (FQS) scan. The explicit request can
come from rcutorture or from any CPU that has too many RCU callbacks
queued (see the qhimark kernel parameter and the RCU_GP_FLAG_OVLD
flag). The aforementioned "short period of time" is specified by the
jiffies_till_first_fqs boot parameter for a given grace period's first
FQS scan and by the jiffies_till_next_fqs for later FQS scans.
Either way, once the wait is over, the GP kthread transitions to
RCU_GP_DOING_FQS.
f. [RCU_GP_DOING_FQS] -> [RCU_GP_CLEANUP]
The RCU_GP_DOING_FQS state performs an FQS scan. Each such scan carries
out two functions for any CPU whose bit is still set in its leaf rcu_node
structure's ->qsmask field, that is, for any CPU that has not yet reported
a quiescent state for the current grace period:
i. Report quiescent states on behalf of CPUs that have been observed
to be idle (from an RCU perspective) since the beginning of the
grace period.
ii. If the current grace period is too old, take various actions to
encourage holdout CPUs to pass through quiescent states, including
enlisting the aid of any calls to cond_resched() and might_sleep(),
and even including IPIing the holdout CPUs.
These checks are skipped for any leaf rcu_node structure with a all-zero
->qsmask field, however such structures are subject to RCU priority
boosting if there are tasks on a given structure blocking the current
grace period. The end of the grace period is detected when the root
rcu_node structure's ->qsmask is zero and when there are no longer any
preempted tasks blocking the current grace period. (No, this last check
is not redundant. To see this, consider an rcu_node tree having exactly
one structure that serves as both root and leaf.)
Once the end of the grace period is detected, the GP kthread transitions
to RCU_GP_CLEANUP.
g. [RCU_GP_CLEANUP] -> [RCU_GP_CLEANED]
The RCU_GP_CLEANUP state marks the end of grace period by updating the
rcu_state structure's ->gp_seq field and also all rcu_node structures'
->gp_seq field. As before, the rcu_node tree is traversed in breadth
first order. Once this update is complete, the GP kthread transitions
to the RCU_GP_CLEANED state.
i. [RCU_GP_CLEANED] -> [RCU_GP_INIT]
Once in the RCU_GP_CLEANED state, the GP kthread immediately transitions
into the RCU_GP_INIT state.
j. The role of timers.
If there is at least one idle CPU, and if timers are not firing, the
transition from RCU_GP_DOING_FQS to RCU_GP_CLEANUP will never happen.
Timers can fail to fire for a number of reasons, including issues in
timer configuration, issues in the timer framework, and failure to handle
softirqs (for example, when there is a storm of interrupts). Whatever the
reason, if the timers fail to fire, the GP kthread will never be awakened,
resulting in RCU CPU stall warnings and eventually in OOM.
However, an RCU CPU stall warning has a large number of potential causes,
as documented in Documentation/RCU/stallwarn.rst. This commit therefore
adds analysis to the RCU CPU stall-warning code to emit an additional
message if the cause of the stall is likely to be timer failure.
Signed-off-by: Neeraj Upadhyay <neeraju@codeaurora.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-11-17 00:06:00 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.jiffies_force_qs, jiffies + j);
|
|
|
|
|
/*
|
|
|
|
|
* jiffies_force_qs before RCU_GP_WAIT_FQS state
|
|
|
|
|
* update; required for stall checks.
|
|
|
|
|
*/
|
|
|
|
|
smp_wmb();
|
2018-07-06 09:23:23 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.jiffies_kick_kthreads,
|
2018-11-21 02:43:34 +08:00
|
|
|
|
jiffies + (j ? 3 * j : 2));
|
2018-07-06 09:23:23 +08:00
|
|
|
|
}
|
2020-02-11 01:58:37 +08:00
|
|
|
|
trace_rcu_grace_period(rcu_state.name, rcu_state.gp_seq,
|
2018-07-06 09:23:23 +08:00
|
|
|
|
TPS("fqswait"));
|
rcu: Check and report missed fqs timer wakeup on RCU stall
For a new grace period request, the RCU GP kthread transitions through
following states:
a. [RCU_GP_WAIT_GPS] -> [RCU_GP_DONE_GPS]
The RCU_GP_WAIT_GPS state is where the GP kthread waits for a request
for a new GP. Once it receives a request (for example, when a new RCU
callback is queued), the GP kthread transitions to RCU_GP_DONE_GPS.
b. [RCU_GP_DONE_GPS] -> [RCU_GP_ONOFF]
Grace period initialization starts in rcu_gp_init(), which records the
start of new GP in rcu_state.gp_seq and transitions to RCU_GP_ONOFF.
c. [RCU_GP_ONOFF] -> [RCU_GP_INIT]
The purpose of the RCU_GP_ONOFF state is to apply the online/offline
information that was buffered for any CPUs that recently came online or
went offline. This state is maintained in per-leaf rcu_node bitmasks,
with the buffered state in ->qsmaskinitnext and the state for the upcoming
GP in ->qsmaskinit. At the end of this RCU_GP_ONOFF state, each bit in
->qsmaskinit will correspond to a CPU that must pass through a quiescent
state before the upcoming grace period is allowed to complete.
However, a leaf rcu_node structure with an all-zeroes ->qsmaskinit
cannot necessarily be ignored. In preemptible RCU, there might well be
tasks still in RCU read-side critical sections that were first preempted
while running on one of the CPUs managed by this structure. Such tasks
will be queued on this structure's ->blkd_tasks list. Only after this
list fully drains can this leaf rcu_node structure be ignored, and even
then only if none of its CPUs have come back online in the meantime.
Once that happens, the ->qsmaskinit masks further up the tree will be
updated to exclude this leaf rcu_node structure.
Once the ->qsmaskinitnext and ->qsmaskinit fields have been updated
as needed, the GP kthread transitions to RCU_GP_INIT.
d. [RCU_GP_INIT] -> [RCU_GP_WAIT_FQS]
The purpose of the RCU_GP_INIT state is to copy each ->qsmaskinit to
the ->qsmask field within each rcu_node structure. This copying is done
breadth-first from the root to the leaves. Why not just copy directly
from ->qsmaskinitnext to ->qsmask? Because the ->qsmaskinitnext masks
can change in the meantime as additional CPUs come online or go offline.
Such changes would result in inconsistencies in the ->qsmask fields up and
down the tree, which could in turn result in too-short grace periods or
grace-period hangs. These issues are avoided by snapshotting the leaf
rcu_node structures' ->qsmaskinitnext fields into their ->qsmaskinit
counterparts, generating a consistent set of ->qsmaskinit fields
throughout the tree, and only then copying these consistent ->qsmaskinit
fields to their ->qsmask counterparts.
Once this initialization step is complete, the GP kthread transitions
to RCU_GP_WAIT_FQS, where it waits to do a force-quiescent-state scan
on the one hand or for the end of the grace period on the other.
e. [RCU_GP_WAIT_FQS] -> [RCU_GP_DOING_FQS]
The RCU_GP_WAIT_FQS state waits for one of three things: (1) An
explicit request to do a force-quiescent-state scan, (2) The end of
the grace period, or (3) A short interval of time, after which it
will do a force-quiescent-state (FQS) scan. The explicit request can
come from rcutorture or from any CPU that has too many RCU callbacks
queued (see the qhimark kernel parameter and the RCU_GP_FLAG_OVLD
flag). The aforementioned "short period of time" is specified by the
jiffies_till_first_fqs boot parameter for a given grace period's first
FQS scan and by the jiffies_till_next_fqs for later FQS scans.
Either way, once the wait is over, the GP kthread transitions to
RCU_GP_DOING_FQS.
f. [RCU_GP_DOING_FQS] -> [RCU_GP_CLEANUP]
The RCU_GP_DOING_FQS state performs an FQS scan. Each such scan carries
out two functions for any CPU whose bit is still set in its leaf rcu_node
structure's ->qsmask field, that is, for any CPU that has not yet reported
a quiescent state for the current grace period:
i. Report quiescent states on behalf of CPUs that have been observed
to be idle (from an RCU perspective) since the beginning of the
grace period.
ii. If the current grace period is too old, take various actions to
encourage holdout CPUs to pass through quiescent states, including
enlisting the aid of any calls to cond_resched() and might_sleep(),
and even including IPIing the holdout CPUs.
These checks are skipped for any leaf rcu_node structure with a all-zero
->qsmask field, however such structures are subject to RCU priority
boosting if there are tasks on a given structure blocking the current
grace period. The end of the grace period is detected when the root
rcu_node structure's ->qsmask is zero and when there are no longer any
preempted tasks blocking the current grace period. (No, this last check
is not redundant. To see this, consider an rcu_node tree having exactly
one structure that serves as both root and leaf.)
Once the end of the grace period is detected, the GP kthread transitions
to RCU_GP_CLEANUP.
g. [RCU_GP_CLEANUP] -> [RCU_GP_CLEANED]
The RCU_GP_CLEANUP state marks the end of grace period by updating the
rcu_state structure's ->gp_seq field and also all rcu_node structures'
->gp_seq field. As before, the rcu_node tree is traversed in breadth
first order. Once this update is complete, the GP kthread transitions
to the RCU_GP_CLEANED state.
i. [RCU_GP_CLEANED] -> [RCU_GP_INIT]
Once in the RCU_GP_CLEANED state, the GP kthread immediately transitions
into the RCU_GP_INIT state.
j. The role of timers.
If there is at least one idle CPU, and if timers are not firing, the
transition from RCU_GP_DOING_FQS to RCU_GP_CLEANUP will never happen.
Timers can fail to fire for a number of reasons, including issues in
timer configuration, issues in the timer framework, and failure to handle
softirqs (for example, when there is a storm of interrupts). Whatever the
reason, if the timers fail to fire, the GP kthread will never be awakened,
resulting in RCU CPU stall warnings and eventually in OOM.
However, an RCU CPU stall warning has a large number of potential causes,
as documented in Documentation/RCU/stallwarn.rst. This commit therefore
adds analysis to the RCU CPU stall-warning code to emit an additional
message if the cause of the stall is likely to be timer failure.
Signed-off-by: Neeraj Upadhyay <neeraju@codeaurora.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-11-17 00:06:00 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_state, RCU_GP_WAIT_FQS);
|
2021-06-29 21:55:51 +08:00
|
|
|
|
(void)swait_event_idle_timeout_exclusive(rcu_state.gp_wq,
|
|
|
|
|
rcu_gp_fqs_check_wake(&gf), j);
|
2020-04-02 10:57:52 +08:00
|
|
|
|
rcu_gp_torture_wait();
|
rcu: Check and report missed fqs timer wakeup on RCU stall
For a new grace period request, the RCU GP kthread transitions through
following states:
a. [RCU_GP_WAIT_GPS] -> [RCU_GP_DONE_GPS]
The RCU_GP_WAIT_GPS state is where the GP kthread waits for a request
for a new GP. Once it receives a request (for example, when a new RCU
callback is queued), the GP kthread transitions to RCU_GP_DONE_GPS.
b. [RCU_GP_DONE_GPS] -> [RCU_GP_ONOFF]
Grace period initialization starts in rcu_gp_init(), which records the
start of new GP in rcu_state.gp_seq and transitions to RCU_GP_ONOFF.
c. [RCU_GP_ONOFF] -> [RCU_GP_INIT]
The purpose of the RCU_GP_ONOFF state is to apply the online/offline
information that was buffered for any CPUs that recently came online or
went offline. This state is maintained in per-leaf rcu_node bitmasks,
with the buffered state in ->qsmaskinitnext and the state for the upcoming
GP in ->qsmaskinit. At the end of this RCU_GP_ONOFF state, each bit in
->qsmaskinit will correspond to a CPU that must pass through a quiescent
state before the upcoming grace period is allowed to complete.
However, a leaf rcu_node structure with an all-zeroes ->qsmaskinit
cannot necessarily be ignored. In preemptible RCU, there might well be
tasks still in RCU read-side critical sections that were first preempted
while running on one of the CPUs managed by this structure. Such tasks
will be queued on this structure's ->blkd_tasks list. Only after this
list fully drains can this leaf rcu_node structure be ignored, and even
then only if none of its CPUs have come back online in the meantime.
Once that happens, the ->qsmaskinit masks further up the tree will be
updated to exclude this leaf rcu_node structure.
Once the ->qsmaskinitnext and ->qsmaskinit fields have been updated
as needed, the GP kthread transitions to RCU_GP_INIT.
d. [RCU_GP_INIT] -> [RCU_GP_WAIT_FQS]
The purpose of the RCU_GP_INIT state is to copy each ->qsmaskinit to
the ->qsmask field within each rcu_node structure. This copying is done
breadth-first from the root to the leaves. Why not just copy directly
from ->qsmaskinitnext to ->qsmask? Because the ->qsmaskinitnext masks
can change in the meantime as additional CPUs come online or go offline.
Such changes would result in inconsistencies in the ->qsmask fields up and
down the tree, which could in turn result in too-short grace periods or
grace-period hangs. These issues are avoided by snapshotting the leaf
rcu_node structures' ->qsmaskinitnext fields into their ->qsmaskinit
counterparts, generating a consistent set of ->qsmaskinit fields
throughout the tree, and only then copying these consistent ->qsmaskinit
fields to their ->qsmask counterparts.
Once this initialization step is complete, the GP kthread transitions
to RCU_GP_WAIT_FQS, where it waits to do a force-quiescent-state scan
on the one hand or for the end of the grace period on the other.
e. [RCU_GP_WAIT_FQS] -> [RCU_GP_DOING_FQS]
The RCU_GP_WAIT_FQS state waits for one of three things: (1) An
explicit request to do a force-quiescent-state scan, (2) The end of
the grace period, or (3) A short interval of time, after which it
will do a force-quiescent-state (FQS) scan. The explicit request can
come from rcutorture or from any CPU that has too many RCU callbacks
queued (see the qhimark kernel parameter and the RCU_GP_FLAG_OVLD
flag). The aforementioned "short period of time" is specified by the
jiffies_till_first_fqs boot parameter for a given grace period's first
FQS scan and by the jiffies_till_next_fqs for later FQS scans.
Either way, once the wait is over, the GP kthread transitions to
RCU_GP_DOING_FQS.
f. [RCU_GP_DOING_FQS] -> [RCU_GP_CLEANUP]
The RCU_GP_DOING_FQS state performs an FQS scan. Each such scan carries
out two functions for any CPU whose bit is still set in its leaf rcu_node
structure's ->qsmask field, that is, for any CPU that has not yet reported
a quiescent state for the current grace period:
i. Report quiescent states on behalf of CPUs that have been observed
to be idle (from an RCU perspective) since the beginning of the
grace period.
ii. If the current grace period is too old, take various actions to
encourage holdout CPUs to pass through quiescent states, including
enlisting the aid of any calls to cond_resched() and might_sleep(),
and even including IPIing the holdout CPUs.
These checks are skipped for any leaf rcu_node structure with a all-zero
->qsmask field, however such structures are subject to RCU priority
boosting if there are tasks on a given structure blocking the current
grace period. The end of the grace period is detected when the root
rcu_node structure's ->qsmask is zero and when there are no longer any
preempted tasks blocking the current grace period. (No, this last check
is not redundant. To see this, consider an rcu_node tree having exactly
one structure that serves as both root and leaf.)
Once the end of the grace period is detected, the GP kthread transitions
to RCU_GP_CLEANUP.
g. [RCU_GP_CLEANUP] -> [RCU_GP_CLEANED]
The RCU_GP_CLEANUP state marks the end of grace period by updating the
rcu_state structure's ->gp_seq field and also all rcu_node structures'
->gp_seq field. As before, the rcu_node tree is traversed in breadth
first order. Once this update is complete, the GP kthread transitions
to the RCU_GP_CLEANED state.
i. [RCU_GP_CLEANED] -> [RCU_GP_INIT]
Once in the RCU_GP_CLEANED state, the GP kthread immediately transitions
into the RCU_GP_INIT state.
j. The role of timers.
If there is at least one idle CPU, and if timers are not firing, the
transition from RCU_GP_DOING_FQS to RCU_GP_CLEANUP will never happen.
Timers can fail to fire for a number of reasons, including issues in
timer configuration, issues in the timer framework, and failure to handle
softirqs (for example, when there is a storm of interrupts). Whatever the
reason, if the timers fail to fire, the GP kthread will never be awakened,
resulting in RCU CPU stall warnings and eventually in OOM.
However, an RCU CPU stall warning has a large number of potential causes,
as documented in Documentation/RCU/stallwarn.rst. This commit therefore
adds analysis to the RCU CPU stall-warning code to emit an additional
message if the cause of the stall is likely to be timer failure.
Signed-off-by: Neeraj Upadhyay <neeraju@codeaurora.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-11-17 00:06:00 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_state, RCU_GP_DOING_FQS);
|
2018-07-06 09:23:23 +08:00
|
|
|
|
/* Locking provides needed memory barriers. */
|
2022-06-09 15:13:40 +08:00
|
|
|
|
/*
|
|
|
|
|
* Exit the loop if the root rcu_node structure indicates that the grace period
|
|
|
|
|
* has ended, leave the loop. The rcu_preempt_blocked_readers_cgp(rnp) check
|
|
|
|
|
* is required only for single-node rcu_node trees because readers blocking
|
|
|
|
|
* the current grace period are queued only on leaf rcu_node structures.
|
|
|
|
|
* For multi-node trees, checking the root node's ->qsmask suffices, because a
|
|
|
|
|
* given root node's ->qsmask bit is cleared only when all CPUs and tasks from
|
|
|
|
|
* the corresponding leaf nodes have passed through their quiescent state.
|
|
|
|
|
*/
|
2018-07-06 09:23:23 +08:00
|
|
|
|
if (!READ_ONCE(rnp->qsmask) &&
|
|
|
|
|
!rcu_preempt_blocked_readers_cgp(rnp))
|
|
|
|
|
break;
|
|
|
|
|
/* If time for quiescent-state forcing, do it. */
|
2020-04-11 05:48:20 +08:00
|
|
|
|
if (!time_after(rcu_state.jiffies_force_qs, jiffies) ||
|
2020-06-22 02:37:27 +08:00
|
|
|
|
(gf & (RCU_GP_FLAG_FQS | RCU_GP_FLAG_OVLD))) {
|
2020-02-11 01:58:37 +08:00
|
|
|
|
trace_rcu_grace_period(rcu_state.name, rcu_state.gp_seq,
|
2018-07-06 09:23:23 +08:00
|
|
|
|
TPS("fqsstart"));
|
|
|
|
|
rcu_gp_fqs(first_gp_fqs);
|
2020-02-23 12:07:09 +08:00
|
|
|
|
gf = 0;
|
|
|
|
|
if (first_gp_fqs) {
|
|
|
|
|
first_gp_fqs = false;
|
|
|
|
|
gf = rcu_state.cbovld ? RCU_GP_FLAG_OVLD : 0;
|
|
|
|
|
}
|
2020-02-11 01:58:37 +08:00
|
|
|
|
trace_rcu_grace_period(rcu_state.name, rcu_state.gp_seq,
|
2018-07-06 09:23:23 +08:00
|
|
|
|
TPS("fqsend"));
|
|
|
|
|
cond_resched_tasks_rcu_qs();
|
|
|
|
|
WRITE_ONCE(rcu_state.gp_activity, jiffies);
|
|
|
|
|
ret = 0; /* Force full wait till next FQS. */
|
rcu: Compute jiffies_till_sched_qs from other kernel parameters
The jiffies_till_sched_qs value used to determine how old a grace period
must be before RCU enlists the help of the scheduler to force a quiescent
state on the holdout CPU. Currently, this defaults to HZ/10 regardless of
system size and may be set only at boot time. This can be a problem for
very large systems, because if the values of the jiffies_till_first_fqs
and jiffies_till_next_fqs kernel parameters are left at their defaults,
they are calculated to increase as the number of CPUs actually configured
on the system increases. Thus, on a sufficiently large system, RCU would
enlist the help of the scheduler before the grace-period kthread had a
chance to scan for idle CPUs, which wastes CPU time.
This commit therefore allows jiffies_till_sched_qs to be set, if desired,
but if left as default, computes is as jiffies_till_first_fqs plus twice
jiffies_till_next_fqs, thus allowing three force-quiescent-state scans
for idle CPUs. This scales with the number of CPUs, providing sensible
default values.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-07-26 02:25:23 +08:00
|
|
|
|
j = READ_ONCE(jiffies_till_next_fqs);
|
2018-07-06 09:23:23 +08:00
|
|
|
|
} else {
|
|
|
|
|
/* Deal with stray signal. */
|
|
|
|
|
cond_resched_tasks_rcu_qs();
|
|
|
|
|
WRITE_ONCE(rcu_state.gp_activity, jiffies);
|
|
|
|
|
WARN_ON(signal_pending(current));
|
2020-02-11 01:58:37 +08:00
|
|
|
|
trace_rcu_grace_period(rcu_state.name, rcu_state.gp_seq,
|
2018-07-06 09:23:23 +08:00
|
|
|
|
TPS("fqswaitsig"));
|
|
|
|
|
ret = 1; /* Keep old FQS timing. */
|
|
|
|
|
j = jiffies;
|
|
|
|
|
if (time_after(jiffies, rcu_state.jiffies_force_qs))
|
|
|
|
|
j = 1;
|
|
|
|
|
else
|
|
|
|
|
j = rcu_state.jiffies_force_qs - j;
|
2020-02-23 12:07:09 +08:00
|
|
|
|
gf = 0;
|
2018-07-06 09:23:23 +08:00
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2012-06-23 02:08:41 +08:00
|
|
|
|
/*
|
|
|
|
|
* Clean up after the old grace period.
|
|
|
|
|
*/
|
2021-04-12 01:49:52 +08:00
|
|
|
|
static noinline void rcu_gp_cleanup(void)
|
2012-06-23 02:08:41 +08:00
|
|
|
|
{
|
2019-10-31 02:56:10 +08:00
|
|
|
|
int cpu;
|
rcu: Make callers awaken grace-period kthread
The rcu_start_gp_advanced() function currently uses irq_work_queue()
to defer wakeups of the RCU grace-period kthread. This deferring
is necessary to avoid RCU-scheduler deadlocks involving the rcu_node
structure's lock, meaning that RCU cannot call any of the scheduler's
wake-up functions while holding one of these locks.
Unfortunately, the second and subsequent calls to irq_work_queue() are
ignored, and the first call will be ignored (aside from queuing the work
item) if the scheduler-clock tick is turned off. This is OK for many
uses, especially those where irq_work_queue() is called from an interrupt
or softirq handler, because in those cases the scheduler-clock-tick state
will be re-evaluated, which will turn the scheduler-clock tick back on.
On the next tick, any deferred work will then be processed.
However, this strategy does not always work for RCU, which can be invoked
at process level from idle CPUs. In this case, the tick might never
be turned back on, indefinitely defering a grace-period start request.
Note that the RCU CPU stall detector cannot see this condition, because
there is no RCU grace period in progress. Therefore, we can (and do!)
see long tens-of-seconds stalls in grace-period handling. In theory,
we could see a full grace-period hang, but rcutorture testing to date
has seen only the tens-of-seconds stalls. Event tracing demonstrates
that irq_work_queue() is being called repeatedly to no effect during
these stalls: The "newreq" event appears repeatedly from a task that is
not one of the grace-period kthreads.
In theory, irq_work_queue() might be fixed to avoid this sort of issue,
but RCU's requirements are unusual and it is quite straightforward to pass
wake-up responsibility up through RCU's call chain, so that the wakeup
happens when the offending locks are released.
This commit therefore makes this change. The rcu_start_gp_advanced(),
rcu_start_future_gp(), rcu_accelerate_cbs(), rcu_advance_cbs(),
__note_gp_changes(), and rcu_start_gp() functions now return a boolean
which indicates when a wake-up is needed. A new rcu_gp_kthread_wake()
does the wakeup when it is necessary and safe to do so: No self-wakes,
no wake-ups if the ->gp_flags field indicates there is no need (as in
someone else did the wake-up before we got around to it), and no wake-ups
before the grace-period kthread has been created.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2014-03-12 04:02:16 +08:00
|
|
|
|
bool needgp = false;
|
2019-10-31 02:56:10 +08:00
|
|
|
|
unsigned long gp_duration;
|
2018-04-27 02:52:09 +08:00
|
|
|
|
unsigned long new_gp_seq;
|
2019-05-16 00:56:40 +08:00
|
|
|
|
bool offloaded;
|
2012-06-23 02:08:41 +08:00
|
|
|
|
struct rcu_data *rdp;
|
2018-07-04 08:22:34 +08:00
|
|
|
|
struct rcu_node *rnp = rcu_get_root();
|
2016-02-19 16:46:41 +08:00
|
|
|
|
struct swait_queue_head *sq;
|
2012-06-19 09:36:08 +08:00
|
|
|
|
|
2018-07-06 06:47:01 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_activity, jiffies);
|
2015-10-08 18:24:23 +08:00
|
|
|
|
raw_spin_lock_irq_rcu_node(rnp);
|
2018-10-04 08:25:33 +08:00
|
|
|
|
rcu_state.gp_end = jiffies;
|
|
|
|
|
gp_duration = rcu_state.gp_end - rcu_state.gp_start;
|
2018-07-06 06:47:01 +08:00
|
|
|
|
if (gp_duration > rcu_state.gp_max)
|
|
|
|
|
rcu_state.gp_max = gp_duration;
|
2012-06-19 09:36:08 +08:00
|
|
|
|
|
2012-06-23 02:08:41 +08:00
|
|
|
|
/*
|
|
|
|
|
* We know the grace period is complete, but to everyone else
|
|
|
|
|
* it appears to still be ongoing. But it is also the case
|
|
|
|
|
* that to everyone else it looks like there is nothing that
|
|
|
|
|
* they can do to advance the grace period. It is therefore
|
|
|
|
|
* safe for us to drop the lock in order to mark the grace
|
|
|
|
|
* period as completed in all of the rcu_node structures.
|
|
|
|
|
*/
|
2022-04-14 08:46:15 +08:00
|
|
|
|
rcu_poll_gp_seq_end(&rcu_state.gp_seq_polled_snap);
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_unlock_irq_rcu_node(rnp);
|
2012-06-19 09:36:08 +08:00
|
|
|
|
|
rcu: Fix day-zero grace-period initialization/cleanup race
The current approach to grace-period initialization is vulnerable to
extremely low-probability races. These races stem from the fact that
the old grace period is marked completed on the same traversal through
the rcu_node structure that is marking the start of the new grace period.
This means that some rcu_node structures will believe that the old grace
period is still in effect at the same time that other rcu_node structures
believe that the new grace period has already started.
These sorts of disagreements can result in too-short grace periods,
as shown in the following scenario:
1. CPU 0 completes a grace period, but needs an additional
grace period, so starts initializing one, initializing all
the non-leaf rcu_node structures and the first leaf rcu_node
structure. Because CPU 0 is both completing the old grace
period and starting a new one, it marks the completion of
the old grace period and the start of the new grace period
in a single traversal of the rcu_node structures.
Therefore, CPUs corresponding to the first rcu_node structure
can become aware that the prior grace period has completed, but
CPUs corresponding to the other rcu_node structures will see
this same prior grace period as still being in progress.
2. CPU 1 passes through a quiescent state, and therefore informs
the RCU core. Because its leaf rcu_node structure has already
been initialized, this CPU's quiescent state is applied to the
new (and only partially initialized) grace period.
3. CPU 1 enters an RCU read-side critical section and acquires
a reference to data item A. Note that this CPU believes that
its critical section started after the beginning of the new
grace period, and therefore will not block this new grace period.
4. CPU 16 exits dyntick-idle mode. Because it was in dyntick-idle
mode, other CPUs informed the RCU core of its extended quiescent
state for the past several grace periods. This means that CPU 16
is not yet aware that these past grace periods have ended. Assume
that CPU 16 corresponds to the second leaf rcu_node structure --
which has not yet been made aware of the new grace period.
5. CPU 16 removes data item A from its enclosing data structure
and passes it to call_rcu(), which queues a callback in the
RCU_NEXT_TAIL segment of the callback queue.
6. CPU 16 enters the RCU core, possibly because it has taken a
scheduling-clock interrupt, or alternatively because it has
more than 10,000 callbacks queued. It notes that the second
most recent grace period has completed (recall that because it
corresponds to the second as-yet-uninitialized rcu_node structure,
it cannot yet become aware that the most recent grace period has
completed), and therefore advances its callbacks. The callback
for data item A is therefore in the RCU_NEXT_READY_TAIL segment
of the callback queue.
7. CPU 0 completes initialization of the remaining leaf rcu_node
structures for the new grace period, including the structure
corresponding to CPU 16.
8. CPU 16 again enters the RCU core, again, possibly because it has
taken a scheduling-clock interrupt, or alternatively because
it now has more than 10,000 callbacks queued. It notes that
the most recent grace period has ended, and therefore advances
its callbacks. The callback for data item A is therefore in
the RCU_DONE_TAIL segment of the callback queue.
9. All CPUs other than CPU 1 pass through quiescent states. Because
CPU 1 already passed through its quiescent state, the new grace
period completes. Note that CPU 1 is still in its RCU read-side
critical section, still referencing data item A.
10. Suppose that CPU 2 wais the last CPU to pass through a quiescent
state for the new grace period, and suppose further that CPU 2
did not have any callbacks queued, therefore not needing an
additional grace period. CPU 2 therefore traverses all of the
rcu_node structures, marking the new grace period as completed,
but does not initialize a new grace period.
11. CPU 16 yet again enters the RCU core, yet again possibly because
it has taken a scheduling-clock interrupt, or alternatively
because it now has more than 10,000 callbacks queued. It notes
that the new grace period has ended, and therefore advances
its callbacks. The callback for data item A is therefore in
the RCU_DONE_TAIL segment of the callback queue. This means
that this callback is now considered ready to be invoked.
12. CPU 16 invokes the callback, freeing data item A while CPU 1
is still referencing it.
This scenario represents a day-zero bug for TREE_RCU. This commit
therefore ensures that the old grace period is marked completed in
all leaf rcu_node structures before a new grace period is marked
started in any of them.
That said, it would have been insanely difficult to force this race to
happen before the grace-period initialization process was preemptible.
Therefore, this commit is not a candidate for -stable.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Conflicts:
kernel/rcutree.c
2012-07-07 22:56:57 +08:00
|
|
|
|
/*
|
2018-05-02 05:34:08 +08:00
|
|
|
|
* Propagate new ->gp_seq value to rcu_node structures so that
|
|
|
|
|
* other CPUs don't have to wait until the start of the next grace
|
|
|
|
|
* period to process their callbacks. This also avoids some nasty
|
|
|
|
|
* RCU grace-period initialization races by forcing the end of
|
|
|
|
|
* the current grace period to be completely recorded in all of
|
|
|
|
|
* the rcu_node structures before the beginning of the next grace
|
|
|
|
|
* period is recorded in any of the rcu_node structures.
|
rcu: Fix day-zero grace-period initialization/cleanup race
The current approach to grace-period initialization is vulnerable to
extremely low-probability races. These races stem from the fact that
the old grace period is marked completed on the same traversal through
the rcu_node structure that is marking the start of the new grace period.
This means that some rcu_node structures will believe that the old grace
period is still in effect at the same time that other rcu_node structures
believe that the new grace period has already started.
These sorts of disagreements can result in too-short grace periods,
as shown in the following scenario:
1. CPU 0 completes a grace period, but needs an additional
grace period, so starts initializing one, initializing all
the non-leaf rcu_node structures and the first leaf rcu_node
structure. Because CPU 0 is both completing the old grace
period and starting a new one, it marks the completion of
the old grace period and the start of the new grace period
in a single traversal of the rcu_node structures.
Therefore, CPUs corresponding to the first rcu_node structure
can become aware that the prior grace period has completed, but
CPUs corresponding to the other rcu_node structures will see
this same prior grace period as still being in progress.
2. CPU 1 passes through a quiescent state, and therefore informs
the RCU core. Because its leaf rcu_node structure has already
been initialized, this CPU's quiescent state is applied to the
new (and only partially initialized) grace period.
3. CPU 1 enters an RCU read-side critical section and acquires
a reference to data item A. Note that this CPU believes that
its critical section started after the beginning of the new
grace period, and therefore will not block this new grace period.
4. CPU 16 exits dyntick-idle mode. Because it was in dyntick-idle
mode, other CPUs informed the RCU core of its extended quiescent
state for the past several grace periods. This means that CPU 16
is not yet aware that these past grace periods have ended. Assume
that CPU 16 corresponds to the second leaf rcu_node structure --
which has not yet been made aware of the new grace period.
5. CPU 16 removes data item A from its enclosing data structure
and passes it to call_rcu(), which queues a callback in the
RCU_NEXT_TAIL segment of the callback queue.
6. CPU 16 enters the RCU core, possibly because it has taken a
scheduling-clock interrupt, or alternatively because it has
more than 10,000 callbacks queued. It notes that the second
most recent grace period has completed (recall that because it
corresponds to the second as-yet-uninitialized rcu_node structure,
it cannot yet become aware that the most recent grace period has
completed), and therefore advances its callbacks. The callback
for data item A is therefore in the RCU_NEXT_READY_TAIL segment
of the callback queue.
7. CPU 0 completes initialization of the remaining leaf rcu_node
structures for the new grace period, including the structure
corresponding to CPU 16.
8. CPU 16 again enters the RCU core, again, possibly because it has
taken a scheduling-clock interrupt, or alternatively because
it now has more than 10,000 callbacks queued. It notes that
the most recent grace period has ended, and therefore advances
its callbacks. The callback for data item A is therefore in
the RCU_DONE_TAIL segment of the callback queue.
9. All CPUs other than CPU 1 pass through quiescent states. Because
CPU 1 already passed through its quiescent state, the new grace
period completes. Note that CPU 1 is still in its RCU read-side
critical section, still referencing data item A.
10. Suppose that CPU 2 wais the last CPU to pass through a quiescent
state for the new grace period, and suppose further that CPU 2
did not have any callbacks queued, therefore not needing an
additional grace period. CPU 2 therefore traverses all of the
rcu_node structures, marking the new grace period as completed,
but does not initialize a new grace period.
11. CPU 16 yet again enters the RCU core, yet again possibly because
it has taken a scheduling-clock interrupt, or alternatively
because it now has more than 10,000 callbacks queued. It notes
that the new grace period has ended, and therefore advances
its callbacks. The callback for data item A is therefore in
the RCU_DONE_TAIL segment of the callback queue. This means
that this callback is now considered ready to be invoked.
12. CPU 16 invokes the callback, freeing data item A while CPU 1
is still referencing it.
This scenario represents a day-zero bug for TREE_RCU. This commit
therefore ensures that the old grace period is marked completed in
all leaf rcu_node structures before a new grace period is marked
started in any of them.
That said, it would have been insanely difficult to force this race to
happen before the grace-period initialization process was preemptible.
Therefore, this commit is not a candidate for -stable.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Conflicts:
kernel/rcutree.c
2012-07-07 22:56:57 +08:00
|
|
|
|
*/
|
2018-07-06 06:47:01 +08:00
|
|
|
|
new_gp_seq = rcu_state.gp_seq;
|
2018-04-27 02:52:09 +08:00
|
|
|
|
rcu_seq_end(&new_gp_seq);
|
2018-07-05 05:33:59 +08:00
|
|
|
|
rcu_for_each_node_breadth_first(rnp) {
|
2015-10-08 18:24:23 +08:00
|
|
|
|
raw_spin_lock_irq_rcu_node(rnp);
|
2017-11-28 07:13:56 +08:00
|
|
|
|
if (WARN_ON_ONCE(rcu_preempt_blocked_readers_cgp(rnp)))
|
2018-07-04 08:22:34 +08:00
|
|
|
|
dump_blkd_tasks(rnp, 10);
|
2015-02-09 21:37:47 +08:00
|
|
|
|
WARN_ON_ONCE(rnp->qsmask);
|
2018-04-27 02:52:09 +08:00
|
|
|
|
WRITE_ONCE(rnp->gp_seq, new_gp_seq);
|
2022-07-29 10:58:13 +08:00
|
|
|
|
if (!rnp->parent)
|
|
|
|
|
smp_mb(); // Order against failing poll_state_synchronize_rcu_full().
|
2018-07-04 06:37:16 +08:00
|
|
|
|
rdp = this_cpu_ptr(&rcu_data);
|
2012-12-18 06:21:14 +08:00
|
|
|
|
if (rnp == rdp->mynode)
|
2018-07-04 08:22:34 +08:00
|
|
|
|
needgp = __note_gp_changes(rnp, rdp) || needgp;
|
2013-09-25 06:04:06 +08:00
|
|
|
|
/* smp_mb() provided by prior unlock-lock pair. */
|
2018-07-04 08:22:34 +08:00
|
|
|
|
needgp = rcu_future_gp_cleanup(rnp) || needgp;
|
2019-10-31 02:56:10 +08:00
|
|
|
|
// Reset overload indication for CPUs no longer overloaded
|
|
|
|
|
if (rcu_is_leaf_node(rnp))
|
|
|
|
|
for_each_leaf_node_cpu_mask(rnp, cpu, rnp->cbovldmask) {
|
|
|
|
|
rdp = per_cpu_ptr(&rcu_data, cpu);
|
|
|
|
|
check_cb_ovld_locked(rdp, rnp);
|
|
|
|
|
}
|
2016-02-19 16:46:40 +08:00
|
|
|
|
sq = rcu_nocb_gp_get(rnp);
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_unlock_irq_rcu_node(rnp);
|
2016-02-19 16:46:40 +08:00
|
|
|
|
rcu_nocb_gp_cleanup(sq);
|
2018-03-03 08:35:27 +08:00
|
|
|
|
cond_resched_tasks_rcu_qs();
|
2018-07-06 06:47:01 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_activity, jiffies);
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_gp_slow(gp_cleanup_delay);
|
2012-06-23 02:08:41 +08:00
|
|
|
|
}
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rnp = rcu_get_root();
|
2018-07-06 06:47:01 +08:00
|
|
|
|
raw_spin_lock_irq_rcu_node(rnp); /* GP before ->gp_seq update. */
|
2012-06-23 02:08:41 +08:00
|
|
|
|
|
2018-10-16 01:00:58 +08:00
|
|
|
|
/* Declare grace period done, trace first to use old GP number. */
|
2018-07-06 06:47:01 +08:00
|
|
|
|
trace_rcu_grace_period(rcu_state.name, rcu_state.gp_seq, TPS("end"));
|
2018-10-16 01:00:58 +08:00
|
|
|
|
rcu_seq_end(&rcu_state.gp_seq);
|
2020-03-22 10:52:20 +08:00
|
|
|
|
ASSERT_EXCLUSIVE_WRITER(rcu_state.gp_seq);
|
rcu: Check and report missed fqs timer wakeup on RCU stall
For a new grace period request, the RCU GP kthread transitions through
following states:
a. [RCU_GP_WAIT_GPS] -> [RCU_GP_DONE_GPS]
The RCU_GP_WAIT_GPS state is where the GP kthread waits for a request
for a new GP. Once it receives a request (for example, when a new RCU
callback is queued), the GP kthread transitions to RCU_GP_DONE_GPS.
b. [RCU_GP_DONE_GPS] -> [RCU_GP_ONOFF]
Grace period initialization starts in rcu_gp_init(), which records the
start of new GP in rcu_state.gp_seq and transitions to RCU_GP_ONOFF.
c. [RCU_GP_ONOFF] -> [RCU_GP_INIT]
The purpose of the RCU_GP_ONOFF state is to apply the online/offline
information that was buffered for any CPUs that recently came online or
went offline. This state is maintained in per-leaf rcu_node bitmasks,
with the buffered state in ->qsmaskinitnext and the state for the upcoming
GP in ->qsmaskinit. At the end of this RCU_GP_ONOFF state, each bit in
->qsmaskinit will correspond to a CPU that must pass through a quiescent
state before the upcoming grace period is allowed to complete.
However, a leaf rcu_node structure with an all-zeroes ->qsmaskinit
cannot necessarily be ignored. In preemptible RCU, there might well be
tasks still in RCU read-side critical sections that were first preempted
while running on one of the CPUs managed by this structure. Such tasks
will be queued on this structure's ->blkd_tasks list. Only after this
list fully drains can this leaf rcu_node structure be ignored, and even
then only if none of its CPUs have come back online in the meantime.
Once that happens, the ->qsmaskinit masks further up the tree will be
updated to exclude this leaf rcu_node structure.
Once the ->qsmaskinitnext and ->qsmaskinit fields have been updated
as needed, the GP kthread transitions to RCU_GP_INIT.
d. [RCU_GP_INIT] -> [RCU_GP_WAIT_FQS]
The purpose of the RCU_GP_INIT state is to copy each ->qsmaskinit to
the ->qsmask field within each rcu_node structure. This copying is done
breadth-first from the root to the leaves. Why not just copy directly
from ->qsmaskinitnext to ->qsmask? Because the ->qsmaskinitnext masks
can change in the meantime as additional CPUs come online or go offline.
Such changes would result in inconsistencies in the ->qsmask fields up and
down the tree, which could in turn result in too-short grace periods or
grace-period hangs. These issues are avoided by snapshotting the leaf
rcu_node structures' ->qsmaskinitnext fields into their ->qsmaskinit
counterparts, generating a consistent set of ->qsmaskinit fields
throughout the tree, and only then copying these consistent ->qsmaskinit
fields to their ->qsmask counterparts.
Once this initialization step is complete, the GP kthread transitions
to RCU_GP_WAIT_FQS, where it waits to do a force-quiescent-state scan
on the one hand or for the end of the grace period on the other.
e. [RCU_GP_WAIT_FQS] -> [RCU_GP_DOING_FQS]
The RCU_GP_WAIT_FQS state waits for one of three things: (1) An
explicit request to do a force-quiescent-state scan, (2) The end of
the grace period, or (3) A short interval of time, after which it
will do a force-quiescent-state (FQS) scan. The explicit request can
come from rcutorture or from any CPU that has too many RCU callbacks
queued (see the qhimark kernel parameter and the RCU_GP_FLAG_OVLD
flag). The aforementioned "short period of time" is specified by the
jiffies_till_first_fqs boot parameter for a given grace period's first
FQS scan and by the jiffies_till_next_fqs for later FQS scans.
Either way, once the wait is over, the GP kthread transitions to
RCU_GP_DOING_FQS.
f. [RCU_GP_DOING_FQS] -> [RCU_GP_CLEANUP]
The RCU_GP_DOING_FQS state performs an FQS scan. Each such scan carries
out two functions for any CPU whose bit is still set in its leaf rcu_node
structure's ->qsmask field, that is, for any CPU that has not yet reported
a quiescent state for the current grace period:
i. Report quiescent states on behalf of CPUs that have been observed
to be idle (from an RCU perspective) since the beginning of the
grace period.
ii. If the current grace period is too old, take various actions to
encourage holdout CPUs to pass through quiescent states, including
enlisting the aid of any calls to cond_resched() and might_sleep(),
and even including IPIing the holdout CPUs.
These checks are skipped for any leaf rcu_node structure with a all-zero
->qsmask field, however such structures are subject to RCU priority
boosting if there are tasks on a given structure blocking the current
grace period. The end of the grace period is detected when the root
rcu_node structure's ->qsmask is zero and when there are no longer any
preempted tasks blocking the current grace period. (No, this last check
is not redundant. To see this, consider an rcu_node tree having exactly
one structure that serves as both root and leaf.)
Once the end of the grace period is detected, the GP kthread transitions
to RCU_GP_CLEANUP.
g. [RCU_GP_CLEANUP] -> [RCU_GP_CLEANED]
The RCU_GP_CLEANUP state marks the end of grace period by updating the
rcu_state structure's ->gp_seq field and also all rcu_node structures'
->gp_seq field. As before, the rcu_node tree is traversed in breadth
first order. Once this update is complete, the GP kthread transitions
to the RCU_GP_CLEANED state.
i. [RCU_GP_CLEANED] -> [RCU_GP_INIT]
Once in the RCU_GP_CLEANED state, the GP kthread immediately transitions
into the RCU_GP_INIT state.
j. The role of timers.
If there is at least one idle CPU, and if timers are not firing, the
transition from RCU_GP_DOING_FQS to RCU_GP_CLEANUP will never happen.
Timers can fail to fire for a number of reasons, including issues in
timer configuration, issues in the timer framework, and failure to handle
softirqs (for example, when there is a storm of interrupts). Whatever the
reason, if the timers fail to fire, the GP kthread will never be awakened,
resulting in RCU CPU stall warnings and eventually in OOM.
However, an RCU CPU stall warning has a large number of potential causes,
as documented in Documentation/RCU/stallwarn.rst. This commit therefore
adds analysis to the RCU CPU stall-warning code to emit an additional
message if the cause of the stall is likely to be timer failure.
Signed-off-by: Neeraj Upadhyay <neeraju@codeaurora.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-11-17 00:06:00 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_state, RCU_GP_IDLE);
|
2018-04-12 22:20:30 +08:00
|
|
|
|
/* Check for GP requests since above loop. */
|
2018-07-04 06:37:16 +08:00
|
|
|
|
rdp = this_cpu_ptr(&rcu_data);
|
2018-05-14 11:15:40 +08:00
|
|
|
|
if (!needgp && ULONG_CMP_LT(rnp->gp_seq, rnp->gp_seq_needed)) {
|
2018-05-02 04:08:46 +08:00
|
|
|
|
trace_rcu_this_gp(rnp, rdp, rnp->gp_seq_needed,
|
2018-04-13 02:24:09 +08:00
|
|
|
|
TPS("CleanupMore"));
|
2018-04-12 22:20:30 +08:00
|
|
|
|
needgp = true;
|
|
|
|
|
}
|
rcu: Make callers awaken grace-period kthread
The rcu_start_gp_advanced() function currently uses irq_work_queue()
to defer wakeups of the RCU grace-period kthread. This deferring
is necessary to avoid RCU-scheduler deadlocks involving the rcu_node
structure's lock, meaning that RCU cannot call any of the scheduler's
wake-up functions while holding one of these locks.
Unfortunately, the second and subsequent calls to irq_work_queue() are
ignored, and the first call will be ignored (aside from queuing the work
item) if the scheduler-clock tick is turned off. This is OK for many
uses, especially those where irq_work_queue() is called from an interrupt
or softirq handler, because in those cases the scheduler-clock-tick state
will be re-evaluated, which will turn the scheduler-clock tick back on.
On the next tick, any deferred work will then be processed.
However, this strategy does not always work for RCU, which can be invoked
at process level from idle CPUs. In this case, the tick might never
be turned back on, indefinitely defering a grace-period start request.
Note that the RCU CPU stall detector cannot see this condition, because
there is no RCU grace period in progress. Therefore, we can (and do!)
see long tens-of-seconds stalls in grace-period handling. In theory,
we could see a full grace-period hang, but rcutorture testing to date
has seen only the tens-of-seconds stalls. Event tracing demonstrates
that irq_work_queue() is being called repeatedly to no effect during
these stalls: The "newreq" event appears repeatedly from a task that is
not one of the grace-period kthreads.
In theory, irq_work_queue() might be fixed to avoid this sort of issue,
but RCU's requirements are unusual and it is quite straightforward to pass
wake-up responsibility up through RCU's call chain, so that the wakeup
happens when the offending locks are released.
This commit therefore makes this change. The rcu_start_gp_advanced(),
rcu_start_future_gp(), rcu_accelerate_cbs(), rcu_advance_cbs(),
__note_gp_changes(), and rcu_start_gp() functions now return a boolean
which indicates when a wake-up is needed. A new rcu_gp_kthread_wake()
does the wakeup when it is necessary and safe to do so: No self-wakes,
no wake-ups if the ->gp_flags field indicates there is no need (as in
someone else did the wake-up before we got around to it), and no wake-ups
before the grace-period kthread has been created.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2014-03-12 04:02:16 +08:00
|
|
|
|
/* Advance CBs to reduce false positives below. */
|
2020-11-12 08:51:21 +08:00
|
|
|
|
offloaded = rcu_rdp_is_offloaded(rdp);
|
2019-05-16 00:56:40 +08:00
|
|
|
|
if ((offloaded || !rcu_accelerate_cbs(rnp, rdp)) && needgp) {
|
2022-03-03 03:01:37 +08:00
|
|
|
|
|
|
|
|
|
// We get here if a grace period was needed (“needgp”)
|
|
|
|
|
// and the above call to rcu_accelerate_cbs() did not set
|
|
|
|
|
// the RCU_GP_FLAG_INIT bit in ->gp_state (which records
|
|
|
|
|
// the need for another grace period). The purpose
|
|
|
|
|
// of the “offloaded” check is to avoid invoking
|
|
|
|
|
// rcu_accelerate_cbs() on an offloaded CPU because we do not
|
|
|
|
|
// hold the ->nocb_lock needed to safely access an offloaded
|
|
|
|
|
// ->cblist. We do not want to acquire that lock because
|
|
|
|
|
// it can be heavily contended during callback floods.
|
|
|
|
|
|
2018-07-06 06:47:01 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_flags, RCU_GP_FLAG_INIT);
|
2020-01-04 07:17:12 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_req_activity, jiffies);
|
2022-03-03 03:01:37 +08:00
|
|
|
|
trace_rcu_grace_period(rcu_state.name, rcu_state.gp_seq, TPS("newreq"));
|
2018-04-23 06:06:05 +08:00
|
|
|
|
} else {
|
2022-03-03 03:01:37 +08:00
|
|
|
|
|
|
|
|
|
// We get here either if there is no need for an
|
|
|
|
|
// additional grace period or if rcu_accelerate_cbs() has
|
|
|
|
|
// already set the RCU_GP_FLAG_INIT bit in ->gp_flags.
|
|
|
|
|
// So all we need to do is to clear all of the other
|
|
|
|
|
// ->gp_flags bits.
|
|
|
|
|
|
|
|
|
|
WRITE_ONCE(rcu_state.gp_flags, rcu_state.gp_flags & RCU_GP_FLAG_INIT);
|
2013-08-10 07:02:09 +08:00
|
|
|
|
}
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_unlock_irq_rcu_node(rnp);
|
2020-08-07 10:42:47 +08:00
|
|
|
|
|
|
|
|
|
// If strict, make all CPUs aware of the end of the old grace period.
|
|
|
|
|
if (IS_ENABLED(CONFIG_RCU_STRICT_GRACE_PERIOD))
|
|
|
|
|
on_each_cpu(rcu_strict_gp_boundary, NULL, 0);
|
2012-06-23 02:08:41 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Body of kthread that handles grace periods.
|
|
|
|
|
*/
|
2018-07-04 08:22:34 +08:00
|
|
|
|
static int __noreturn rcu_gp_kthread(void *unused)
|
2012-06-23 02:08:41 +08:00
|
|
|
|
{
|
2015-02-25 03:05:36 +08:00
|
|
|
|
rcu_bind_gp_kthread();
|
2012-06-23 02:08:41 +08:00
|
|
|
|
for (;;) {
|
|
|
|
|
|
|
|
|
|
/* Handle grace-period start. */
|
|
|
|
|
for (;;) {
|
2020-02-11 01:58:37 +08:00
|
|
|
|
trace_rcu_grace_period(rcu_state.name, rcu_state.gp_seq,
|
2013-08-10 03:19:29 +08:00
|
|
|
|
TPS("reqwait"));
|
rcu: Check and report missed fqs timer wakeup on RCU stall
For a new grace period request, the RCU GP kthread transitions through
following states:
a. [RCU_GP_WAIT_GPS] -> [RCU_GP_DONE_GPS]
The RCU_GP_WAIT_GPS state is where the GP kthread waits for a request
for a new GP. Once it receives a request (for example, when a new RCU
callback is queued), the GP kthread transitions to RCU_GP_DONE_GPS.
b. [RCU_GP_DONE_GPS] -> [RCU_GP_ONOFF]
Grace period initialization starts in rcu_gp_init(), which records the
start of new GP in rcu_state.gp_seq and transitions to RCU_GP_ONOFF.
c. [RCU_GP_ONOFF] -> [RCU_GP_INIT]
The purpose of the RCU_GP_ONOFF state is to apply the online/offline
information that was buffered for any CPUs that recently came online or
went offline. This state is maintained in per-leaf rcu_node bitmasks,
with the buffered state in ->qsmaskinitnext and the state for the upcoming
GP in ->qsmaskinit. At the end of this RCU_GP_ONOFF state, each bit in
->qsmaskinit will correspond to a CPU that must pass through a quiescent
state before the upcoming grace period is allowed to complete.
However, a leaf rcu_node structure with an all-zeroes ->qsmaskinit
cannot necessarily be ignored. In preemptible RCU, there might well be
tasks still in RCU read-side critical sections that were first preempted
while running on one of the CPUs managed by this structure. Such tasks
will be queued on this structure's ->blkd_tasks list. Only after this
list fully drains can this leaf rcu_node structure be ignored, and even
then only if none of its CPUs have come back online in the meantime.
Once that happens, the ->qsmaskinit masks further up the tree will be
updated to exclude this leaf rcu_node structure.
Once the ->qsmaskinitnext and ->qsmaskinit fields have been updated
as needed, the GP kthread transitions to RCU_GP_INIT.
d. [RCU_GP_INIT] -> [RCU_GP_WAIT_FQS]
The purpose of the RCU_GP_INIT state is to copy each ->qsmaskinit to
the ->qsmask field within each rcu_node structure. This copying is done
breadth-first from the root to the leaves. Why not just copy directly
from ->qsmaskinitnext to ->qsmask? Because the ->qsmaskinitnext masks
can change in the meantime as additional CPUs come online or go offline.
Such changes would result in inconsistencies in the ->qsmask fields up and
down the tree, which could in turn result in too-short grace periods or
grace-period hangs. These issues are avoided by snapshotting the leaf
rcu_node structures' ->qsmaskinitnext fields into their ->qsmaskinit
counterparts, generating a consistent set of ->qsmaskinit fields
throughout the tree, and only then copying these consistent ->qsmaskinit
fields to their ->qsmask counterparts.
Once this initialization step is complete, the GP kthread transitions
to RCU_GP_WAIT_FQS, where it waits to do a force-quiescent-state scan
on the one hand or for the end of the grace period on the other.
e. [RCU_GP_WAIT_FQS] -> [RCU_GP_DOING_FQS]
The RCU_GP_WAIT_FQS state waits for one of three things: (1) An
explicit request to do a force-quiescent-state scan, (2) The end of
the grace period, or (3) A short interval of time, after which it
will do a force-quiescent-state (FQS) scan. The explicit request can
come from rcutorture or from any CPU that has too many RCU callbacks
queued (see the qhimark kernel parameter and the RCU_GP_FLAG_OVLD
flag). The aforementioned "short period of time" is specified by the
jiffies_till_first_fqs boot parameter for a given grace period's first
FQS scan and by the jiffies_till_next_fqs for later FQS scans.
Either way, once the wait is over, the GP kthread transitions to
RCU_GP_DOING_FQS.
f. [RCU_GP_DOING_FQS] -> [RCU_GP_CLEANUP]
The RCU_GP_DOING_FQS state performs an FQS scan. Each such scan carries
out two functions for any CPU whose bit is still set in its leaf rcu_node
structure's ->qsmask field, that is, for any CPU that has not yet reported
a quiescent state for the current grace period:
i. Report quiescent states on behalf of CPUs that have been observed
to be idle (from an RCU perspective) since the beginning of the
grace period.
ii. If the current grace period is too old, take various actions to
encourage holdout CPUs to pass through quiescent states, including
enlisting the aid of any calls to cond_resched() and might_sleep(),
and even including IPIing the holdout CPUs.
These checks are skipped for any leaf rcu_node structure with a all-zero
->qsmask field, however such structures are subject to RCU priority
boosting if there are tasks on a given structure blocking the current
grace period. The end of the grace period is detected when the root
rcu_node structure's ->qsmask is zero and when there are no longer any
preempted tasks blocking the current grace period. (No, this last check
is not redundant. To see this, consider an rcu_node tree having exactly
one structure that serves as both root and leaf.)
Once the end of the grace period is detected, the GP kthread transitions
to RCU_GP_CLEANUP.
g. [RCU_GP_CLEANUP] -> [RCU_GP_CLEANED]
The RCU_GP_CLEANUP state marks the end of grace period by updating the
rcu_state structure's ->gp_seq field and also all rcu_node structures'
->gp_seq field. As before, the rcu_node tree is traversed in breadth
first order. Once this update is complete, the GP kthread transitions
to the RCU_GP_CLEANED state.
i. [RCU_GP_CLEANED] -> [RCU_GP_INIT]
Once in the RCU_GP_CLEANED state, the GP kthread immediately transitions
into the RCU_GP_INIT state.
j. The role of timers.
If there is at least one idle CPU, and if timers are not firing, the
transition from RCU_GP_DOING_FQS to RCU_GP_CLEANUP will never happen.
Timers can fail to fire for a number of reasons, including issues in
timer configuration, issues in the timer framework, and failure to handle
softirqs (for example, when there is a storm of interrupts). Whatever the
reason, if the timers fail to fire, the GP kthread will never be awakened,
resulting in RCU CPU stall warnings and eventually in OOM.
However, an RCU CPU stall warning has a large number of potential causes,
as documented in Documentation/RCU/stallwarn.rst. This commit therefore
adds analysis to the RCU CPU stall-warning code to emit an additional
message if the cause of the stall is likely to be timer failure.
Signed-off-by: Neeraj Upadhyay <neeraju@codeaurora.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-11-17 00:06:00 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_state, RCU_GP_WAIT_GPS);
|
2018-07-06 06:47:01 +08:00
|
|
|
|
swait_event_idle_exclusive(rcu_state.gp_wq,
|
|
|
|
|
READ_ONCE(rcu_state.gp_flags) &
|
|
|
|
|
RCU_GP_FLAG_INIT);
|
2020-04-02 10:57:52 +08:00
|
|
|
|
rcu_gp_torture_wait();
|
rcu: Check and report missed fqs timer wakeup on RCU stall
For a new grace period request, the RCU GP kthread transitions through
following states:
a. [RCU_GP_WAIT_GPS] -> [RCU_GP_DONE_GPS]
The RCU_GP_WAIT_GPS state is where the GP kthread waits for a request
for a new GP. Once it receives a request (for example, when a new RCU
callback is queued), the GP kthread transitions to RCU_GP_DONE_GPS.
b. [RCU_GP_DONE_GPS] -> [RCU_GP_ONOFF]
Grace period initialization starts in rcu_gp_init(), which records the
start of new GP in rcu_state.gp_seq and transitions to RCU_GP_ONOFF.
c. [RCU_GP_ONOFF] -> [RCU_GP_INIT]
The purpose of the RCU_GP_ONOFF state is to apply the online/offline
information that was buffered for any CPUs that recently came online or
went offline. This state is maintained in per-leaf rcu_node bitmasks,
with the buffered state in ->qsmaskinitnext and the state for the upcoming
GP in ->qsmaskinit. At the end of this RCU_GP_ONOFF state, each bit in
->qsmaskinit will correspond to a CPU that must pass through a quiescent
state before the upcoming grace period is allowed to complete.
However, a leaf rcu_node structure with an all-zeroes ->qsmaskinit
cannot necessarily be ignored. In preemptible RCU, there might well be
tasks still in RCU read-side critical sections that were first preempted
while running on one of the CPUs managed by this structure. Such tasks
will be queued on this structure's ->blkd_tasks list. Only after this
list fully drains can this leaf rcu_node structure be ignored, and even
then only if none of its CPUs have come back online in the meantime.
Once that happens, the ->qsmaskinit masks further up the tree will be
updated to exclude this leaf rcu_node structure.
Once the ->qsmaskinitnext and ->qsmaskinit fields have been updated
as needed, the GP kthread transitions to RCU_GP_INIT.
d. [RCU_GP_INIT] -> [RCU_GP_WAIT_FQS]
The purpose of the RCU_GP_INIT state is to copy each ->qsmaskinit to
the ->qsmask field within each rcu_node structure. This copying is done
breadth-first from the root to the leaves. Why not just copy directly
from ->qsmaskinitnext to ->qsmask? Because the ->qsmaskinitnext masks
can change in the meantime as additional CPUs come online or go offline.
Such changes would result in inconsistencies in the ->qsmask fields up and
down the tree, which could in turn result in too-short grace periods or
grace-period hangs. These issues are avoided by snapshotting the leaf
rcu_node structures' ->qsmaskinitnext fields into their ->qsmaskinit
counterparts, generating a consistent set of ->qsmaskinit fields
throughout the tree, and only then copying these consistent ->qsmaskinit
fields to their ->qsmask counterparts.
Once this initialization step is complete, the GP kthread transitions
to RCU_GP_WAIT_FQS, where it waits to do a force-quiescent-state scan
on the one hand or for the end of the grace period on the other.
e. [RCU_GP_WAIT_FQS] -> [RCU_GP_DOING_FQS]
The RCU_GP_WAIT_FQS state waits for one of three things: (1) An
explicit request to do a force-quiescent-state scan, (2) The end of
the grace period, or (3) A short interval of time, after which it
will do a force-quiescent-state (FQS) scan. The explicit request can
come from rcutorture or from any CPU that has too many RCU callbacks
queued (see the qhimark kernel parameter and the RCU_GP_FLAG_OVLD
flag). The aforementioned "short period of time" is specified by the
jiffies_till_first_fqs boot parameter for a given grace period's first
FQS scan and by the jiffies_till_next_fqs for later FQS scans.
Either way, once the wait is over, the GP kthread transitions to
RCU_GP_DOING_FQS.
f. [RCU_GP_DOING_FQS] -> [RCU_GP_CLEANUP]
The RCU_GP_DOING_FQS state performs an FQS scan. Each such scan carries
out two functions for any CPU whose bit is still set in its leaf rcu_node
structure's ->qsmask field, that is, for any CPU that has not yet reported
a quiescent state for the current grace period:
i. Report quiescent states on behalf of CPUs that have been observed
to be idle (from an RCU perspective) since the beginning of the
grace period.
ii. If the current grace period is too old, take various actions to
encourage holdout CPUs to pass through quiescent states, including
enlisting the aid of any calls to cond_resched() and might_sleep(),
and even including IPIing the holdout CPUs.
These checks are skipped for any leaf rcu_node structure with a all-zero
->qsmask field, however such structures are subject to RCU priority
boosting if there are tasks on a given structure blocking the current
grace period. The end of the grace period is detected when the root
rcu_node structure's ->qsmask is zero and when there are no longer any
preempted tasks blocking the current grace period. (No, this last check
is not redundant. To see this, consider an rcu_node tree having exactly
one structure that serves as both root and leaf.)
Once the end of the grace period is detected, the GP kthread transitions
to RCU_GP_CLEANUP.
g. [RCU_GP_CLEANUP] -> [RCU_GP_CLEANED]
The RCU_GP_CLEANUP state marks the end of grace period by updating the
rcu_state structure's ->gp_seq field and also all rcu_node structures'
->gp_seq field. As before, the rcu_node tree is traversed in breadth
first order. Once this update is complete, the GP kthread transitions
to the RCU_GP_CLEANED state.
i. [RCU_GP_CLEANED] -> [RCU_GP_INIT]
Once in the RCU_GP_CLEANED state, the GP kthread immediately transitions
into the RCU_GP_INIT state.
j. The role of timers.
If there is at least one idle CPU, and if timers are not firing, the
transition from RCU_GP_DOING_FQS to RCU_GP_CLEANUP will never happen.
Timers can fail to fire for a number of reasons, including issues in
timer configuration, issues in the timer framework, and failure to handle
softirqs (for example, when there is a storm of interrupts). Whatever the
reason, if the timers fail to fire, the GP kthread will never be awakened,
resulting in RCU CPU stall warnings and eventually in OOM.
However, an RCU CPU stall warning has a large number of potential causes,
as documented in Documentation/RCU/stallwarn.rst. This commit therefore
adds analysis to the RCU CPU stall-warning code to emit an additional
message if the cause of the stall is likely to be timer failure.
Signed-off-by: Neeraj Upadhyay <neeraju@codeaurora.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-11-17 00:06:00 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_state, RCU_GP_DONE_GPS);
|
2013-09-25 06:04:06 +08:00
|
|
|
|
/* Locking provides needed memory barrier. */
|
2018-07-04 08:22:34 +08:00
|
|
|
|
if (rcu_gp_init())
|
2012-06-23 02:08:41 +08:00
|
|
|
|
break;
|
2018-03-03 08:35:27 +08:00
|
|
|
|
cond_resched_tasks_rcu_qs();
|
2018-07-06 06:47:01 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_activity, jiffies);
|
2014-08-15 01:28:23 +08:00
|
|
|
|
WARN_ON(signal_pending(current));
|
2020-02-11 01:58:37 +08:00
|
|
|
|
trace_rcu_grace_period(rcu_state.name, rcu_state.gp_seq,
|
2013-08-10 03:19:29 +08:00
|
|
|
|
TPS("reqwaitsig"));
|
2012-06-23 02:08:41 +08:00
|
|
|
|
}
|
2012-06-21 08:07:14 +08:00
|
|
|
|
|
2012-06-23 08:06:26 +08:00
|
|
|
|
/* Handle quiescent-state forcing. */
|
2018-07-06 09:23:23 +08:00
|
|
|
|
rcu_gp_fqs_loop();
|
2012-06-23 08:06:26 +08:00
|
|
|
|
|
|
|
|
|
/* Handle grace-period end. */
|
rcu: Check and report missed fqs timer wakeup on RCU stall
For a new grace period request, the RCU GP kthread transitions through
following states:
a. [RCU_GP_WAIT_GPS] -> [RCU_GP_DONE_GPS]
The RCU_GP_WAIT_GPS state is where the GP kthread waits for a request
for a new GP. Once it receives a request (for example, when a new RCU
callback is queued), the GP kthread transitions to RCU_GP_DONE_GPS.
b. [RCU_GP_DONE_GPS] -> [RCU_GP_ONOFF]
Grace period initialization starts in rcu_gp_init(), which records the
start of new GP in rcu_state.gp_seq and transitions to RCU_GP_ONOFF.
c. [RCU_GP_ONOFF] -> [RCU_GP_INIT]
The purpose of the RCU_GP_ONOFF state is to apply the online/offline
information that was buffered for any CPUs that recently came online or
went offline. This state is maintained in per-leaf rcu_node bitmasks,
with the buffered state in ->qsmaskinitnext and the state for the upcoming
GP in ->qsmaskinit. At the end of this RCU_GP_ONOFF state, each bit in
->qsmaskinit will correspond to a CPU that must pass through a quiescent
state before the upcoming grace period is allowed to complete.
However, a leaf rcu_node structure with an all-zeroes ->qsmaskinit
cannot necessarily be ignored. In preemptible RCU, there might well be
tasks still in RCU read-side critical sections that were first preempted
while running on one of the CPUs managed by this structure. Such tasks
will be queued on this structure's ->blkd_tasks list. Only after this
list fully drains can this leaf rcu_node structure be ignored, and even
then only if none of its CPUs have come back online in the meantime.
Once that happens, the ->qsmaskinit masks further up the tree will be
updated to exclude this leaf rcu_node structure.
Once the ->qsmaskinitnext and ->qsmaskinit fields have been updated
as needed, the GP kthread transitions to RCU_GP_INIT.
d. [RCU_GP_INIT] -> [RCU_GP_WAIT_FQS]
The purpose of the RCU_GP_INIT state is to copy each ->qsmaskinit to
the ->qsmask field within each rcu_node structure. This copying is done
breadth-first from the root to the leaves. Why not just copy directly
from ->qsmaskinitnext to ->qsmask? Because the ->qsmaskinitnext masks
can change in the meantime as additional CPUs come online or go offline.
Such changes would result in inconsistencies in the ->qsmask fields up and
down the tree, which could in turn result in too-short grace periods or
grace-period hangs. These issues are avoided by snapshotting the leaf
rcu_node structures' ->qsmaskinitnext fields into their ->qsmaskinit
counterparts, generating a consistent set of ->qsmaskinit fields
throughout the tree, and only then copying these consistent ->qsmaskinit
fields to their ->qsmask counterparts.
Once this initialization step is complete, the GP kthread transitions
to RCU_GP_WAIT_FQS, where it waits to do a force-quiescent-state scan
on the one hand or for the end of the grace period on the other.
e. [RCU_GP_WAIT_FQS] -> [RCU_GP_DOING_FQS]
The RCU_GP_WAIT_FQS state waits for one of three things: (1) An
explicit request to do a force-quiescent-state scan, (2) The end of
the grace period, or (3) A short interval of time, after which it
will do a force-quiescent-state (FQS) scan. The explicit request can
come from rcutorture or from any CPU that has too many RCU callbacks
queued (see the qhimark kernel parameter and the RCU_GP_FLAG_OVLD
flag). The aforementioned "short period of time" is specified by the
jiffies_till_first_fqs boot parameter for a given grace period's first
FQS scan and by the jiffies_till_next_fqs for later FQS scans.
Either way, once the wait is over, the GP kthread transitions to
RCU_GP_DOING_FQS.
f. [RCU_GP_DOING_FQS] -> [RCU_GP_CLEANUP]
The RCU_GP_DOING_FQS state performs an FQS scan. Each such scan carries
out two functions for any CPU whose bit is still set in its leaf rcu_node
structure's ->qsmask field, that is, for any CPU that has not yet reported
a quiescent state for the current grace period:
i. Report quiescent states on behalf of CPUs that have been observed
to be idle (from an RCU perspective) since the beginning of the
grace period.
ii. If the current grace period is too old, take various actions to
encourage holdout CPUs to pass through quiescent states, including
enlisting the aid of any calls to cond_resched() and might_sleep(),
and even including IPIing the holdout CPUs.
These checks are skipped for any leaf rcu_node structure with a all-zero
->qsmask field, however such structures are subject to RCU priority
boosting if there are tasks on a given structure blocking the current
grace period. The end of the grace period is detected when the root
rcu_node structure's ->qsmask is zero and when there are no longer any
preempted tasks blocking the current grace period. (No, this last check
is not redundant. To see this, consider an rcu_node tree having exactly
one structure that serves as both root and leaf.)
Once the end of the grace period is detected, the GP kthread transitions
to RCU_GP_CLEANUP.
g. [RCU_GP_CLEANUP] -> [RCU_GP_CLEANED]
The RCU_GP_CLEANUP state marks the end of grace period by updating the
rcu_state structure's ->gp_seq field and also all rcu_node structures'
->gp_seq field. As before, the rcu_node tree is traversed in breadth
first order. Once this update is complete, the GP kthread transitions
to the RCU_GP_CLEANED state.
i. [RCU_GP_CLEANED] -> [RCU_GP_INIT]
Once in the RCU_GP_CLEANED state, the GP kthread immediately transitions
into the RCU_GP_INIT state.
j. The role of timers.
If there is at least one idle CPU, and if timers are not firing, the
transition from RCU_GP_DOING_FQS to RCU_GP_CLEANUP will never happen.
Timers can fail to fire for a number of reasons, including issues in
timer configuration, issues in the timer framework, and failure to handle
softirqs (for example, when there is a storm of interrupts). Whatever the
reason, if the timers fail to fire, the GP kthread will never be awakened,
resulting in RCU CPU stall warnings and eventually in OOM.
However, an RCU CPU stall warning has a large number of potential causes,
as documented in Documentation/RCU/stallwarn.rst. This commit therefore
adds analysis to the RCU CPU stall-warning code to emit an additional
message if the cause of the stall is likely to be timer failure.
Signed-off-by: Neeraj Upadhyay <neeraju@codeaurora.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-11-17 00:06:00 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_state, RCU_GP_CLEANUP);
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_gp_cleanup();
|
rcu: Check and report missed fqs timer wakeup on RCU stall
For a new grace period request, the RCU GP kthread transitions through
following states:
a. [RCU_GP_WAIT_GPS] -> [RCU_GP_DONE_GPS]
The RCU_GP_WAIT_GPS state is where the GP kthread waits for a request
for a new GP. Once it receives a request (for example, when a new RCU
callback is queued), the GP kthread transitions to RCU_GP_DONE_GPS.
b. [RCU_GP_DONE_GPS] -> [RCU_GP_ONOFF]
Grace period initialization starts in rcu_gp_init(), which records the
start of new GP in rcu_state.gp_seq and transitions to RCU_GP_ONOFF.
c. [RCU_GP_ONOFF] -> [RCU_GP_INIT]
The purpose of the RCU_GP_ONOFF state is to apply the online/offline
information that was buffered for any CPUs that recently came online or
went offline. This state is maintained in per-leaf rcu_node bitmasks,
with the buffered state in ->qsmaskinitnext and the state for the upcoming
GP in ->qsmaskinit. At the end of this RCU_GP_ONOFF state, each bit in
->qsmaskinit will correspond to a CPU that must pass through a quiescent
state before the upcoming grace period is allowed to complete.
However, a leaf rcu_node structure with an all-zeroes ->qsmaskinit
cannot necessarily be ignored. In preemptible RCU, there might well be
tasks still in RCU read-side critical sections that were first preempted
while running on one of the CPUs managed by this structure. Such tasks
will be queued on this structure's ->blkd_tasks list. Only after this
list fully drains can this leaf rcu_node structure be ignored, and even
then only if none of its CPUs have come back online in the meantime.
Once that happens, the ->qsmaskinit masks further up the tree will be
updated to exclude this leaf rcu_node structure.
Once the ->qsmaskinitnext and ->qsmaskinit fields have been updated
as needed, the GP kthread transitions to RCU_GP_INIT.
d. [RCU_GP_INIT] -> [RCU_GP_WAIT_FQS]
The purpose of the RCU_GP_INIT state is to copy each ->qsmaskinit to
the ->qsmask field within each rcu_node structure. This copying is done
breadth-first from the root to the leaves. Why not just copy directly
from ->qsmaskinitnext to ->qsmask? Because the ->qsmaskinitnext masks
can change in the meantime as additional CPUs come online or go offline.
Such changes would result in inconsistencies in the ->qsmask fields up and
down the tree, which could in turn result in too-short grace periods or
grace-period hangs. These issues are avoided by snapshotting the leaf
rcu_node structures' ->qsmaskinitnext fields into their ->qsmaskinit
counterparts, generating a consistent set of ->qsmaskinit fields
throughout the tree, and only then copying these consistent ->qsmaskinit
fields to their ->qsmask counterparts.
Once this initialization step is complete, the GP kthread transitions
to RCU_GP_WAIT_FQS, where it waits to do a force-quiescent-state scan
on the one hand or for the end of the grace period on the other.
e. [RCU_GP_WAIT_FQS] -> [RCU_GP_DOING_FQS]
The RCU_GP_WAIT_FQS state waits for one of three things: (1) An
explicit request to do a force-quiescent-state scan, (2) The end of
the grace period, or (3) A short interval of time, after which it
will do a force-quiescent-state (FQS) scan. The explicit request can
come from rcutorture or from any CPU that has too many RCU callbacks
queued (see the qhimark kernel parameter and the RCU_GP_FLAG_OVLD
flag). The aforementioned "short period of time" is specified by the
jiffies_till_first_fqs boot parameter for a given grace period's first
FQS scan and by the jiffies_till_next_fqs for later FQS scans.
Either way, once the wait is over, the GP kthread transitions to
RCU_GP_DOING_FQS.
f. [RCU_GP_DOING_FQS] -> [RCU_GP_CLEANUP]
The RCU_GP_DOING_FQS state performs an FQS scan. Each such scan carries
out two functions for any CPU whose bit is still set in its leaf rcu_node
structure's ->qsmask field, that is, for any CPU that has not yet reported
a quiescent state for the current grace period:
i. Report quiescent states on behalf of CPUs that have been observed
to be idle (from an RCU perspective) since the beginning of the
grace period.
ii. If the current grace period is too old, take various actions to
encourage holdout CPUs to pass through quiescent states, including
enlisting the aid of any calls to cond_resched() and might_sleep(),
and even including IPIing the holdout CPUs.
These checks are skipped for any leaf rcu_node structure with a all-zero
->qsmask field, however such structures are subject to RCU priority
boosting if there are tasks on a given structure blocking the current
grace period. The end of the grace period is detected when the root
rcu_node structure's ->qsmask is zero and when there are no longer any
preempted tasks blocking the current grace period. (No, this last check
is not redundant. To see this, consider an rcu_node tree having exactly
one structure that serves as both root and leaf.)
Once the end of the grace period is detected, the GP kthread transitions
to RCU_GP_CLEANUP.
g. [RCU_GP_CLEANUP] -> [RCU_GP_CLEANED]
The RCU_GP_CLEANUP state marks the end of grace period by updating the
rcu_state structure's ->gp_seq field and also all rcu_node structures'
->gp_seq field. As before, the rcu_node tree is traversed in breadth
first order. Once this update is complete, the GP kthread transitions
to the RCU_GP_CLEANED state.
i. [RCU_GP_CLEANED] -> [RCU_GP_INIT]
Once in the RCU_GP_CLEANED state, the GP kthread immediately transitions
into the RCU_GP_INIT state.
j. The role of timers.
If there is at least one idle CPU, and if timers are not firing, the
transition from RCU_GP_DOING_FQS to RCU_GP_CLEANUP will never happen.
Timers can fail to fire for a number of reasons, including issues in
timer configuration, issues in the timer framework, and failure to handle
softirqs (for example, when there is a storm of interrupts). Whatever the
reason, if the timers fail to fire, the GP kthread will never be awakened,
resulting in RCU CPU stall warnings and eventually in OOM.
However, an RCU CPU stall warning has a large number of potential causes,
as documented in Documentation/RCU/stallwarn.rst. This commit therefore
adds analysis to the RCU CPU stall-warning code to emit an additional
message if the cause of the stall is likely to be timer failure.
Signed-off-by: Neeraj Upadhyay <neeraju@codeaurora.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-11-17 00:06:00 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_state, RCU_GP_CLEANED);
|
2012-06-19 09:36:08 +08:00
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
rcu: Merge preemptable-RCU functionality into hierarchical RCU
Create a kernel/rcutree_plugin.h file that contains definitions
for preemptable RCU (or, under the #else branch of the #ifdef,
empty definitions for the classic non-preemptable semantics).
These definitions fit into plugins defined in kernel/rcutree.c
for this purpose.
This variant of preemptable RCU uses a new algorithm whose
read-side expense is roughly that of classic hierarchical RCU
under CONFIG_PREEMPT. This new algorithm's update-side expense
is similar to that of classic hierarchical RCU, and, in absence
of read-side preemption or blocking, is exactly that of classic
hierarchical RCU. Perhaps more important, this new algorithm
has a much simpler implementation, saving well over 1,000 lines
of code compared to mainline's implementation of preemptable
RCU, which will hopefully be retired in favor of this new
algorithm.
The simplifications are obtained by maintaining per-task
nesting state for running tasks, and using a simple
lock-protected algorithm to handle accounting when tasks block
within RCU read-side critical sections, making use of lessons
learned while creating numerous user-level RCU implementations
over the past 18 months.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: laijs@cn.fujitsu.com
Cc: dipankar@in.ibm.com
Cc: akpm@linux-foundation.org
Cc: mathieu.desnoyers@polymtl.ca
Cc: josht@linux.vnet.ibm.com
Cc: dvhltc@us.ibm.com
Cc: niv@us.ibm.com
Cc: peterz@infradead.org
Cc: rostedt@goodmis.org
LKML-Reference: <12509746134003-git-send-email->
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-08-23 04:56:52 +08:00
|
|
|
|
/*
|
2018-07-08 09:12:26 +08:00
|
|
|
|
* Report a full set of quiescent states to the rcu_state data structure.
|
|
|
|
|
* Invoke rcu_gp_kthread_wake() to awaken the grace-period kthread if
|
|
|
|
|
* another grace period is required. Whether we wake the grace-period
|
|
|
|
|
* kthread or it awakens itself for the next round of quiescent-state
|
|
|
|
|
* forcing, that kthread will clean up after the just-completed grace
|
|
|
|
|
* period. Note that the caller must hold rnp->lock, which is released
|
|
|
|
|
* before return.
|
rcu: Merge preemptable-RCU functionality into hierarchical RCU
Create a kernel/rcutree_plugin.h file that contains definitions
for preemptable RCU (or, under the #else branch of the #ifdef,
empty definitions for the classic non-preemptable semantics).
These definitions fit into plugins defined in kernel/rcutree.c
for this purpose.
This variant of preemptable RCU uses a new algorithm whose
read-side expense is roughly that of classic hierarchical RCU
under CONFIG_PREEMPT. This new algorithm's update-side expense
is similar to that of classic hierarchical RCU, and, in absence
of read-side preemption or blocking, is exactly that of classic
hierarchical RCU. Perhaps more important, this new algorithm
has a much simpler implementation, saving well over 1,000 lines
of code compared to mainline's implementation of preemptable
RCU, which will hopefully be retired in favor of this new
algorithm.
The simplifications are obtained by maintaining per-task
nesting state for running tasks, and using a simple
lock-protected algorithm to handle accounting when tasks block
within RCU read-side critical sections, making use of lessons
learned while creating numerous user-level RCU implementations
over the past 18 months.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: laijs@cn.fujitsu.com
Cc: dipankar@in.ibm.com
Cc: akpm@linux-foundation.org
Cc: mathieu.desnoyers@polymtl.ca
Cc: josht@linux.vnet.ibm.com
Cc: dvhltc@us.ibm.com
Cc: niv@us.ibm.com
Cc: peterz@infradead.org
Cc: rostedt@goodmis.org
LKML-Reference: <12509746134003-git-send-email->
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-08-23 04:56:52 +08:00
|
|
|
|
*/
|
2018-07-04 08:22:34 +08:00
|
|
|
|
static void rcu_report_qs_rsp(unsigned long flags)
|
2018-07-04 08:22:34 +08:00
|
|
|
|
__releases(rcu_get_root()->lock)
|
rcu: Merge preemptable-RCU functionality into hierarchical RCU
Create a kernel/rcutree_plugin.h file that contains definitions
for preemptable RCU (or, under the #else branch of the #ifdef,
empty definitions for the classic non-preemptable semantics).
These definitions fit into plugins defined in kernel/rcutree.c
for this purpose.
This variant of preemptable RCU uses a new algorithm whose
read-side expense is roughly that of classic hierarchical RCU
under CONFIG_PREEMPT. This new algorithm's update-side expense
is similar to that of classic hierarchical RCU, and, in absence
of read-side preemption or blocking, is exactly that of classic
hierarchical RCU. Perhaps more important, this new algorithm
has a much simpler implementation, saving well over 1,000 lines
of code compared to mainline's implementation of preemptable
RCU, which will hopefully be retired in favor of this new
algorithm.
The simplifications are obtained by maintaining per-task
nesting state for running tasks, and using a simple
lock-protected algorithm to handle accounting when tasks block
within RCU read-side critical sections, making use of lessons
learned while creating numerous user-level RCU implementations
over the past 18 months.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: laijs@cn.fujitsu.com
Cc: dipankar@in.ibm.com
Cc: akpm@linux-foundation.org
Cc: mathieu.desnoyers@polymtl.ca
Cc: josht@linux.vnet.ibm.com
Cc: dvhltc@us.ibm.com
Cc: niv@us.ibm.com
Cc: peterz@infradead.org
Cc: rostedt@goodmis.org
LKML-Reference: <12509746134003-git-send-email->
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-08-23 04:56:52 +08:00
|
|
|
|
{
|
2018-07-04 08:22:34 +08:00
|
|
|
|
raw_lockdep_assert_held_rcu_node(rcu_get_root());
|
2018-07-04 08:22:34 +08:00
|
|
|
|
WARN_ON_ONCE(!rcu_gp_in_progress());
|
2018-07-06 06:47:01 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_flags,
|
|
|
|
|
READ_ONCE(rcu_state.gp_flags) | RCU_GP_FLAG_FQS);
|
2018-07-04 08:22:34 +08:00
|
|
|
|
raw_spin_unlock_irqrestore_rcu_node(rcu_get_root(), flags);
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_gp_kthread_wake();
|
rcu: Merge preemptable-RCU functionality into hierarchical RCU
Create a kernel/rcutree_plugin.h file that contains definitions
for preemptable RCU (or, under the #else branch of the #ifdef,
empty definitions for the classic non-preemptable semantics).
These definitions fit into plugins defined in kernel/rcutree.c
for this purpose.
This variant of preemptable RCU uses a new algorithm whose
read-side expense is roughly that of classic hierarchical RCU
under CONFIG_PREEMPT. This new algorithm's update-side expense
is similar to that of classic hierarchical RCU, and, in absence
of read-side preemption or blocking, is exactly that of classic
hierarchical RCU. Perhaps more important, this new algorithm
has a much simpler implementation, saving well over 1,000 lines
of code compared to mainline's implementation of preemptable
RCU, which will hopefully be retired in favor of this new
algorithm.
The simplifications are obtained by maintaining per-task
nesting state for running tasks, and using a simple
lock-protected algorithm to handle accounting when tasks block
within RCU read-side critical sections, making use of lessons
learned while creating numerous user-level RCU implementations
over the past 18 months.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: laijs@cn.fujitsu.com
Cc: dipankar@in.ibm.com
Cc: akpm@linux-foundation.org
Cc: mathieu.desnoyers@polymtl.ca
Cc: josht@linux.vnet.ibm.com
Cc: dvhltc@us.ibm.com
Cc: niv@us.ibm.com
Cc: peterz@infradead.org
Cc: rostedt@goodmis.org
LKML-Reference: <12509746134003-git-send-email->
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-08-23 04:56:52 +08:00
|
|
|
|
}
|
|
|
|
|
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
/*
|
2009-12-03 04:10:13 +08:00
|
|
|
|
* Similar to rcu_report_qs_rdp(), for which it is a helper function.
|
|
|
|
|
* Allows quiescent states for a group of CPUs to be reported at one go
|
|
|
|
|
* to the specified rcu_node structure, though all the CPUs in the group
|
rcu: Associate quiescent-state reports with grace period
As noted in earlier commit logs, CPU hotplug operations running
concurrently with grace-period initialization can result in a given
leaf rcu_node structure having all CPUs offline and no blocked readers,
but with this rcu_node structure nevertheless blocking the current
grace period. Therefore, the quiescent-state forcing code now checks
for this situation and repairs it.
Unfortunately, this checking can result in false positives, for example,
when the last task has just removed itself from this leaf rcu_node
structure, but has not yet started clearing the ->qsmask bits further
up the structure. This means that the grace-period kthread (which
forces quiescent states) and some other task might be attempting to
concurrently clear these ->qsmask bits. This is usually not a problem:
One of these tasks will be the first to acquire the upper-level rcu_node
structure's lock and with therefore clear the bit, and the other task,
seeing the bit already cleared, will stop trying to clear bits.
Sadly, this means that the following unusual sequence of events -can-
result in a problem:
1. The grace-period kthread wins, and clears the ->qsmask bits.
2. This is the last thing blocking the current grace period, so
that the grace-period kthread clears ->qsmask bits all the way
to the root and finds that the root ->qsmask field is now zero.
3. Another grace period is required, so that the grace period kthread
initializes it, including setting all the needed qsmask bits.
4. The leaf rcu_node structure (the one that started this whole
mess) is blocking this new grace period, either because it
has at least one online CPU or because there is at least one
task that had blocked within an RCU read-side critical section
while running on one of this leaf rcu_node structure's CPUs.
(And yes, that CPU might well have gone offline before the
grace period in step (3) above started, which can mean that
there is a task on the leaf rcu_node structure's ->blkd_tasks
list, but ->qsmask equal to zero.)
5. The other kthread didn't get around to trying to clear the upper
level ->qsmask bits until all the above had happened. This means
that it now sees bits set in the upper-level ->qsmask field, so it
proceeds to clear them. Too bad that it is doing so on behalf of
a quiescent state that does not apply to the current grace period!
This sequence of events can result in the new grace period being too
short. It can also result in the new grace period ending before the
leaf rcu_node structure's ->qsmask bits have been cleared, which will
result in splats during initialization of the next grace period. In
addition, it can result in tasks blocking the new grace period still
being queued at the start of the next grace period, which will result
in other splats. Sasha's testing turned up another of these splats,
as did rcutorture testing. (And yes, rcutorture is being adjusted to
make these splats show up more quickly. Which probably is having the
undesirable side effect of making other problems show up less quickly.
Can't have everything!)
Reported-by: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: <stable@vger.kernel.org> # 4.0.x
Tested-by: Sasha Levin <sasha.levin@oracle.com>
2015-03-16 00:19:35 +08:00
|
|
|
|
* must be represented by the same rcu_node structure (which need not be a
|
|
|
|
|
* leaf rcu_node structure, though it often will be). The gps parameter
|
|
|
|
|
* is the grace-period snapshot, which means that the quiescent states
|
2018-04-28 05:54:46 +08:00
|
|
|
|
* are valid only if rnp->gp_seq is equal to gps. That structure's lock
|
rcu: Associate quiescent-state reports with grace period
As noted in earlier commit logs, CPU hotplug operations running
concurrently with grace-period initialization can result in a given
leaf rcu_node structure having all CPUs offline and no blocked readers,
but with this rcu_node structure nevertheless blocking the current
grace period. Therefore, the quiescent-state forcing code now checks
for this situation and repairs it.
Unfortunately, this checking can result in false positives, for example,
when the last task has just removed itself from this leaf rcu_node
structure, but has not yet started clearing the ->qsmask bits further
up the structure. This means that the grace-period kthread (which
forces quiescent states) and some other task might be attempting to
concurrently clear these ->qsmask bits. This is usually not a problem:
One of these tasks will be the first to acquire the upper-level rcu_node
structure's lock and with therefore clear the bit, and the other task,
seeing the bit already cleared, will stop trying to clear bits.
Sadly, this means that the following unusual sequence of events -can-
result in a problem:
1. The grace-period kthread wins, and clears the ->qsmask bits.
2. This is the last thing blocking the current grace period, so
that the grace-period kthread clears ->qsmask bits all the way
to the root and finds that the root ->qsmask field is now zero.
3. Another grace period is required, so that the grace period kthread
initializes it, including setting all the needed qsmask bits.
4. The leaf rcu_node structure (the one that started this whole
mess) is blocking this new grace period, either because it
has at least one online CPU or because there is at least one
task that had blocked within an RCU read-side critical section
while running on one of this leaf rcu_node structure's CPUs.
(And yes, that CPU might well have gone offline before the
grace period in step (3) above started, which can mean that
there is a task on the leaf rcu_node structure's ->blkd_tasks
list, but ->qsmask equal to zero.)
5. The other kthread didn't get around to trying to clear the upper
level ->qsmask bits until all the above had happened. This means
that it now sees bits set in the upper-level ->qsmask field, so it
proceeds to clear them. Too bad that it is doing so on behalf of
a quiescent state that does not apply to the current grace period!
This sequence of events can result in the new grace period being too
short. It can also result in the new grace period ending before the
leaf rcu_node structure's ->qsmask bits have been cleared, which will
result in splats during initialization of the next grace period. In
addition, it can result in tasks blocking the new grace period still
being queued at the start of the next grace period, which will result
in other splats. Sasha's testing turned up another of these splats,
as did rcutorture testing. (And yes, rcutorture is being adjusted to
make these splats show up more quickly. Which probably is having the
undesirable side effect of making other problems show up less quickly.
Can't have everything!)
Reported-by: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: <stable@vger.kernel.org> # 4.0.x
Tested-by: Sasha Levin <sasha.levin@oracle.com>
2015-03-16 00:19:35 +08:00
|
|
|
|
* must be held upon entry, and it is released before return.
|
rcu: Fix grace-period hangs from mid-init task resume
Without special fail-safe quiescent-state-propagation checks, grace-period
hangs can result from the following scenario:
1. A task running on a given CPU is preempted in its RCU read-side
critical section.
2. That CPU goes offline, and there are now no online CPUs
corresponding to that CPU's leaf rcu_node structure.
3. The rcu_gp_init() function does the first phase of grace-period
initialization, and sets the aforementioned leaf rcu_node
structure's ->qsmaskinit field to all zeroes. Because there
is a blocked task, it does not propagate the zeroing of either
->qsmaskinit or ->qsmaskinitnext up the rcu_node tree.
4. The task resumes on some other CPU and exits its critical section.
There is no grace period in progress, so the resulting quiescent
state is not reported up the tree.
5. The rcu_gp_init() function does the second phase of grace-period
initialization, which results in the leaf rcu_node structure
being initialized to expect no further quiescent states, but
with that structure's parent expecting a quiescent-state report.
The parent will never receive a quiescent state from this leaf
rcu_node structure, so the grace period will hang, resulting in
RCU CPU stall warnings.
It would be good to get rid of the special fail-safe quiescent-state
propagation checks. This commit therefore checks the leaf rcu_node
structure's ->wait_blkd_tasks field during grace-period initialization.
If this flag is set, the rcu_report_qs_rnp() is invoked to immediately
report the possible quiescent state. While in the neighborhood, this
commit also report quiescent states for any CPUs that went offline between
the two phases of grace-period initialization, thus reducing grace-period
delays and hopefully eventually allowing removal of offline-CPU checks
from the force-quiescent-state code path.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-05-08 00:34:17 +08:00
|
|
|
|
*
|
|
|
|
|
* As a special case, if mask is zero, the bit-already-cleared check is
|
|
|
|
|
* disabled. This allows propagating quiescent state due to resumed tasks
|
|
|
|
|
* during grace-period initialization.
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
*/
|
2018-07-04 08:22:34 +08:00
|
|
|
|
static void rcu_report_qs_rnp(unsigned long mask, struct rcu_node *rnp,
|
|
|
|
|
unsigned long gps, unsigned long flags)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
__releases(rnp->lock)
|
|
|
|
|
{
|
rcu: Associate quiescent-state reports with grace period
As noted in earlier commit logs, CPU hotplug operations running
concurrently with grace-period initialization can result in a given
leaf rcu_node structure having all CPUs offline and no blocked readers,
but with this rcu_node structure nevertheless blocking the current
grace period. Therefore, the quiescent-state forcing code now checks
for this situation and repairs it.
Unfortunately, this checking can result in false positives, for example,
when the last task has just removed itself from this leaf rcu_node
structure, but has not yet started clearing the ->qsmask bits further
up the structure. This means that the grace-period kthread (which
forces quiescent states) and some other task might be attempting to
concurrently clear these ->qsmask bits. This is usually not a problem:
One of these tasks will be the first to acquire the upper-level rcu_node
structure's lock and with therefore clear the bit, and the other task,
seeing the bit already cleared, will stop trying to clear bits.
Sadly, this means that the following unusual sequence of events -can-
result in a problem:
1. The grace-period kthread wins, and clears the ->qsmask bits.
2. This is the last thing blocking the current grace period, so
that the grace-period kthread clears ->qsmask bits all the way
to the root and finds that the root ->qsmask field is now zero.
3. Another grace period is required, so that the grace period kthread
initializes it, including setting all the needed qsmask bits.
4. The leaf rcu_node structure (the one that started this whole
mess) is blocking this new grace period, either because it
has at least one online CPU or because there is at least one
task that had blocked within an RCU read-side critical section
while running on one of this leaf rcu_node structure's CPUs.
(And yes, that CPU might well have gone offline before the
grace period in step (3) above started, which can mean that
there is a task on the leaf rcu_node structure's ->blkd_tasks
list, but ->qsmask equal to zero.)
5. The other kthread didn't get around to trying to clear the upper
level ->qsmask bits until all the above had happened. This means
that it now sees bits set in the upper-level ->qsmask field, so it
proceeds to clear them. Too bad that it is doing so on behalf of
a quiescent state that does not apply to the current grace period!
This sequence of events can result in the new grace period being too
short. It can also result in the new grace period ending before the
leaf rcu_node structure's ->qsmask bits have been cleared, which will
result in splats during initialization of the next grace period. In
addition, it can result in tasks blocking the new grace period still
being queued at the start of the next grace period, which will result
in other splats. Sasha's testing turned up another of these splats,
as did rcutorture testing. (And yes, rcutorture is being adjusted to
make these splats show up more quickly. Which probably is having the
undesirable side effect of making other problems show up less quickly.
Can't have everything!)
Reported-by: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: <stable@vger.kernel.org> # 4.0.x
Tested-by: Sasha Levin <sasha.levin@oracle.com>
2015-03-16 00:19:35 +08:00
|
|
|
|
unsigned long oldmask = 0;
|
2009-09-19 00:50:17 +08:00
|
|
|
|
struct rcu_node *rnp_c;
|
|
|
|
|
|
2018-01-17 22:24:30 +08:00
|
|
|
|
raw_lockdep_assert_held_rcu_node(rnp);
|
2017-04-29 03:32:15 +08:00
|
|
|
|
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
/* Walk up the rcu_node hierarchy. */
|
|
|
|
|
for (;;) {
|
rcu: Fix grace-period hangs from mid-init task resume
Without special fail-safe quiescent-state-propagation checks, grace-period
hangs can result from the following scenario:
1. A task running on a given CPU is preempted in its RCU read-side
critical section.
2. That CPU goes offline, and there are now no online CPUs
corresponding to that CPU's leaf rcu_node structure.
3. The rcu_gp_init() function does the first phase of grace-period
initialization, and sets the aforementioned leaf rcu_node
structure's ->qsmaskinit field to all zeroes. Because there
is a blocked task, it does not propagate the zeroing of either
->qsmaskinit or ->qsmaskinitnext up the rcu_node tree.
4. The task resumes on some other CPU and exits its critical section.
There is no grace period in progress, so the resulting quiescent
state is not reported up the tree.
5. The rcu_gp_init() function does the second phase of grace-period
initialization, which results in the leaf rcu_node structure
being initialized to expect no further quiescent states, but
with that structure's parent expecting a quiescent-state report.
The parent will never receive a quiescent state from this leaf
rcu_node structure, so the grace period will hang, resulting in
RCU CPU stall warnings.
It would be good to get rid of the special fail-safe quiescent-state
propagation checks. This commit therefore checks the leaf rcu_node
structure's ->wait_blkd_tasks field during grace-period initialization.
If this flag is set, the rcu_report_qs_rnp() is invoked to immediately
report the possible quiescent state. While in the neighborhood, this
commit also report quiescent states for any CPUs that went offline between
the two phases of grace-period initialization, thus reducing grace-period
delays and hopefully eventually allowing removal of offline-CPU checks
from the force-quiescent-state code path.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-05-08 00:34:17 +08:00
|
|
|
|
if ((!(rnp->qsmask & mask) && mask) || rnp->gp_seq != gps) {
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
rcu: Associate quiescent-state reports with grace period
As noted in earlier commit logs, CPU hotplug operations running
concurrently with grace-period initialization can result in a given
leaf rcu_node structure having all CPUs offline and no blocked readers,
but with this rcu_node structure nevertheless blocking the current
grace period. Therefore, the quiescent-state forcing code now checks
for this situation and repairs it.
Unfortunately, this checking can result in false positives, for example,
when the last task has just removed itself from this leaf rcu_node
structure, but has not yet started clearing the ->qsmask bits further
up the structure. This means that the grace-period kthread (which
forces quiescent states) and some other task might be attempting to
concurrently clear these ->qsmask bits. This is usually not a problem:
One of these tasks will be the first to acquire the upper-level rcu_node
structure's lock and with therefore clear the bit, and the other task,
seeing the bit already cleared, will stop trying to clear bits.
Sadly, this means that the following unusual sequence of events -can-
result in a problem:
1. The grace-period kthread wins, and clears the ->qsmask bits.
2. This is the last thing blocking the current grace period, so
that the grace-period kthread clears ->qsmask bits all the way
to the root and finds that the root ->qsmask field is now zero.
3. Another grace period is required, so that the grace period kthread
initializes it, including setting all the needed qsmask bits.
4. The leaf rcu_node structure (the one that started this whole
mess) is blocking this new grace period, either because it
has at least one online CPU or because there is at least one
task that had blocked within an RCU read-side critical section
while running on one of this leaf rcu_node structure's CPUs.
(And yes, that CPU might well have gone offline before the
grace period in step (3) above started, which can mean that
there is a task on the leaf rcu_node structure's ->blkd_tasks
list, but ->qsmask equal to zero.)
5. The other kthread didn't get around to trying to clear the upper
level ->qsmask bits until all the above had happened. This means
that it now sees bits set in the upper-level ->qsmask field, so it
proceeds to clear them. Too bad that it is doing so on behalf of
a quiescent state that does not apply to the current grace period!
This sequence of events can result in the new grace period being too
short. It can also result in the new grace period ending before the
leaf rcu_node structure's ->qsmask bits have been cleared, which will
result in splats during initialization of the next grace period. In
addition, it can result in tasks blocking the new grace period still
being queued at the start of the next grace period, which will result
in other splats. Sasha's testing turned up another of these splats,
as did rcutorture testing. (And yes, rcutorture is being adjusted to
make these splats show up more quickly. Which probably is having the
undesirable side effect of making other problems show up less quickly.
Can't have everything!)
Reported-by: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: <stable@vger.kernel.org> # 4.0.x
Tested-by: Sasha Levin <sasha.levin@oracle.com>
2015-03-16 00:19:35 +08:00
|
|
|
|
/*
|
|
|
|
|
* Our bit has already been cleared, or the
|
|
|
|
|
* relevant grace period is already over, so done.
|
|
|
|
|
*/
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
return;
|
|
|
|
|
}
|
rcu: Associate quiescent-state reports with grace period
As noted in earlier commit logs, CPU hotplug operations running
concurrently with grace-period initialization can result in a given
leaf rcu_node structure having all CPUs offline and no blocked readers,
but with this rcu_node structure nevertheless blocking the current
grace period. Therefore, the quiescent-state forcing code now checks
for this situation and repairs it.
Unfortunately, this checking can result in false positives, for example,
when the last task has just removed itself from this leaf rcu_node
structure, but has not yet started clearing the ->qsmask bits further
up the structure. This means that the grace-period kthread (which
forces quiescent states) and some other task might be attempting to
concurrently clear these ->qsmask bits. This is usually not a problem:
One of these tasks will be the first to acquire the upper-level rcu_node
structure's lock and with therefore clear the bit, and the other task,
seeing the bit already cleared, will stop trying to clear bits.
Sadly, this means that the following unusual sequence of events -can-
result in a problem:
1. The grace-period kthread wins, and clears the ->qsmask bits.
2. This is the last thing blocking the current grace period, so
that the grace-period kthread clears ->qsmask bits all the way
to the root and finds that the root ->qsmask field is now zero.
3. Another grace period is required, so that the grace period kthread
initializes it, including setting all the needed qsmask bits.
4. The leaf rcu_node structure (the one that started this whole
mess) is blocking this new grace period, either because it
has at least one online CPU or because there is at least one
task that had blocked within an RCU read-side critical section
while running on one of this leaf rcu_node structure's CPUs.
(And yes, that CPU might well have gone offline before the
grace period in step (3) above started, which can mean that
there is a task on the leaf rcu_node structure's ->blkd_tasks
list, but ->qsmask equal to zero.)
5. The other kthread didn't get around to trying to clear the upper
level ->qsmask bits until all the above had happened. This means
that it now sees bits set in the upper-level ->qsmask field, so it
proceeds to clear them. Too bad that it is doing so on behalf of
a quiescent state that does not apply to the current grace period!
This sequence of events can result in the new grace period being too
short. It can also result in the new grace period ending before the
leaf rcu_node structure's ->qsmask bits have been cleared, which will
result in splats during initialization of the next grace period. In
addition, it can result in tasks blocking the new grace period still
being queued at the start of the next grace period, which will result
in other splats. Sasha's testing turned up another of these splats,
as did rcutorture testing. (And yes, rcutorture is being adjusted to
make these splats show up more quickly. Which probably is having the
undesirable side effect of making other problems show up less quickly.
Can't have everything!)
Reported-by: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: <stable@vger.kernel.org> # 4.0.x
Tested-by: Sasha Levin <sasha.levin@oracle.com>
2015-03-16 00:19:35 +08:00
|
|
|
|
WARN_ON_ONCE(oldmask); /* Any child must be all zeroed! */
|
2018-04-14 08:11:44 +08:00
|
|
|
|
WARN_ON_ONCE(!rcu_is_leaf_node(rnp) &&
|
2017-07-12 12:52:31 +08:00
|
|
|
|
rcu_preempt_blocked_readers_cgp(rnp));
|
2020-01-04 03:38:51 +08:00
|
|
|
|
WRITE_ONCE(rnp->qsmask, rnp->qsmask & ~mask);
|
2018-07-06 07:15:38 +08:00
|
|
|
|
trace_rcu_quiescent_state_report(rcu_state.name, rnp->gp_seq,
|
rcu: Add grace-period, quiescent-state, and call_rcu trace events
Add trace events to record grace-period start and end, quiescent states,
CPUs noticing grace-period start and end, grace-period initialization,
call_rcu() invocation, tasks blocking in RCU read-side critical sections,
tasks exiting those same critical sections, force_quiescent_state()
detection of dyntick-idle and offline CPUs, CPUs entering and leaving
dyntick-idle mode (except from NMIs), CPUs coming online and going
offline, and CPUs being kicked for staying in dyntick-idle mode for too
long (as in many weeks, even on 32-bit systems).
Signed-off-by: Paul E. McKenney <paul.mckenney@linaro.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
rcu: Add the rcu flavor to callback trace events
The earlier trace events for registering RCU callbacks and for invoking
them did not include the RCU flavor (rcu_bh, rcu_preempt, or rcu_sched).
This commit adds the RCU flavor to those trace events.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2011-06-25 21:36:56 +08:00
|
|
|
|
mask, rnp->qsmask, rnp->level,
|
|
|
|
|
rnp->grplo, rnp->grphi,
|
|
|
|
|
!!rnp->gp_tasks);
|
2011-02-08 04:47:15 +08:00
|
|
|
|
if (rnp->qsmask != 0 || rcu_preempt_blocked_readers_cgp(rnp)) {
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
|
|
|
|
/* Other bits still set at this level, so done. */
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
return;
|
|
|
|
|
}
|
2018-04-29 09:50:06 +08:00
|
|
|
|
rnp->completedqs = rnp->gp_seq;
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
mask = rnp->grpmask;
|
|
|
|
|
if (rnp->parent == NULL) {
|
|
|
|
|
|
|
|
|
|
/* No more levels. Exit loop holding root lock. */
|
|
|
|
|
|
|
|
|
|
break;
|
|
|
|
|
}
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
|
2009-09-19 00:50:17 +08:00
|
|
|
|
rnp_c = rnp;
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
rnp = rnp->parent;
|
2015-10-08 18:24:23 +08:00
|
|
|
|
raw_spin_lock_irqsave_rcu_node(rnp, flags);
|
2020-01-04 06:53:31 +08:00
|
|
|
|
oldmask = READ_ONCE(rnp_c->qsmask);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Get here if we are the last CPU to pass through a quiescent
|
2009-12-03 04:10:13 +08:00
|
|
|
|
* state for this grace period. Invoke rcu_report_qs_rsp()
|
rcu: Merge preemptable-RCU functionality into hierarchical RCU
Create a kernel/rcutree_plugin.h file that contains definitions
for preemptable RCU (or, under the #else branch of the #ifdef,
empty definitions for the classic non-preemptable semantics).
These definitions fit into plugins defined in kernel/rcutree.c
for this purpose.
This variant of preemptable RCU uses a new algorithm whose
read-side expense is roughly that of classic hierarchical RCU
under CONFIG_PREEMPT. This new algorithm's update-side expense
is similar to that of classic hierarchical RCU, and, in absence
of read-side preemption or blocking, is exactly that of classic
hierarchical RCU. Perhaps more important, this new algorithm
has a much simpler implementation, saving well over 1,000 lines
of code compared to mainline's implementation of preemptable
RCU, which will hopefully be retired in favor of this new
algorithm.
The simplifications are obtained by maintaining per-task
nesting state for running tasks, and using a simple
lock-protected algorithm to handle accounting when tasks block
within RCU read-side critical sections, making use of lessons
learned while creating numerous user-level RCU implementations
over the past 18 months.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: laijs@cn.fujitsu.com
Cc: dipankar@in.ibm.com
Cc: akpm@linux-foundation.org
Cc: mathieu.desnoyers@polymtl.ca
Cc: josht@linux.vnet.ibm.com
Cc: dvhltc@us.ibm.com
Cc: niv@us.ibm.com
Cc: peterz@infradead.org
Cc: rostedt@goodmis.org
LKML-Reference: <12509746134003-git-send-email->
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-08-23 04:56:52 +08:00
|
|
|
|
* to clean up and start the next grace period if one is needed.
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
*/
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_report_qs_rsp(flags); /* releases rnp->lock. */
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
|
|
|
|
|
2015-02-24 00:59:29 +08:00
|
|
|
|
/*
|
|
|
|
|
* Record a quiescent state for all tasks that were previously queued
|
|
|
|
|
* on the specified rcu_node structure and that were blocking the current
|
2018-07-08 09:12:26 +08:00
|
|
|
|
* RCU grace period. The caller must hold the corresponding rnp->lock with
|
2015-02-24 00:59:29 +08:00
|
|
|
|
* irqs disabled, and this lock is released upon return, but irqs remain
|
|
|
|
|
* disabled.
|
|
|
|
|
*/
|
2018-05-04 05:30:02 +08:00
|
|
|
|
static void __maybe_unused
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_report_unblock_qs_rnp(struct rcu_node *rnp, unsigned long flags)
|
2015-02-24 00:59:29 +08:00
|
|
|
|
__releases(rnp->lock)
|
|
|
|
|
{
|
rcu: Associate quiescent-state reports with grace period
As noted in earlier commit logs, CPU hotplug operations running
concurrently with grace-period initialization can result in a given
leaf rcu_node structure having all CPUs offline and no blocked readers,
but with this rcu_node structure nevertheless blocking the current
grace period. Therefore, the quiescent-state forcing code now checks
for this situation and repairs it.
Unfortunately, this checking can result in false positives, for example,
when the last task has just removed itself from this leaf rcu_node
structure, but has not yet started clearing the ->qsmask bits further
up the structure. This means that the grace-period kthread (which
forces quiescent states) and some other task might be attempting to
concurrently clear these ->qsmask bits. This is usually not a problem:
One of these tasks will be the first to acquire the upper-level rcu_node
structure's lock and with therefore clear the bit, and the other task,
seeing the bit already cleared, will stop trying to clear bits.
Sadly, this means that the following unusual sequence of events -can-
result in a problem:
1. The grace-period kthread wins, and clears the ->qsmask bits.
2. This is the last thing blocking the current grace period, so
that the grace-period kthread clears ->qsmask bits all the way
to the root and finds that the root ->qsmask field is now zero.
3. Another grace period is required, so that the grace period kthread
initializes it, including setting all the needed qsmask bits.
4. The leaf rcu_node structure (the one that started this whole
mess) is blocking this new grace period, either because it
has at least one online CPU or because there is at least one
task that had blocked within an RCU read-side critical section
while running on one of this leaf rcu_node structure's CPUs.
(And yes, that CPU might well have gone offline before the
grace period in step (3) above started, which can mean that
there is a task on the leaf rcu_node structure's ->blkd_tasks
list, but ->qsmask equal to zero.)
5. The other kthread didn't get around to trying to clear the upper
level ->qsmask bits until all the above had happened. This means
that it now sees bits set in the upper-level ->qsmask field, so it
proceeds to clear them. Too bad that it is doing so on behalf of
a quiescent state that does not apply to the current grace period!
This sequence of events can result in the new grace period being too
short. It can also result in the new grace period ending before the
leaf rcu_node structure's ->qsmask bits have been cleared, which will
result in splats during initialization of the next grace period. In
addition, it can result in tasks blocking the new grace period still
being queued at the start of the next grace period, which will result
in other splats. Sasha's testing turned up another of these splats,
as did rcutorture testing. (And yes, rcutorture is being adjusted to
make these splats show up more quickly. Which probably is having the
undesirable side effect of making other problems show up less quickly.
Can't have everything!)
Reported-by: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: <stable@vger.kernel.org> # 4.0.x
Tested-by: Sasha Levin <sasha.levin@oracle.com>
2015-03-16 00:19:35 +08:00
|
|
|
|
unsigned long gps;
|
2015-02-24 00:59:29 +08:00
|
|
|
|
unsigned long mask;
|
|
|
|
|
struct rcu_node *rnp_p;
|
|
|
|
|
|
2018-01-17 22:24:30 +08:00
|
|
|
|
raw_lockdep_assert_held_rcu_node(rnp);
|
2019-10-15 18:28:48 +08:00
|
|
|
|
if (WARN_ON_ONCE(!IS_ENABLED(CONFIG_PREEMPT_RCU)) ||
|
2018-04-28 05:05:27 +08:00
|
|
|
|
WARN_ON_ONCE(rcu_preempt_blocked_readers_cgp(rnp)) ||
|
|
|
|
|
rnp->qsmask != 0) {
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
|
2015-02-24 00:59:29 +08:00
|
|
|
|
return; /* Still need more quiescent states! */
|
|
|
|
|
}
|
|
|
|
|
|
2018-05-02 06:00:10 +08:00
|
|
|
|
rnp->completedqs = rnp->gp_seq;
|
2015-02-24 00:59:29 +08:00
|
|
|
|
rnp_p = rnp->parent;
|
|
|
|
|
if (rnp_p == NULL) {
|
|
|
|
|
/*
|
2015-03-09 05:52:27 +08:00
|
|
|
|
* Only one rcu_node structure in the tree, so don't
|
|
|
|
|
* try to report up to its nonexistent parent!
|
2015-02-24 00:59:29 +08:00
|
|
|
|
*/
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_report_qs_rsp(flags);
|
2015-02-24 00:59:29 +08:00
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
2018-04-28 05:54:46 +08:00
|
|
|
|
/* Report up the rest of the hierarchy, tracking current ->gp_seq. */
|
|
|
|
|
gps = rnp->gp_seq;
|
2015-02-24 00:59:29 +08:00
|
|
|
|
mask = rnp->grpmask;
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_unlock_rcu_node(rnp); /* irqs remain disabled. */
|
2015-10-08 18:24:23 +08:00
|
|
|
|
raw_spin_lock_rcu_node(rnp_p); /* irqs already disabled. */
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_report_qs_rnp(mask, rnp_p, gps, flags);
|
2015-02-24 00:59:29 +08:00
|
|
|
|
}
|
|
|
|
|
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
/*
|
2009-12-03 04:10:13 +08:00
|
|
|
|
* Record a quiescent state for the specified CPU to that CPU's rcu_data
|
2016-01-28 14:44:45 +08:00
|
|
|
|
* structure. This must be called from the specified CPU.
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
*/
|
|
|
|
|
static void
|
2020-08-21 02:26:14 +08:00
|
|
|
|
rcu_report_qs_rdp(struct rcu_data *rdp)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
{
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
unsigned long mask;
|
2021-10-19 08:08:10 +08:00
|
|
|
|
bool needacc = false;
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
struct rcu_node *rnp;
|
|
|
|
|
|
2020-08-21 02:26:14 +08:00
|
|
|
|
WARN_ON_ONCE(rdp->cpu != smp_processor_id());
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
rnp = rdp->mynode;
|
2015-10-08 18:24:23 +08:00
|
|
|
|
raw_spin_lock_irqsave_rcu_node(rnp, flags);
|
2018-04-28 05:54:46 +08:00
|
|
|
|
if (rdp->cpu_no_qs.b.norm || rdp->gp_seq != rnp->gp_seq ||
|
|
|
|
|
rdp->gpwrap) {
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
|
|
|
|
/*
|
rcu: Simplify quiescent-state accounting
There is often a delay between the time that a CPU passes through a
quiescent state and the time that this quiescent state is reported to the
RCU core. It is quite possible that the grace period ended before the
quiescent state could be reported, for example, some other CPU might have
deduced that this CPU passed through dyntick-idle mode. It is critically
important that quiescent state be counted only against the grace period
that was in effect at the time that the quiescent state was detected.
Previously, this was handled by recording the number of the last grace
period to complete when passing through a quiescent state. The RCU
core then checks this number against the current value, and rejects
the quiescent state if there is a mismatch. However, one additional
possibility must be accounted for, namely that the quiescent state was
recorded after the prior grace period completed but before the current
grace period started. In this case, the RCU core must reject the
quiescent state, but the recorded number will match. This is handled
when the CPU becomes aware of a new grace period -- at that point,
it invalidates any prior quiescent state.
This works, but is a bit indirect. The new approach records the current
grace period, and the RCU core checks to see (1) that this is still the
current grace period and (2) that this grace period has not yet ended.
This approach simplifies reasoning about correctness, and this commit
changes over to this new approach.
Signed-off-by: Paul E. McKenney <paul.mckenney@linaro.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2011-06-27 15:17:43 +08:00
|
|
|
|
* The grace period in which this quiescent state was
|
|
|
|
|
* recorded has ended, so don't report it upwards.
|
|
|
|
|
* We will instead need a new quiescent state that lies
|
|
|
|
|
* within the current grace period.
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
*/
|
2015-08-07 06:16:57 +08:00
|
|
|
|
rdp->cpu_no_qs.b.norm = true; /* need qs for new gp. */
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
mask = rdp->grpmask;
|
2020-08-21 02:26:14 +08:00
|
|
|
|
rdp->core_needs_qs = false;
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
if ((rnp->qsmask & mask) == 0) {
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
} else {
|
|
|
|
|
/*
|
|
|
|
|
* This GP can't end until cpu checks in, so all of our
|
|
|
|
|
* callbacks can be processed during the next GP.
|
2021-10-19 08:08:09 +08:00
|
|
|
|
*
|
2021-10-19 08:08:10 +08:00
|
|
|
|
* NOCB kthreads have their own way to deal with that...
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
*/
|
2021-10-19 08:08:10 +08:00
|
|
|
|
if (!rcu_rdp_is_offloaded(rdp)) {
|
2023-01-18 15:30:14 +08:00
|
|
|
|
/*
|
|
|
|
|
* The current GP has not yet ended, so it
|
|
|
|
|
* should not be possible for rcu_accelerate_cbs()
|
|
|
|
|
* to return true. So complain, but don't awaken.
|
|
|
|
|
*/
|
|
|
|
|
WARN_ON_ONCE(rcu_accelerate_cbs(rnp, rdp));
|
2021-10-19 08:08:10 +08:00
|
|
|
|
} else if (!rcu_segcblist_completely_offloaded(&rdp->cblist)) {
|
|
|
|
|
/*
|
|
|
|
|
* ...but NOCB kthreads may miss or delay callbacks acceleration
|
|
|
|
|
* if in the middle of a (de-)offloading process.
|
|
|
|
|
*/
|
|
|
|
|
needacc = true;
|
|
|
|
|
}
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
2019-09-06 01:26:41 +08:00
|
|
|
|
rcu_disable_urgency_upon_qs(rdp);
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_report_qs_rnp(mask, rnp, rnp->gp_seq, flags);
|
rcu: Associate quiescent-state reports with grace period
As noted in earlier commit logs, CPU hotplug operations running
concurrently with grace-period initialization can result in a given
leaf rcu_node structure having all CPUs offline and no blocked readers,
but with this rcu_node structure nevertheless blocking the current
grace period. Therefore, the quiescent-state forcing code now checks
for this situation and repairs it.
Unfortunately, this checking can result in false positives, for example,
when the last task has just removed itself from this leaf rcu_node
structure, but has not yet started clearing the ->qsmask bits further
up the structure. This means that the grace-period kthread (which
forces quiescent states) and some other task might be attempting to
concurrently clear these ->qsmask bits. This is usually not a problem:
One of these tasks will be the first to acquire the upper-level rcu_node
structure's lock and with therefore clear the bit, and the other task,
seeing the bit already cleared, will stop trying to clear bits.
Sadly, this means that the following unusual sequence of events -can-
result in a problem:
1. The grace-period kthread wins, and clears the ->qsmask bits.
2. This is the last thing blocking the current grace period, so
that the grace-period kthread clears ->qsmask bits all the way
to the root and finds that the root ->qsmask field is now zero.
3. Another grace period is required, so that the grace period kthread
initializes it, including setting all the needed qsmask bits.
4. The leaf rcu_node structure (the one that started this whole
mess) is blocking this new grace period, either because it
has at least one online CPU or because there is at least one
task that had blocked within an RCU read-side critical section
while running on one of this leaf rcu_node structure's CPUs.
(And yes, that CPU might well have gone offline before the
grace period in step (3) above started, which can mean that
there is a task on the leaf rcu_node structure's ->blkd_tasks
list, but ->qsmask equal to zero.)
5. The other kthread didn't get around to trying to clear the upper
level ->qsmask bits until all the above had happened. This means
that it now sees bits set in the upper-level ->qsmask field, so it
proceeds to clear them. Too bad that it is doing so on behalf of
a quiescent state that does not apply to the current grace period!
This sequence of events can result in the new grace period being too
short. It can also result in the new grace period ending before the
leaf rcu_node structure's ->qsmask bits have been cleared, which will
result in splats during initialization of the next grace period. In
addition, it can result in tasks blocking the new grace period still
being queued at the start of the next grace period, which will result
in other splats. Sasha's testing turned up another of these splats,
as did rcutorture testing. (And yes, rcutorture is being adjusted to
make these splats show up more quickly. Which probably is having the
undesirable side effect of making other problems show up less quickly.
Can't have everything!)
Reported-by: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: <stable@vger.kernel.org> # 4.0.x
Tested-by: Sasha Levin <sasha.levin@oracle.com>
2015-03-16 00:19:35 +08:00
|
|
|
|
/* ^^^ Released rnp->lock */
|
2021-10-19 08:08:10 +08:00
|
|
|
|
|
|
|
|
|
if (needacc) {
|
|
|
|
|
rcu_nocb_lock_irqsave(rdp, flags);
|
|
|
|
|
rcu_accelerate_cbs_unlocked(rnp, rdp);
|
|
|
|
|
rcu_nocb_unlock_irqrestore(rdp, flags);
|
|
|
|
|
}
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Check to see if there is a new grace period of which this CPU
|
|
|
|
|
* is not yet aware, and if so, set up local rcu_data state for it.
|
|
|
|
|
* Otherwise, see if this CPU has just passed through its first
|
|
|
|
|
* quiescent state for this grace period, and record that fact if so.
|
|
|
|
|
*/
|
|
|
|
|
static void
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_check_quiescent_state(struct rcu_data *rdp)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
{
|
2013-03-20 03:38:24 +08:00
|
|
|
|
/* Check for grace-period ends and beginnings. */
|
2018-07-04 08:22:34 +08:00
|
|
|
|
note_gp_changes(rdp);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Does this CPU still need to do its part for current grace period?
|
|
|
|
|
* If no, return and let the other CPUs do their part as well.
|
|
|
|
|
*/
|
2015-08-07 02:31:51 +08:00
|
|
|
|
if (!rdp->core_needs_qs)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Was there a quiescent state since the beginning of the grace
|
|
|
|
|
* period? If no, then exit and wait for the next call.
|
|
|
|
|
*/
|
2016-12-01 03:21:21 +08:00
|
|
|
|
if (rdp->cpu_no_qs.b.norm)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
return;
|
|
|
|
|
|
2009-12-03 04:10:13 +08:00
|
|
|
|
/*
|
|
|
|
|
* Tell RCU we are done (but rcu_report_qs_rdp() will be the
|
|
|
|
|
* judge of that).
|
|
|
|
|
*/
|
2020-08-21 02:26:14 +08:00
|
|
|
|
rcu_report_qs_rdp(rdp);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
|
|
|
|
|
2023-03-22 07:43:54 +08:00
|
|
|
|
/* Return true if callback-invocation time limit exceeded. */
|
2023-04-01 00:05:56 +08:00
|
|
|
|
static bool rcu_do_batch_check_time(long count, long tlimit,
|
|
|
|
|
bool jlimit_check, unsigned long jlimit)
|
2023-03-22 07:43:54 +08:00
|
|
|
|
{
|
|
|
|
|
// Invoke local_clock() only once per 32 consecutive callbacks.
|
2023-04-01 00:05:56 +08:00
|
|
|
|
return unlikely(tlimit) &&
|
|
|
|
|
(!likely(count & 31) ||
|
|
|
|
|
(IS_ENABLED(CONFIG_RCU_DOUBLE_CHECK_CB_TIME) &&
|
|
|
|
|
jlimit_check && time_after(jiffies, jlimit))) &&
|
|
|
|
|
local_clock() >= tlimit;
|
2023-03-22 07:43:54 +08:00
|
|
|
|
}
|
|
|
|
|
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
/*
|
|
|
|
|
* Invoke any RCU callbacks that have made it to the end of their grace
|
2021-03-23 13:29:10 +08:00
|
|
|
|
* period. Throttle as specified by rdp->blimit.
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
*/
|
2018-07-04 08:22:34 +08:00
|
|
|
|
static void rcu_do_batch(struct rcu_data *rdp)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
{
|
2023-04-01 00:05:56 +08:00
|
|
|
|
long bl;
|
|
|
|
|
long count = 0;
|
2020-06-24 08:09:27 +08:00
|
|
|
|
int div;
|
2020-11-19 00:15:41 +08:00
|
|
|
|
bool __maybe_unused empty;
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
unsigned long flags;
|
2023-04-01 00:05:56 +08:00
|
|
|
|
unsigned long jlimit;
|
|
|
|
|
bool jlimit_check = false;
|
|
|
|
|
long pending;
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
struct rcu_cblist rcl = RCU_CBLIST_INITIALIZER(rcl);
|
2023-04-01 00:05:56 +08:00
|
|
|
|
struct rcu_head *rhp;
|
|
|
|
|
long tlimit = 0;
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
2012-12-04 05:52:00 +08:00
|
|
|
|
/* If no callbacks are ready, just return. */
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
if (!rcu_segcblist_ready_cbs(&rdp->cblist)) {
|
2018-07-06 06:54:02 +08:00
|
|
|
|
trace_rcu_batch_start(rcu_state.name,
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
rcu_segcblist_n_cbs(&rdp->cblist), 0);
|
2018-07-06 06:54:02 +08:00
|
|
|
|
trace_rcu_batch_end(rcu_state.name, 0,
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
!rcu_segcblist_empty(&rdp->cblist),
|
2011-12-08 08:32:40 +08:00
|
|
|
|
need_resched(), is_idle_task(current),
|
2022-04-29 20:42:22 +08:00
|
|
|
|
rcu_is_callbacks_kthread(rdp));
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
return;
|
2011-06-18 06:53:19 +08:00
|
|
|
|
}
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
|
|
|
|
/*
|
2021-10-19 08:08:12 +08:00
|
|
|
|
* Extract the list of ready callbacks, disabling IRQs to prevent
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
* races with call_rcu() from interrupt handlers. Leave the
|
|
|
|
|
* callback counts, as rcu_barrier() needs to be conservative.
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
*/
|
2021-10-19 08:08:12 +08:00
|
|
|
|
rcu_nocb_lock_irqsave(rdp, flags);
|
2012-01-11 06:23:29 +08:00
|
|
|
|
WARN_ON_ONCE(cpu_is_offline(smp_processor_id()));
|
2022-11-15 01:40:19 +08:00
|
|
|
|
pending = rcu_segcblist_get_seglen(&rdp->cblist, RCU_DONE_TAIL);
|
2020-06-24 08:09:27 +08:00
|
|
|
|
div = READ_ONCE(rcu_divisor);
|
|
|
|
|
div = div < 0 ? 7 : div > sizeof(long) * 8 - 2 ? sizeof(long) * 8 - 2 : div;
|
|
|
|
|
bl = max(rdp->blimit, pending >> div);
|
2023-03-22 07:43:54 +08:00
|
|
|
|
if ((in_serving_softirq() || rdp->rcu_cpu_kthread_status == RCU_KTHREAD_RUNNING) &&
|
2023-04-01 00:05:56 +08:00
|
|
|
|
(IS_ENABLED(CONFIG_RCU_DOUBLE_CHECK_CB_TIME) || unlikely(bl > 100))) {
|
|
|
|
|
const long npj = NSEC_PER_SEC / HZ;
|
2020-06-24 08:49:40 +08:00
|
|
|
|
long rrn = READ_ONCE(rcu_resched_ns);
|
|
|
|
|
|
|
|
|
|
rrn = rrn < NSEC_PER_MSEC ? NSEC_PER_MSEC : rrn > NSEC_PER_SEC ? NSEC_PER_SEC : rrn;
|
|
|
|
|
tlimit = local_clock() + rrn;
|
2023-04-01 00:05:56 +08:00
|
|
|
|
jlimit = jiffies + (rrn + npj + 1) / npj;
|
|
|
|
|
jlimit_check = true;
|
2020-06-24 08:49:40 +08:00
|
|
|
|
}
|
2018-07-06 06:54:02 +08:00
|
|
|
|
trace_rcu_batch_start(rcu_state.name,
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
rcu_segcblist_n_cbs(&rdp->cblist), bl);
|
|
|
|
|
rcu_segcblist_extract_done_cbs(&rdp->cblist, &rcl);
|
2021-10-19 08:08:11 +08:00
|
|
|
|
if (rcu_rdp_is_offloaded(rdp))
|
2019-05-28 20:54:26 +08:00
|
|
|
|
rdp->qlen_last_fqs_check = rcu_segcblist_n_cbs(&rdp->cblist);
|
2020-11-15 03:31:32 +08:00
|
|
|
|
|
|
|
|
|
trace_rcu_segcb_stats(&rdp->cblist, TPS("SegCbDequeued"));
|
2019-05-16 00:56:40 +08:00
|
|
|
|
rcu_nocb_unlock_irqrestore(rdp, flags);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
|
|
|
|
/* Invoke callbacks. */
|
2019-07-29 02:50:56 +08:00
|
|
|
|
tick_dep_set_task(current, TICK_DEP_BIT_RCU);
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
rhp = rcu_cblist_dequeue(&rcl);
|
2020-11-15 03:31:32 +08:00
|
|
|
|
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
for (; rhp; rhp = rcu_cblist_dequeue(&rcl)) {
|
2019-08-31 00:36:32 +08:00
|
|
|
|
rcu_callback_t f;
|
|
|
|
|
|
rcu/tree: Make rcu_do_batch count how many callbacks were executed
The rcu_do_batch() function extracts the ready-to-invoke callbacks
from the rcu_segcblist located in the ->cblist field of the current
CPU's rcu_data structure. These callbacks are first moved to a local
(unsegmented) rcu_cblist. The rcu_do_batch() function then uses this
rcu_cblist's ->len field to count how many CBs it has invoked, but it
does so by counting that field down from zero. Finally, this function
negates the value in this ->len field (resulting in a positive number)
and subtracts the result from the ->len field of the current CPU's
->cblist field.
Except that it is sometimes necessary for rcu_do_batch() to stop invoking
callbacks mid-stream, despite there being more ready to invoke, for
example, if a high-priority task wakes up. In this case the remaining
not-yet-invoked callbacks are requeued back onto the CPU's ->cblist,
but remain in the ready-to-invoke segment of that list. As above, the
negative of the local rcu_cblist's ->len field is still subtracted from
the ->len field of the current CPU's ->cblist field.
The design of counting down from 0 is confusing and error-prone, plus
use of a positive count will make it easier to provide a uniform and
consistent API to deal with the per-segment counts that are added
later in this series. For example, rcu_segcblist_extract_done_cbs()
can unconditionally populate the resulting unsegmented list's ->len
field during extraction.
This commit therefore explicitly counts how many callbacks were executed
in rcu_do_batch() itself, counting up from zero, and then uses that
to update the per-CPU segcb list's ->len field, without relying on the
downcounting of rcl->len from zero.
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Reviewed-by: Neeraj Upadhyay <neeraju@codeaurora.org>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-11-03 22:25:57 +08:00
|
|
|
|
count++;
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
debug_rcu_head_unqueue(rhp);
|
2019-08-31 00:36:32 +08:00
|
|
|
|
|
|
|
|
|
rcu_lock_acquire(&rcu_callback_map);
|
|
|
|
|
trace_rcu_invoke_callback(rcu_state.name, rhp);
|
|
|
|
|
|
|
|
|
|
f = rhp->func;
|
|
|
|
|
WRITE_ONCE(rhp->func, (rcu_callback_t)0L);
|
|
|
|
|
f(rhp);
|
|
|
|
|
|
|
|
|
|
rcu_lock_release(&rcu_callback_map);
|
|
|
|
|
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
/*
|
|
|
|
|
* Stop only if limit reached and CPU has something to do.
|
|
|
|
|
*/
|
2021-10-19 08:08:14 +08:00
|
|
|
|
if (in_serving_softirq()) {
|
|
|
|
|
if (count >= bl && (need_resched() || !is_idle_task(current)))
|
|
|
|
|
break;
|
2021-10-19 08:08:15 +08:00
|
|
|
|
/*
|
|
|
|
|
* Make sure we don't spend too much time here and deprive other
|
|
|
|
|
* softirq vectors of CPU cycles.
|
|
|
|
|
*/
|
2023-04-01 00:05:56 +08:00
|
|
|
|
if (rcu_do_batch_check_time(count, tlimit, jlimit_check, jlimit))
|
2021-10-19 08:08:15 +08:00
|
|
|
|
break;
|
2021-10-19 08:08:14 +08:00
|
|
|
|
} else {
|
2023-03-22 07:43:54 +08:00
|
|
|
|
// In rcuc/rcuoc context, so no worries about
|
|
|
|
|
// depriving other softirq vectors of CPU cycles.
|
2019-05-16 00:56:40 +08:00
|
|
|
|
local_bh_enable();
|
|
|
|
|
lockdep_assert_irqs_enabled();
|
|
|
|
|
cond_resched_tasks_rcu_qs();
|
|
|
|
|
lockdep_assert_irqs_enabled();
|
|
|
|
|
local_bh_disable();
|
2023-03-22 07:43:54 +08:00
|
|
|
|
// But rcuc kthreads can delay quiescent-state
|
|
|
|
|
// reporting, so check time limits for them.
|
|
|
|
|
if (rdp->rcu_cpu_kthread_status == RCU_KTHREAD_RUNNING &&
|
2023-04-01 00:05:56 +08:00
|
|
|
|
rcu_do_batch_check_time(count, tlimit, jlimit_check, jlimit)) {
|
2023-03-22 07:43:54 +08:00
|
|
|
|
rdp->rcu_cpu_has_work = 1;
|
|
|
|
|
break;
|
|
|
|
|
}
|
2019-05-16 00:56:40 +08:00
|
|
|
|
}
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
|
|
|
|
|
2021-10-19 08:08:12 +08:00
|
|
|
|
rcu_nocb_lock_irqsave(rdp, flags);
|
2020-05-02 07:49:48 +08:00
|
|
|
|
rdp->n_cbs_invoked += count;
|
2018-07-06 06:54:02 +08:00
|
|
|
|
trace_rcu_batch_end(rcu_state.name, count, !!rcl.head, need_resched(),
|
2022-04-29 20:42:22 +08:00
|
|
|
|
is_idle_task(current), rcu_is_callbacks_kthread(rdp));
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
/* Update counts and requeue any remaining callbacks. */
|
|
|
|
|
rcu_segcblist_insert_done_cbs(&rdp->cblist, &rcl);
|
rcu/tree: Make rcu_do_batch count how many callbacks were executed
The rcu_do_batch() function extracts the ready-to-invoke callbacks
from the rcu_segcblist located in the ->cblist field of the current
CPU's rcu_data structure. These callbacks are first moved to a local
(unsegmented) rcu_cblist. The rcu_do_batch() function then uses this
rcu_cblist's ->len field to count how many CBs it has invoked, but it
does so by counting that field down from zero. Finally, this function
negates the value in this ->len field (resulting in a positive number)
and subtracts the result from the ->len field of the current CPU's
->cblist field.
Except that it is sometimes necessary for rcu_do_batch() to stop invoking
callbacks mid-stream, despite there being more ready to invoke, for
example, if a high-priority task wakes up. In this case the remaining
not-yet-invoked callbacks are requeued back onto the CPU's ->cblist,
but remain in the ready-to-invoke segment of that list. As above, the
negative of the local rcu_cblist's ->len field is still subtracted from
the ->len field of the current CPU's ->cblist field.
The design of counting down from 0 is confusing and error-prone, plus
use of a positive count will make it easier to provide a uniform and
consistent API to deal with the per-segment counts that are added
later in this series. For example, rcu_segcblist_extract_done_cbs()
can unconditionally populate the resulting unsegmented list's ->len
field during extraction.
This commit therefore explicitly counts how many callbacks were executed
in rcu_do_batch() itself, counting up from zero, and then uses that
to update the per-CPU segcb list's ->len field, without relying on the
downcounting of rcl->len from zero.
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Reviewed-by: Neeraj Upadhyay <neeraju@codeaurora.org>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-11-03 22:25:57 +08:00
|
|
|
|
rcu_segcblist_add_len(&rdp->cblist, -count);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
|
|
|
|
/* Reinstate batch limit if we have worked down the excess. */
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
count = rcu_segcblist_n_cbs(&rdp->cblist);
|
2019-04-11 08:01:39 +08:00
|
|
|
|
if (rdp->blimit >= DEFAULT_MAX_RCU_BLIMIT && count <= qlowmark)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
rdp->blimit = blimit;
|
|
|
|
|
|
2009-10-15 01:15:55 +08:00
|
|
|
|
/* Reset ->qlen_last_fqs_check trigger if enough CBs have drained. */
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
if (count == 0 && rdp->qlen_last_fqs_check != 0) {
|
2009-10-15 01:15:55 +08:00
|
|
|
|
rdp->qlen_last_fqs_check = 0;
|
2021-07-20 21:16:27 +08:00
|
|
|
|
rdp->n_force_qs_snap = READ_ONCE(rcu_state.n_force_qs);
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
} else if (count < rdp->qlen_last_fqs_check - qhimark)
|
|
|
|
|
rdp->qlen_last_fqs_check = count;
|
2017-10-20 05:52:41 +08:00
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* The following usually indicates a double call_rcu(). To track
|
|
|
|
|
* this down, try building with CONFIG_DEBUG_OBJECTS_RCU_HEAD=y.
|
|
|
|
|
*/
|
2020-11-19 00:15:41 +08:00
|
|
|
|
empty = rcu_segcblist_empty(&rdp->cblist);
|
|
|
|
|
WARN_ON_ONCE(count == 0 && !empty);
|
rcu/nocb: Add bypass callback queueing
Use of the rcu_data structure's segmented ->cblist for no-CBs CPUs
takes advantage of unrelated grace periods, thus reducing the memory
footprint in the face of floods of call_rcu() invocations. However,
the ->cblist field is a more-complex rcu_segcblist structure which must
be protected via locking. Even though there are only three entities
which can acquire this lock (the CPU invoking call_rcu(), the no-CBs
grace-period kthread, and the no-CBs callbacks kthread), the contention
on this lock is excessive under heavy stress.
This commit therefore greatly reduces contention by provisioning
an rcu_cblist structure field named ->nocb_bypass within the
rcu_data structure. Each no-CBs CPU is permitted only a limited
number of enqueues onto the ->cblist per jiffy, controlled by a new
nocb_nobypass_lim_per_jiffy kernel boot parameter that defaults to
about 16 enqueues per millisecond (16 * 1000 / HZ). When that limit is
exceeded, the CPU instead enqueues onto the new ->nocb_bypass.
The ->nocb_bypass is flushed into the ->cblist every jiffy or when
the number of callbacks on ->nocb_bypass exceeds qhimark, whichever
happens first. During call_rcu() floods, this flushing is carried out
by the CPU during the course of its call_rcu() invocations. However,
a CPU could simply stop invoking call_rcu() at any time. The no-CBs
grace-period kthread therefore carries out less-aggressive flushing
(every few jiffies or when the number of callbacks on ->nocb_bypass
exceeds (2 * qhimark), whichever comes first). This means that the
no-CBs grace-period kthread cannot be permitted to do unbounded waits
while there are callbacks on ->nocb_bypass. A ->nocb_bypass_timer is
used to provide the needed wakeups.
[ paulmck: Apply Coverity feedback reported by Colin Ian King. ]
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
2019-07-03 07:03:33 +08:00
|
|
|
|
WARN_ON_ONCE(!IS_ENABLED(CONFIG_RCU_NOCB_CPU) &&
|
2020-11-19 00:15:41 +08:00
|
|
|
|
count != 0 && empty);
|
|
|
|
|
WARN_ON_ONCE(count == 0 && rcu_segcblist_n_segment_cbs(&rdp->cblist) != 0);
|
|
|
|
|
WARN_ON_ONCE(!empty && rcu_segcblist_n_segment_cbs(&rdp->cblist) == 0);
|
2009-10-15 01:15:55 +08:00
|
|
|
|
|
2019-05-16 00:56:40 +08:00
|
|
|
|
rcu_nocb_unlock_irqrestore(rdp, flags);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
2019-07-29 02:50:56 +08:00
|
|
|
|
tick_dep_clear_task(current, TICK_DEP_BIT_RCU);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
2018-11-22 03:35:03 +08:00
|
|
|
|
* This function is invoked from each scheduling-clock interrupt,
|
|
|
|
|
* and checks to see if this CPU is in a non-context-switch quiescent
|
|
|
|
|
* state, for example, user mode or idle loop. It also schedules RCU
|
|
|
|
|
* core processing. If the current grace period has gone on too long,
|
|
|
|
|
* it will ask the scheduler to manufacture a context switch for the sole
|
2021-03-31 04:47:42 +08:00
|
|
|
|
* purpose of providing the needed quiescent state.
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
*/
|
2018-11-22 03:35:03 +08:00
|
|
|
|
void rcu_sched_clock_irq(int user)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
{
|
2022-02-24 09:29:37 +08:00
|
|
|
|
unsigned long j;
|
|
|
|
|
|
|
|
|
|
if (IS_ENABLED(CONFIG_PROVE_RCU)) {
|
|
|
|
|
j = jiffies;
|
|
|
|
|
WARN_ON_ONCE(time_before(j, __this_cpu_read(rcu_data.last_sched_clock)));
|
|
|
|
|
__this_cpu_write(rcu_data.last_sched_clock, j);
|
|
|
|
|
}
|
2013-07-13 05:18:47 +08:00
|
|
|
|
trace_rcu_utilization(TPS("Start scheduler-tick"));
|
2020-11-20 02:13:06 +08:00
|
|
|
|
lockdep_assert_irqs_disabled();
|
2018-07-06 08:59:36 +08:00
|
|
|
|
raw_cpu_inc(rcu_data.ticks_this_gp);
|
rcu: Make need_resched() respond to urgent RCU-QS needs
The per-CPU rcu_dynticks.rcu_urgent_qs variable communicates an urgent
need for an RCU quiescent state from the force-quiescent-state processing
within the grace-period kthread to context switches and to cond_resched().
Unfortunately, such urgent needs are not communicated to need_resched(),
which is sometimes used to decide when to invoke cond_resched(), for
but one example, within the KVM vcpu_run() function. As of v4.15, this
can result in synchronize_sched() being delayed by up to ten seconds,
which can be problematic, to say nothing of annoying.
This commit therefore checks rcu_dynticks.rcu_urgent_qs from within
rcu_check_callbacks(), which is invoked from the scheduling-clock
interrupt handler. If the current task is not an idle task and is
not executing in usermode, a context switch is forced, and either way,
the rcu_dynticks.rcu_urgent_qs variable is set to false. If the current
task is an idle task, then RCU's dyntick-idle code will detect the
quiescent state, so no further action is required. Similarly, if the
task is executing in usermode, other code in rcu_check_callbacks() and
its called functions will report the corresponding quiescent state.
Reported-by: Marius Hillenbrand <mhillenb@amazon.de>
Reported-by: David Woodhouse <dwmw2@infradead.org>
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-07-10 04:47:30 +08:00
|
|
|
|
/* The load-acquire pairs with the store-release setting to true. */
|
2018-08-04 12:00:38 +08:00
|
|
|
|
if (smp_load_acquire(this_cpu_ptr(&rcu_data.rcu_urgent_qs))) {
|
rcu: Make need_resched() respond to urgent RCU-QS needs
The per-CPU rcu_dynticks.rcu_urgent_qs variable communicates an urgent
need for an RCU quiescent state from the force-quiescent-state processing
within the grace-period kthread to context switches and to cond_resched().
Unfortunately, such urgent needs are not communicated to need_resched(),
which is sometimes used to decide when to invoke cond_resched(), for
but one example, within the KVM vcpu_run() function. As of v4.15, this
can result in synchronize_sched() being delayed by up to ten seconds,
which can be problematic, to say nothing of annoying.
This commit therefore checks rcu_dynticks.rcu_urgent_qs from within
rcu_check_callbacks(), which is invoked from the scheduling-clock
interrupt handler. If the current task is not an idle task and is
not executing in usermode, a context switch is forced, and either way,
the rcu_dynticks.rcu_urgent_qs variable is set to false. If the current
task is an idle task, then RCU's dyntick-idle code will detect the
quiescent state, so no further action is required. Similarly, if the
task is executing in usermode, other code in rcu_check_callbacks() and
its called functions will report the corresponding quiescent state.
Reported-by: Marius Hillenbrand <mhillenb@amazon.de>
Reported-by: David Woodhouse <dwmw2@infradead.org>
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-07-10 04:47:30 +08:00
|
|
|
|
/* Idle and userspace execution already are quiescent states. */
|
2018-07-10 06:50:16 +08:00
|
|
|
|
if (!rcu_is_cpu_rrupt_from_idle() && !user) {
|
rcu: Make need_resched() respond to urgent RCU-QS needs
The per-CPU rcu_dynticks.rcu_urgent_qs variable communicates an urgent
need for an RCU quiescent state from the force-quiescent-state processing
within the grace-period kthread to context switches and to cond_resched().
Unfortunately, such urgent needs are not communicated to need_resched(),
which is sometimes used to decide when to invoke cond_resched(), for
but one example, within the KVM vcpu_run() function. As of v4.15, this
can result in synchronize_sched() being delayed by up to ten seconds,
which can be problematic, to say nothing of annoying.
This commit therefore checks rcu_dynticks.rcu_urgent_qs from within
rcu_check_callbacks(), which is invoked from the scheduling-clock
interrupt handler. If the current task is not an idle task and is
not executing in usermode, a context switch is forced, and either way,
the rcu_dynticks.rcu_urgent_qs variable is set to false. If the current
task is an idle task, then RCU's dyntick-idle code will detect the
quiescent state, so no further action is required. Similarly, if the
task is executing in usermode, other code in rcu_check_callbacks() and
its called functions will report the corresponding quiescent state.
Reported-by: Marius Hillenbrand <mhillenb@amazon.de>
Reported-by: David Woodhouse <dwmw2@infradead.org>
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-07-10 04:47:30 +08:00
|
|
|
|
set_tsk_need_resched(current);
|
|
|
|
|
set_preempt_need_resched();
|
|
|
|
|
}
|
2018-08-04 12:00:38 +08:00
|
|
|
|
__this_cpu_write(rcu_data.rcu_urgent_qs, false);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
2018-11-22 03:35:03 +08:00
|
|
|
|
rcu_flavor_sched_clock_irq(user);
|
2019-09-14 18:39:22 +08:00
|
|
|
|
if (rcu_pending(user))
|
2011-06-16 06:47:09 +08:00
|
|
|
|
invoke_rcu_core();
|
rcu-tasks: Make RCU Tasks Trace check for userspace execution
Userspace execution is a valid quiescent state for RCU Tasks Trace,
but the scheduling-clock interrupt does not currently report such
quiescent states.
Of course, the scheduling-clock interrupt is not strictly speaking
userspace execution. However, the only way that this code is not
in a quiescent state is if something invoked rcu_read_lock_trace(),
and that would be reflected in the ->trc_reader_nesting field in
the task_struct structure. Furthermore, this field is checked by
rcu_tasks_trace_qs(), which is invoked by rcu_tasks_qs() which is in
turn invoked by rcu_note_voluntary_context_switch() in kernels building
at least one of the RCU Tasks flavors. It is therefore safe to invoke
rcu_tasks_trace_qs() from the rcu_sched_clock_irq().
But rcu_tasks_qs() also invokes rcu_tasks_classic_qs() for RCU
Tasks, which lacks the read-side markers provided by RCU Tasks Trace.
This raises the possibility that an RCU Tasks grace period could start
after the interrupt from userspace execution, but before the call to
rcu_sched_clock_irq(). However, it turns out that this is safe because
the RCU Tasks grace period waits for an RCU grace period, which will
wait for the entire scheduling-clock interrupt handler, including any
RCU Tasks read-side critical section that this handler might contain.
This commit therefore updates the rcu_sched_clock_irq() function's
check for usermode execution and its call to rcu_tasks_classic_qs()
to instead check for both usermode execution and interrupt from idle,
and to instead call rcu_note_voluntary_context_switch(). This
consolidates code and provides more faster RCU Tasks Trace
reporting of quiescent states in kernels that do scheduling-clock
interrupts for userspace execution.
[ paulmck: Consolidate checks into rcu_sched_clock_irq(). ]
Signed-off-by: Zqiang <qiang1.zhang@intel.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2022-07-19 12:39:00 +08:00
|
|
|
|
if (user || rcu_is_cpu_rrupt_from_idle())
|
|
|
|
|
rcu_note_voluntary_context_switch(current);
|
2020-11-20 02:13:06 +08:00
|
|
|
|
lockdep_assert_irqs_disabled();
|
2018-05-11 16:30:34 +08:00
|
|
|
|
|
2013-07-13 05:18:47 +08:00
|
|
|
|
trace_rcu_utilization(TPS("End scheduler-tick"));
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
2019-03-20 11:33:00 +08:00
|
|
|
|
* Scan the leaf rcu_node structures. For each structure on which all
|
|
|
|
|
* CPUs have reported a quiescent state and on which there are tasks
|
|
|
|
|
* blocking the current grace period, initiate RCU priority boosting.
|
|
|
|
|
* Otherwise, invoke the specified function to check dyntick state for
|
|
|
|
|
* each CPU that has not yet reported a quiescent state.
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
*/
|
2018-07-06 08:55:14 +08:00
|
|
|
|
static void force_qs_rnp(int (*f)(struct rcu_data *rdp))
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
{
|
|
|
|
|
int cpu;
|
|
|
|
|
unsigned long flags;
|
2009-09-28 22:46:33 +08:00
|
|
|
|
struct rcu_node *rnp;
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
2019-10-31 02:56:10 +08:00
|
|
|
|
rcu_state.cbovld = rcu_state.cbovldnext;
|
|
|
|
|
rcu_state.cbovldnext = false;
|
2018-07-05 05:33:59 +08:00
|
|
|
|
rcu_for_each_leaf_node(rnp) {
|
2023-10-31 16:53:08 +08:00
|
|
|
|
unsigned long mask = 0;
|
|
|
|
|
unsigned long rsmask = 0;
|
|
|
|
|
|
2018-03-03 08:35:27 +08:00
|
|
|
|
cond_resched_tasks_rcu_qs();
|
2015-10-08 18:24:23 +08:00
|
|
|
|
raw_spin_lock_irqsave_rcu_node(rnp, flags);
|
2019-10-31 02:56:10 +08:00
|
|
|
|
rcu_state.cbovldnext |= !!rnp->cbovldmask;
|
2009-09-28 22:46:33 +08:00
|
|
|
|
if (rnp->qsmask == 0) {
|
2020-06-23 02:07:03 +08:00
|
|
|
|
if (rcu_preempt_blocked_readers_cgp(rnp)) {
|
2015-03-09 05:52:27 +08:00
|
|
|
|
/*
|
|
|
|
|
* No point in scanning bits because they
|
|
|
|
|
* are all zero. But we might need to
|
|
|
|
|
* priority-boost blocked readers.
|
|
|
|
|
*/
|
|
|
|
|
rcu_initiate_boost(rnp, flags);
|
|
|
|
|
/* rcu_initiate_boost() releases rnp->lock */
|
|
|
|
|
continue;
|
|
|
|
|
}
|
2018-05-03 02:07:02 +08:00
|
|
|
|
raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
|
|
|
|
|
continue;
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
2019-10-31 00:37:11 +08:00
|
|
|
|
for_each_leaf_node_cpu_mask(rnp, cpu, rnp->qsmask) {
|
2023-10-31 16:53:08 +08:00
|
|
|
|
struct rcu_data *rdp;
|
|
|
|
|
int ret;
|
|
|
|
|
|
2019-10-31 00:37:11 +08:00
|
|
|
|
rdp = per_cpu_ptr(&rcu_data, cpu);
|
2023-10-31 16:53:08 +08:00
|
|
|
|
ret = f(rdp);
|
|
|
|
|
if (ret > 0) {
|
2019-10-31 00:37:11 +08:00
|
|
|
|
mask |= rdp->grpmask;
|
|
|
|
|
rcu_disable_urgency_upon_qs(rdp);
|
2013-06-22 07:37:22 +08:00
|
|
|
|
}
|
2023-10-31 16:53:08 +08:00
|
|
|
|
if (ret < 0)
|
|
|
|
|
rsmask |= rdp->grpmask;
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
2010-01-05 07:09:08 +08:00
|
|
|
|
if (mask != 0) {
|
2018-04-28 05:54:46 +08:00
|
|
|
|
/* Idle/offline CPUs, report (releases rnp->lock). */
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_report_qs_rnp(mask, rnp, rnp->gp_seq, flags);
|
rcu: Process offlining and onlining only at grace-period start
Races between CPU hotplug and grace periods can be difficult to resolve,
so the ->onoff_mutex is used to exclude the two events. Unfortunately,
this means that it is impossible for an outgoing CPU to perform the
last bits of its offlining from its last pass through the idle loop,
because sleeplocks cannot be acquired in that context.
This commit avoids these problems by buffering online and offline events
in a new ->qsmaskinitnext field in the leaf rcu_node structures. When a
grace period starts, the events accumulated in this mask are applied to
the ->qsmaskinit field, and, if needed, up the rcu_node tree. The special
case of all CPUs corresponding to a given leaf rcu_node structure being
offline while there are still elements in that structure's ->blkd_tasks
list is handled using a new ->wait_blkd_tasks field. In this case,
propagating the offline bits up the tree is deferred until the beginning
of the grace period after all of the tasks have exited their RCU read-side
critical sections and removed themselves from the list, at which point
the ->wait_blkd_tasks flag is cleared. If one of that leaf rcu_node
structure's CPUs comes back online before the list empties, then the
->wait_blkd_tasks flag is simply cleared.
This of course means that RCU's notion of which CPUs are offline can be
out of date. This is OK because RCU need only wait on CPUs that were
online at the time that the grace period started. In addition, RCU's
force-quiescent-state actions will handle the case where a CPU goes
offline after the grace period starts.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2015-01-24 13:52:37 +08:00
|
|
|
|
} else {
|
|
|
|
|
/* Nothing to do here, so just drop the lock. */
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
2023-10-31 16:53:08 +08:00
|
|
|
|
|
|
|
|
|
for_each_leaf_node_cpu_mask(rnp, cpu, rsmask)
|
|
|
|
|
resched_cpu(cpu);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Force quiescent states on reluctant CPUs, and also detect which
|
|
|
|
|
* CPUs are in dyntick-idle mode.
|
|
|
|
|
*/
|
2018-11-29 08:57:54 +08:00
|
|
|
|
void rcu_force_quiescent_state(void)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
{
|
|
|
|
|
unsigned long flags;
|
2012-06-27 08:00:35 +08:00
|
|
|
|
bool ret;
|
|
|
|
|
struct rcu_node *rnp;
|
|
|
|
|
struct rcu_node *rnp_old = NULL;
|
|
|
|
|
|
|
|
|
|
/* Funnel through hierarchy to reduce memory contention. */
|
2022-10-13 12:41:48 +08:00
|
|
|
|
rnp = raw_cpu_read(rcu_data.mynode);
|
2012-06-27 08:00:35 +08:00
|
|
|
|
for (; rnp != NULL; rnp = rnp->parent) {
|
2018-07-06 07:15:38 +08:00
|
|
|
|
ret = (READ_ONCE(rcu_state.gp_flags) & RCU_GP_FLAG_FQS) ||
|
2019-08-13 07:14:00 +08:00
|
|
|
|
!raw_spin_trylock(&rnp->fqslock);
|
2012-06-27 08:00:35 +08:00
|
|
|
|
if (rnp_old != NULL)
|
|
|
|
|
raw_spin_unlock(&rnp_old->fqslock);
|
2018-01-11 05:10:49 +08:00
|
|
|
|
if (ret)
|
2012-06-27 08:00:35 +08:00
|
|
|
|
return;
|
|
|
|
|
rnp_old = rnp;
|
|
|
|
|
}
|
2018-07-04 08:22:34 +08:00
|
|
|
|
/* rnp_old == rcu_get_root(), rnp == NULL. */
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
2012-06-27 08:00:35 +08:00
|
|
|
|
/* Reached the root of the rcu_node tree, acquire lock. */
|
2015-10-08 18:24:23 +08:00
|
|
|
|
raw_spin_lock_irqsave_rcu_node(rnp_old, flags);
|
2012-06-27 08:00:35 +08:00
|
|
|
|
raw_spin_unlock(&rnp_old->fqslock);
|
2018-07-06 07:15:38 +08:00
|
|
|
|
if (READ_ONCE(rcu_state.gp_flags) & RCU_GP_FLAG_FQS) {
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_unlock_irqrestore_rcu_node(rnp_old, flags);
|
2012-06-23 08:06:26 +08:00
|
|
|
|
return; /* Someone beat us to it. */
|
2010-01-05 07:09:09 +08:00
|
|
|
|
}
|
2018-07-06 07:15:38 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_flags,
|
|
|
|
|
READ_ONCE(rcu_state.gp_flags) | RCU_GP_FLAG_FQS);
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_unlock_irqrestore_rcu_node(rnp_old, flags);
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_gp_kthread_wake();
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
2018-11-29 08:57:54 +08:00
|
|
|
|
EXPORT_SYMBOL_GPL(rcu_force_quiescent_state);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
2020-08-08 22:56:31 +08:00
|
|
|
|
// Workqueue handler for an RCU reader for kernels enforcing struct RCU
|
|
|
|
|
// grace periods.
|
|
|
|
|
static void strict_work_handler(struct work_struct *work)
|
|
|
|
|
{
|
|
|
|
|
rcu_read_lock();
|
|
|
|
|
rcu_read_unlock();
|
|
|
|
|
}
|
|
|
|
|
|
2018-11-22 04:42:12 +08:00
|
|
|
|
/* Perform RCU core processing work for the current CPU. */
|
2019-03-21 05:13:33 +08:00
|
|
|
|
static __latent_entropy void rcu_core(void)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
{
|
|
|
|
|
unsigned long flags;
|
2018-07-04 06:37:16 +08:00
|
|
|
|
struct rcu_data *rdp = raw_cpu_ptr(&rcu_data);
|
2018-04-22 11:44:11 +08:00
|
|
|
|
struct rcu_node *rnp = rdp->mynode;
|
2021-10-19 08:08:08 +08:00
|
|
|
|
/*
|
|
|
|
|
* On RT rcu_core() can be preempted when IRQs aren't disabled.
|
|
|
|
|
* Therefore this function can race with concurrent NOCB (de-)offloading
|
|
|
|
|
* on this CPU and the below condition must be considered volatile.
|
|
|
|
|
* However if we race with:
|
|
|
|
|
*
|
|
|
|
|
* _ Offloading: In the worst case we accelerate or process callbacks
|
|
|
|
|
* concurrently with NOCB kthreads. We are guaranteed to
|
|
|
|
|
* call rcu_nocb_lock() if that happens.
|
|
|
|
|
*
|
|
|
|
|
* _ Deoffloading: In the worst case we miss callbacks acceleration or
|
|
|
|
|
* processing. This is fine because the early stage
|
|
|
|
|
* of deoffloading invokes rcu_core() after setting
|
|
|
|
|
* SEGCBLIST_RCU_CORE. So we guarantee that we'll process
|
|
|
|
|
* what could have been dismissed without the need to wait
|
|
|
|
|
* for the next rcu_pending() check in the next jiffy.
|
|
|
|
|
*/
|
2020-11-13 20:13:27 +08:00
|
|
|
|
const bool do_batch = !rcu_segcblist_completely_offloaded(&rdp->cblist);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
2018-07-04 08:22:34 +08:00
|
|
|
|
if (cpu_is_offline(smp_processor_id()))
|
|
|
|
|
return;
|
|
|
|
|
trace_rcu_utilization(TPS("Start RCU core"));
|
2017-03-26 03:46:01 +08:00
|
|
|
|
WARN_ON_ONCE(!rdp->beenonline);
|
2009-08-16 00:53:48 +08:00
|
|
|
|
|
rcu: Defer reporting RCU-preempt quiescent states when disabled
This commit defers reporting of RCU-preempt quiescent states at
rcu_read_unlock_special() time when any of interrupts, softirq, or
preemption are disabled. These deferred quiescent states are reported
at a later RCU_SOFTIRQ, context switch, idle entry, or CPU-hotplug
offline operation. Of course, if another RCU read-side critical
section has started in the meantime, the reporting of the quiescent
state will be further deferred.
This also means that disabling preemption, interrupts, and/or
softirqs will act as an RCU-preempt read-side critical section.
This is enforced by checking preempt_count() as needed.
Some special cases must be handled on an ad-hoc basis, for example,
context switch is a quiescent state even though both the scheduler and
do_exit() disable preemption. In these cases, additional calls to
rcu_preempt_deferred_qs() override the preemption disabling. Similar
logic overrides disabled interrupts in rcu_preempt_check_callbacks()
because in this case the quiescent state happened just before the
corresponding scheduling-clock interrupt.
In theory, this change lifts a long-standing restriction that required
that if interrupts were disabled across a call to rcu_read_unlock()
that the matching rcu_read_lock() also be contained within that
interrupts-disabled region of code. Because the reporting of the
corresponding RCU-preempt quiescent state is now deferred until
after interrupts have been enabled, it is no longer possible for this
situation to result in deadlocks involving the scheduler's runqueue and
priority-inheritance locks. This may allow some code simplification that
might reduce interrupt latency a bit. Unfortunately, in practice this
would also defer deboosting a low-priority task that had been subjected
to RCU priority boosting, so real-time-response considerations might
well force this restriction to remain in place.
Because RCU-preempt grace periods are now blocked not only by RCU
read-side critical sections, but also by disabling of interrupts,
preemption, and softirqs, it will be possible to eliminate RCU-bh and
RCU-sched in favor of RCU-preempt in CONFIG_PREEMPT=y kernels. This may
require some additional plumbing to provide the network denial-of-service
guarantees that have been traditionally provided by RCU-bh. Once these
are in place, CONFIG_PREEMPT=n kernels will be able to fold RCU-bh
into RCU-sched. This would mean that all kernels would have but
one flavor of RCU, which would open the door to significant code
cleanup.
Moving to a single flavor of RCU would also have the beneficial effect
of reducing the NOCB kthreads by at least a factor of two.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
[ paulmck: Apply rcu_read_unlock_special() preempt_count() feedback
from Joel Fernandes. ]
[ paulmck: Adjust rcu_eqs_enter() call to rcu_preempt_deferred_qs() in
response to bug reports from kbuild test robot. ]
[ paulmck: Fix bug located by kbuild test robot involving recursion
via rcu_preempt_deferred_qs(). ]
2018-06-22 03:50:01 +08:00
|
|
|
|
/* Report any deferred quiescent states if preemption enabled. */
|
2021-09-30 02:09:34 +08:00
|
|
|
|
if (IS_ENABLED(CONFIG_PREEMPT_COUNT) && (!(preempt_count() & PREEMPT_MASK))) {
|
rcu: Defer reporting RCU-preempt quiescent states when disabled
This commit defers reporting of RCU-preempt quiescent states at
rcu_read_unlock_special() time when any of interrupts, softirq, or
preemption are disabled. These deferred quiescent states are reported
at a later RCU_SOFTIRQ, context switch, idle entry, or CPU-hotplug
offline operation. Of course, if another RCU read-side critical
section has started in the meantime, the reporting of the quiescent
state will be further deferred.
This also means that disabling preemption, interrupts, and/or
softirqs will act as an RCU-preempt read-side critical section.
This is enforced by checking preempt_count() as needed.
Some special cases must be handled on an ad-hoc basis, for example,
context switch is a quiescent state even though both the scheduler and
do_exit() disable preemption. In these cases, additional calls to
rcu_preempt_deferred_qs() override the preemption disabling. Similar
logic overrides disabled interrupts in rcu_preempt_check_callbacks()
because in this case the quiescent state happened just before the
corresponding scheduling-clock interrupt.
In theory, this change lifts a long-standing restriction that required
that if interrupts were disabled across a call to rcu_read_unlock()
that the matching rcu_read_lock() also be contained within that
interrupts-disabled region of code. Because the reporting of the
corresponding RCU-preempt quiescent state is now deferred until
after interrupts have been enabled, it is no longer possible for this
situation to result in deadlocks involving the scheduler's runqueue and
priority-inheritance locks. This may allow some code simplification that
might reduce interrupt latency a bit. Unfortunately, in practice this
would also defer deboosting a low-priority task that had been subjected
to RCU priority boosting, so real-time-response considerations might
well force this restriction to remain in place.
Because RCU-preempt grace periods are now blocked not only by RCU
read-side critical sections, but also by disabling of interrupts,
preemption, and softirqs, it will be possible to eliminate RCU-bh and
RCU-sched in favor of RCU-preempt in CONFIG_PREEMPT=y kernels. This may
require some additional plumbing to provide the network denial-of-service
guarantees that have been traditionally provided by RCU-bh. Once these
are in place, CONFIG_PREEMPT=n kernels will be able to fold RCU-bh
into RCU-sched. This would mean that all kernels would have but
one flavor of RCU, which would open the door to significant code
cleanup.
Moving to a single flavor of RCU would also have the beneficial effect
of reducing the NOCB kthreads by at least a factor of two.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
[ paulmck: Apply rcu_read_unlock_special() preempt_count() feedback
from Joel Fernandes. ]
[ paulmck: Adjust rcu_eqs_enter() call to rcu_preempt_deferred_qs() in
response to bug reports from kbuild test robot. ]
[ paulmck: Fix bug located by kbuild test robot involving recursion
via rcu_preempt_deferred_qs(). ]
2018-06-22 03:50:01 +08:00
|
|
|
|
rcu_preempt_deferred_qs(current);
|
2018-07-27 04:44:00 +08:00
|
|
|
|
} else if (rcu_preempt_need_deferred_qs(current)) {
|
|
|
|
|
set_tsk_need_resched(current);
|
|
|
|
|
set_preempt_need_resched();
|
|
|
|
|
}
|
rcu: Defer reporting RCU-preempt quiescent states when disabled
This commit defers reporting of RCU-preempt quiescent states at
rcu_read_unlock_special() time when any of interrupts, softirq, or
preemption are disabled. These deferred quiescent states are reported
at a later RCU_SOFTIRQ, context switch, idle entry, or CPU-hotplug
offline operation. Of course, if another RCU read-side critical
section has started in the meantime, the reporting of the quiescent
state will be further deferred.
This also means that disabling preemption, interrupts, and/or
softirqs will act as an RCU-preempt read-side critical section.
This is enforced by checking preempt_count() as needed.
Some special cases must be handled on an ad-hoc basis, for example,
context switch is a quiescent state even though both the scheduler and
do_exit() disable preemption. In these cases, additional calls to
rcu_preempt_deferred_qs() override the preemption disabling. Similar
logic overrides disabled interrupts in rcu_preempt_check_callbacks()
because in this case the quiescent state happened just before the
corresponding scheduling-clock interrupt.
In theory, this change lifts a long-standing restriction that required
that if interrupts were disabled across a call to rcu_read_unlock()
that the matching rcu_read_lock() also be contained within that
interrupts-disabled region of code. Because the reporting of the
corresponding RCU-preempt quiescent state is now deferred until
after interrupts have been enabled, it is no longer possible for this
situation to result in deadlocks involving the scheduler's runqueue and
priority-inheritance locks. This may allow some code simplification that
might reduce interrupt latency a bit. Unfortunately, in practice this
would also defer deboosting a low-priority task that had been subjected
to RCU priority boosting, so real-time-response considerations might
well force this restriction to remain in place.
Because RCU-preempt grace periods are now blocked not only by RCU
read-side critical sections, but also by disabling of interrupts,
preemption, and softirqs, it will be possible to eliminate RCU-bh and
RCU-sched in favor of RCU-preempt in CONFIG_PREEMPT=y kernels. This may
require some additional plumbing to provide the network denial-of-service
guarantees that have been traditionally provided by RCU-bh. Once these
are in place, CONFIG_PREEMPT=n kernels will be able to fold RCU-bh
into RCU-sched. This would mean that all kernels would have but
one flavor of RCU, which would open the door to significant code
cleanup.
Moving to a single flavor of RCU would also have the beneficial effect
of reducing the NOCB kthreads by at least a factor of two.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
[ paulmck: Apply rcu_read_unlock_special() preempt_count() feedback
from Joel Fernandes. ]
[ paulmck: Adjust rcu_eqs_enter() call to rcu_preempt_deferred_qs() in
response to bug reports from kbuild test robot. ]
[ paulmck: Fix bug located by kbuild test robot involving recursion
via rcu_preempt_deferred_qs(). ]
2018-06-22 03:50:01 +08:00
|
|
|
|
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
/* Update RCU state based on any recent quiescent states. */
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_check_quiescent_state(rdp);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
rcu: Switch __rcu_process_callbacks() to rcu_accelerate_cbs()
The __rcu_process_callbacks() function currently checks to see if
the current CPU needs a grace period and also if there is any other
reason to kick off a new grace period. This is one of the fail-safe
checks that has been rendered unnecessary by the changes that increase
the accuracy of rcu_gp_cleanup()'s estimate as to whether another grace
period is required. Because this particular fail-safe involved acquiring
the root rcu_node structure's ->lock, which has seen excessive contention
in real life, this fail-safe needs to go.
However, one check must remain, namely the check for newly arrived
RCU callbacks that have not yet been associated with a grace period.
One might hope that the checks in __note_gp_changes(), which is invoked
indirectly from rcu_check_quiescent_state(), would suffice, but this
function won't be invoked at all if RCU is idle. It is therefore necessary
to replace the fail-safe checks with a simpler check for newly arrived
callbacks during an RCU idle period, which is exactly what this commit
does. This change removes the final call to rcu_start_gp(), so this
function is removed as well.
Note that lockless use of cpu_needs_another_gp() is racy, but that
these races are harmless in this case. If RCU really is idle, the
values will not change, so the return value from cpu_needs_another_gp()
will be correct. If RCU is not idle, the resulting redundant call to
rcu_accelerate_cbs() will be harmless, and might even have the benefit
of reducing grace-period latency a bit.
This commit also moves interrupt disabling into the "if" statement to
improve real-time response a bit.
Reported-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
2018-04-12 00:51:20 +08:00
|
|
|
|
/* No grace period and unregistered callbacks? */
|
2018-07-04 08:22:34 +08:00
|
|
|
|
if (!rcu_gp_in_progress() &&
|
2020-11-13 20:13:28 +08:00
|
|
|
|
rcu_segcblist_is_enabled(&rdp->cblist) && do_batch) {
|
|
|
|
|
rcu_nocb_lock_irqsave(rdp, flags);
|
2018-05-02 07:29:47 +08:00
|
|
|
|
if (!rcu_segcblist_restempty(&rdp->cblist, RCU_NEXT_READY_TAIL))
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_accelerate_cbs_unlocked(rnp, rdp);
|
2020-11-13 20:13:28 +08:00
|
|
|
|
rcu_nocb_unlock_irqrestore(rdp, flags);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
|
|
|
|
|
2018-10-02 06:42:44 +08:00
|
|
|
|
rcu_check_gp_start_stall(rnp, rdp, rcu_jiffies_till_stall_check());
|
2018-04-22 11:44:11 +08:00
|
|
|
|
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
/* If there are callbacks ready, invoke them. */
|
2020-11-13 20:13:27 +08:00
|
|
|
|
if (do_batch && rcu_segcblist_ready_cbs(&rdp->cblist) &&
|
2021-10-19 08:08:16 +08:00
|
|
|
|
likely(READ_ONCE(rcu_scheduler_fully_active))) {
|
2019-03-25 23:36:03 +08:00
|
|
|
|
rcu_do_batch(rdp);
|
2021-10-19 08:08:16 +08:00
|
|
|
|
/* Re-invoke RCU core processing if there are callbacks remaining. */
|
|
|
|
|
if (rcu_segcblist_ready_cbs(&rdp->cblist))
|
|
|
|
|
invoke_rcu_core();
|
|
|
|
|
}
|
rcu: Break call_rcu() deadlock involving scheduler and perf
Dave Jones got the following lockdep splat:
> ======================================================
> [ INFO: possible circular locking dependency detected ]
> 3.12.0-rc3+ #92 Not tainted
> -------------------------------------------------------
> trinity-child2/15191 is trying to acquire lock:
> (&rdp->nocb_wq){......}, at: [<ffffffff8108ff43>] __wake_up+0x23/0x50
>
> but task is already holding lock:
> (&ctx->lock){-.-...}, at: [<ffffffff81154c19>] perf_event_exit_task+0x109/0x230
>
> which lock already depends on the new lock.
>
>
> the existing dependency chain (in reverse order) is:
>
> -> #3 (&ctx->lock){-.-...}:
> [<ffffffff810cc243>] lock_acquire+0x93/0x200
> [<ffffffff81733f90>] _raw_spin_lock+0x40/0x80
> [<ffffffff811500ff>] __perf_event_task_sched_out+0x2df/0x5e0
> [<ffffffff81091b83>] perf_event_task_sched_out+0x93/0xa0
> [<ffffffff81732052>] __schedule+0x1d2/0xa20
> [<ffffffff81732f30>] preempt_schedule_irq+0x50/0xb0
> [<ffffffff817352b6>] retint_kernel+0x26/0x30
> [<ffffffff813eed04>] tty_flip_buffer_push+0x34/0x50
> [<ffffffff813f0504>] pty_write+0x54/0x60
> [<ffffffff813e900d>] n_tty_write+0x32d/0x4e0
> [<ffffffff813e5838>] tty_write+0x158/0x2d0
> [<ffffffff811c4850>] vfs_write+0xc0/0x1f0
> [<ffffffff811c52cc>] SyS_write+0x4c/0xa0
> [<ffffffff8173d4e4>] tracesys+0xdd/0xe2
>
> -> #2 (&rq->lock){-.-.-.}:
> [<ffffffff810cc243>] lock_acquire+0x93/0x200
> [<ffffffff81733f90>] _raw_spin_lock+0x40/0x80
> [<ffffffff810980b2>] wake_up_new_task+0xc2/0x2e0
> [<ffffffff81054336>] do_fork+0x126/0x460
> [<ffffffff81054696>] kernel_thread+0x26/0x30
> [<ffffffff8171ff93>] rest_init+0x23/0x140
> [<ffffffff81ee1e4b>] start_kernel+0x3f6/0x403
> [<ffffffff81ee1571>] x86_64_start_reservations+0x2a/0x2c
> [<ffffffff81ee1664>] x86_64_start_kernel+0xf1/0xf4
>
> -> #1 (&p->pi_lock){-.-.-.}:
> [<ffffffff810cc243>] lock_acquire+0x93/0x200
> [<ffffffff8173419b>] _raw_spin_lock_irqsave+0x4b/0x90
> [<ffffffff810979d1>] try_to_wake_up+0x31/0x350
> [<ffffffff81097d62>] default_wake_function+0x12/0x20
> [<ffffffff81084af8>] autoremove_wake_function+0x18/0x40
> [<ffffffff8108ea38>] __wake_up_common+0x58/0x90
> [<ffffffff8108ff59>] __wake_up+0x39/0x50
> [<ffffffff8110d4f8>] __call_rcu_nocb_enqueue+0xa8/0xc0
> [<ffffffff81111450>] __call_rcu+0x140/0x820
> [<ffffffff81111b8d>] call_rcu+0x1d/0x20
> [<ffffffff81093697>] cpu_attach_domain+0x287/0x360
> [<ffffffff81099d7e>] build_sched_domains+0xe5e/0x10a0
> [<ffffffff81efa7fc>] sched_init_smp+0x3b7/0x47a
> [<ffffffff81ee1f4e>] kernel_init_freeable+0xf6/0x202
> [<ffffffff817200be>] kernel_init+0xe/0x190
> [<ffffffff8173d22c>] ret_from_fork+0x7c/0xb0
>
> -> #0 (&rdp->nocb_wq){......}:
> [<ffffffff810cb7ca>] __lock_acquire+0x191a/0x1be0
> [<ffffffff810cc243>] lock_acquire+0x93/0x200
> [<ffffffff8173419b>] _raw_spin_lock_irqsave+0x4b/0x90
> [<ffffffff8108ff43>] __wake_up+0x23/0x50
> [<ffffffff8110d4f8>] __call_rcu_nocb_enqueue+0xa8/0xc0
> [<ffffffff81111450>] __call_rcu+0x140/0x820
> [<ffffffff81111bb0>] kfree_call_rcu+0x20/0x30
> [<ffffffff81149abf>] put_ctx+0x4f/0x70
> [<ffffffff81154c3e>] perf_event_exit_task+0x12e/0x230
> [<ffffffff81056b8d>] do_exit+0x30d/0xcc0
> [<ffffffff8105893c>] do_group_exit+0x4c/0xc0
> [<ffffffff810589c4>] SyS_exit_group+0x14/0x20
> [<ffffffff8173d4e4>] tracesys+0xdd/0xe2
>
> other info that might help us debug this:
>
> Chain exists of:
> &rdp->nocb_wq --> &rq->lock --> &ctx->lock
>
> Possible unsafe locking scenario:
>
> CPU0 CPU1
> ---- ----
> lock(&ctx->lock);
> lock(&rq->lock);
> lock(&ctx->lock);
> lock(&rdp->nocb_wq);
>
> *** DEADLOCK ***
>
> 1 lock held by trinity-child2/15191:
> #0: (&ctx->lock){-.-...}, at: [<ffffffff81154c19>] perf_event_exit_task+0x109/0x230
>
> stack backtrace:
> CPU: 2 PID: 15191 Comm: trinity-child2 Not tainted 3.12.0-rc3+ #92
> ffffffff82565b70 ffff880070c2dbf8 ffffffff8172a363 ffffffff824edf40
> ffff880070c2dc38 ffffffff81726741 ffff880070c2dc90 ffff88022383b1c0
> ffff88022383aac0 0000000000000000 ffff88022383b188 ffff88022383b1c0
> Call Trace:
> [<ffffffff8172a363>] dump_stack+0x4e/0x82
> [<ffffffff81726741>] print_circular_bug+0x200/0x20f
> [<ffffffff810cb7ca>] __lock_acquire+0x191a/0x1be0
> [<ffffffff810c6439>] ? get_lock_stats+0x19/0x60
> [<ffffffff8100b2f4>] ? native_sched_clock+0x24/0x80
> [<ffffffff810cc243>] lock_acquire+0x93/0x200
> [<ffffffff8108ff43>] ? __wake_up+0x23/0x50
> [<ffffffff8173419b>] _raw_spin_lock_irqsave+0x4b/0x90
> [<ffffffff8108ff43>] ? __wake_up+0x23/0x50
> [<ffffffff8108ff43>] __wake_up+0x23/0x50
> [<ffffffff8110d4f8>] __call_rcu_nocb_enqueue+0xa8/0xc0
> [<ffffffff81111450>] __call_rcu+0x140/0x820
> [<ffffffff8109bc8f>] ? local_clock+0x3f/0x50
> [<ffffffff81111bb0>] kfree_call_rcu+0x20/0x30
> [<ffffffff81149abf>] put_ctx+0x4f/0x70
> [<ffffffff81154c3e>] perf_event_exit_task+0x12e/0x230
> [<ffffffff81056b8d>] do_exit+0x30d/0xcc0
> [<ffffffff810c9af5>] ? trace_hardirqs_on_caller+0x115/0x1e0
> [<ffffffff810c9bcd>] ? trace_hardirqs_on+0xd/0x10
> [<ffffffff8105893c>] do_group_exit+0x4c/0xc0
> [<ffffffff810589c4>] SyS_exit_group+0x14/0x20
> [<ffffffff8173d4e4>] tracesys+0xdd/0xe2
The underlying problem is that perf is invoking call_rcu() with the
scheduler locks held, but in NOCB mode, call_rcu() will with high
probability invoke the scheduler -- which just might want to use its
locks. The reason that call_rcu() needs to invoke the scheduler is
to wake up the corresponding rcuo callback-offload kthread, which
does the job of starting up a grace period and invoking the callbacks
afterwards.
One solution (championed on a related problem by Lai Jiangshan) is to
simply defer the wakeup to some point where scheduler locks are no longer
held. Since we don't want to unnecessarily incur the cost of such
deferral, the task before us is threefold:
1. Determine when it is likely that a relevant scheduler lock is held.
2. Defer the wakeup in such cases.
3. Ensure that all deferred wakeups eventually happen, preferably
sooner rather than later.
We use irqs_disabled_flags() as a proxy for relevant scheduler locks
being held. This works because the relevant locks are always acquired
with interrupts disabled. We may defer more often than needed, but that
is at least safe.
The wakeup deferral is tracked via a new field in the per-CPU and
per-RCU-flavor rcu_data structure, namely ->nocb_defer_wakeup.
This flag is checked by the RCU core processing. The __rcu_pending()
function now checks this flag, which causes rcu_check_callbacks()
to initiate RCU core processing at each scheduling-clock interrupt
where this flag is set. Of course this is not sufficient because
scheduling-clock interrupts are often turned off (the things we used to
be able to count on!). So the flags are also checked on entry to any
state that RCU considers to be idle, which includes both NO_HZ_IDLE idle
state and NO_HZ_FULL user-mode-execution state.
This approach should allow call_rcu() to be invoked regardless of what
locks you might be holding, the key word being "should".
Reported-by: Dave Jones <davej@redhat.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
2013-10-05 05:33:34 +08:00
|
|
|
|
|
|
|
|
|
/* Do any needed deferred wakeups of rcuo kthreads. */
|
|
|
|
|
do_nocb_deferred_wakeup(rdp);
|
2013-07-13 05:18:47 +08:00
|
|
|
|
trace_rcu_utilization(TPS("End RCU core"));
|
2020-08-08 22:56:31 +08:00
|
|
|
|
|
|
|
|
|
// If strict GPs, schedule an RCU reader in a clean environment.
|
|
|
|
|
if (IS_ENABLED(CONFIG_RCU_STRICT_GRACE_PERIOD))
|
|
|
|
|
queue_work_on(rdp->cpu, rcu_gp_wq, &rdp->strict_work);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
|
|
|
|
|
2019-03-21 05:13:33 +08:00
|
|
|
|
static void rcu_core_si(struct softirq_action *h)
|
|
|
|
|
{
|
|
|
|
|
rcu_core();
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void rcu_wake_cond(struct task_struct *t, int status)
|
|
|
|
|
{
|
|
|
|
|
/*
|
|
|
|
|
* If the thread is yielding, only wake it when this
|
|
|
|
|
* is invoked from idle
|
|
|
|
|
*/
|
|
|
|
|
if (t && (status != RCU_KTHREAD_YIELDING || is_idle_task(current)))
|
|
|
|
|
wake_up_process(t);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void invoke_rcu_core_kthread(void)
|
|
|
|
|
{
|
|
|
|
|
struct task_struct *t;
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
|
|
local_irq_save(flags);
|
|
|
|
|
__this_cpu_write(rcu_data.rcu_cpu_has_work, 1);
|
|
|
|
|
t = __this_cpu_read(rcu_data.rcu_cpu_kthread_task);
|
|
|
|
|
if (t != NULL && t != current)
|
|
|
|
|
rcu_wake_cond(t, __this_cpu_read(rcu_data.rcu_cpu_kthread_status));
|
|
|
|
|
local_irq_restore(flags);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Wake up this CPU's rcuc kthread to do RCU core processing.
|
|
|
|
|
*/
|
2011-06-16 06:47:09 +08:00
|
|
|
|
static void invoke_rcu_core(void)
|
rcu: Use softirq to address performance regression
Commit a26ac2455ffcf3(rcu: move TREE_RCU from softirq to kthread)
introduced performance regression. In an AIM7 test, this commit degraded
performance by about 40%.
The commit runs rcu callbacks in a kthread instead of softirq. We observed
high rate of context switch which is caused by this. Out test system has
64 CPUs and HZ is 1000, so we saw more than 64k context switch per second
which is caused by RCU's per-CPU kthread. A trace showed that most of
the time the RCU per-CPU kthread doesn't actually handle any callbacks,
but instead just does a very small amount of work handling grace periods.
This means that RCU's per-CPU kthreads are making the scheduler do quite
a bit of work in order to allow a very small amount of RCU-related
processing to be done.
Alex Shi's analysis determined that this slowdown is due to lock
contention within the scheduler. Unfortunately, as Peter Zijlstra points
out, the scheduler's real-time semantics require global action, which
means that this contention is inherent in real-time scheduling. (Yes,
perhaps someone will come up with a workaround -- otherwise, -rt is not
going to do well on large SMP systems -- but this patch will work around
this issue in the meantime. And "the meantime" might well be forever.)
This patch therefore re-introduces softirq processing to RCU, but only
for core RCU work. RCU callbacks are still executed in kthread context,
so that only a small amount of RCU work runs in softirq context in the
common case. This should minimize ksoftirqd execution, allowing us to
skip boosting of ksoftirqd for CONFIG_RCU_BOOST=y kernels.
Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Tested-by: "Alex,Shi" <alex.shi@intel.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2011-06-14 13:26:25 +08:00
|
|
|
|
{
|
2019-03-21 05:13:33 +08:00
|
|
|
|
if (!cpu_online(smp_processor_id()))
|
|
|
|
|
return;
|
|
|
|
|
if (use_softirq)
|
2013-02-05 04:14:24 +08:00
|
|
|
|
raise_softirq(RCU_SOFTIRQ);
|
2019-03-21 05:13:33 +08:00
|
|
|
|
else
|
|
|
|
|
invoke_rcu_core_kthread();
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void rcu_cpu_kthread_park(unsigned int cpu)
|
|
|
|
|
{
|
|
|
|
|
per_cpu(rcu_data.rcu_cpu_kthread_status, cpu) = RCU_KTHREAD_OFFCPU;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int rcu_cpu_kthread_should_run(unsigned int cpu)
|
|
|
|
|
{
|
|
|
|
|
return __this_cpu_read(rcu_data.rcu_cpu_has_work);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Per-CPU kernel thread that invokes RCU callbacks. This replaces
|
|
|
|
|
* the RCU softirq used in configurations of RCU that do not support RCU
|
|
|
|
|
* priority boosting.
|
|
|
|
|
*/
|
|
|
|
|
static void rcu_cpu_kthread(unsigned int cpu)
|
|
|
|
|
{
|
|
|
|
|
unsigned int *statusp = this_cpu_ptr(&rcu_data.rcu_cpu_kthread_status);
|
|
|
|
|
char work, *workp = this_cpu_ptr(&rcu_data.rcu_cpu_has_work);
|
2022-01-25 10:47:44 +08:00
|
|
|
|
unsigned long *j = this_cpu_ptr(&rcu_data.rcuc_activity);
|
2019-03-21 05:13:33 +08:00
|
|
|
|
int spincnt;
|
|
|
|
|
|
2019-10-15 18:23:57 +08:00
|
|
|
|
trace_rcu_utilization(TPS("Start CPU kthread@rcu_run"));
|
2019-03-21 05:13:33 +08:00
|
|
|
|
for (spincnt = 0; spincnt < 10; spincnt++) {
|
2022-01-25 10:47:44 +08:00
|
|
|
|
WRITE_ONCE(*j, jiffies);
|
2019-03-21 05:13:33 +08:00
|
|
|
|
local_bh_disable();
|
|
|
|
|
*statusp = RCU_KTHREAD_RUNNING;
|
|
|
|
|
local_irq_disable();
|
|
|
|
|
work = *workp;
|
2023-04-08 07:47:34 +08:00
|
|
|
|
WRITE_ONCE(*workp, 0);
|
2019-03-21 05:13:33 +08:00
|
|
|
|
local_irq_enable();
|
|
|
|
|
if (work)
|
|
|
|
|
rcu_core();
|
|
|
|
|
local_bh_enable();
|
2023-04-08 07:47:34 +08:00
|
|
|
|
if (!READ_ONCE(*workp)) {
|
2019-03-21 05:13:33 +08:00
|
|
|
|
trace_rcu_utilization(TPS("End CPU kthread@rcu_wait"));
|
|
|
|
|
*statusp = RCU_KTHREAD_WAITING;
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
*statusp = RCU_KTHREAD_YIELDING;
|
|
|
|
|
trace_rcu_utilization(TPS("Start CPU kthread@rcu_yield"));
|
2020-05-08 06:44:46 +08:00
|
|
|
|
schedule_timeout_idle(2);
|
2019-03-21 05:13:33 +08:00
|
|
|
|
trace_rcu_utilization(TPS("End CPU kthread@rcu_yield"));
|
|
|
|
|
*statusp = RCU_KTHREAD_WAITING;
|
2022-01-25 10:47:44 +08:00
|
|
|
|
WRITE_ONCE(*j, jiffies);
|
2019-03-21 05:13:33 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static struct smp_hotplug_thread rcu_cpu_thread_spec = {
|
|
|
|
|
.store = &rcu_data.rcu_cpu_kthread_task,
|
|
|
|
|
.thread_should_run = rcu_cpu_kthread_should_run,
|
|
|
|
|
.thread_fn = rcu_cpu_kthread,
|
|
|
|
|
.thread_comm = "rcuc/%u",
|
|
|
|
|
.setup = rcu_cpu_kthread_setup,
|
|
|
|
|
.park = rcu_cpu_kthread_park,
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Spawn per-CPU RCU core processing kthreads.
|
|
|
|
|
*/
|
|
|
|
|
static int __init rcu_spawn_core_kthreads(void)
|
|
|
|
|
{
|
|
|
|
|
int cpu;
|
|
|
|
|
|
|
|
|
|
for_each_possible_cpu(cpu)
|
|
|
|
|
per_cpu(rcu_data.rcu_cpu_has_work, cpu) = 0;
|
2021-12-29 00:05:10 +08:00
|
|
|
|
if (use_softirq)
|
2019-03-21 05:13:33 +08:00
|
|
|
|
return 0;
|
|
|
|
|
WARN_ONCE(smpboot_register_percpu_thread(&rcu_cpu_thread_spec),
|
|
|
|
|
"%s: Could not start rcuc kthread, OOM is now expected behavior\n", __func__);
|
|
|
|
|
return 0;
|
rcu: Use softirq to address performance regression
Commit a26ac2455ffcf3(rcu: move TREE_RCU from softirq to kthread)
introduced performance regression. In an AIM7 test, this commit degraded
performance by about 40%.
The commit runs rcu callbacks in a kthread instead of softirq. We observed
high rate of context switch which is caused by this. Out test system has
64 CPUs and HZ is 1000, so we saw more than 64k context switch per second
which is caused by RCU's per-CPU kthread. A trace showed that most of
the time the RCU per-CPU kthread doesn't actually handle any callbacks,
but instead just does a very small amount of work handling grace periods.
This means that RCU's per-CPU kthreads are making the scheduler do quite
a bit of work in order to allow a very small amount of RCU-related
processing to be done.
Alex Shi's analysis determined that this slowdown is due to lock
contention within the scheduler. Unfortunately, as Peter Zijlstra points
out, the scheduler's real-time semantics require global action, which
means that this contention is inherent in real-time scheduling. (Yes,
perhaps someone will come up with a workaround -- otherwise, -rt is not
going to do well on large SMP systems -- but this patch will work around
this issue in the meantime. And "the meantime" might well be forever.)
This patch therefore re-introduces softirq processing to RCU, but only
for core RCU work. RCU callbacks are still executed in kthread context,
so that only a small amount of RCU work runs in softirq context in the
common case. This should minimize ksoftirqd execution, allowing us to
skip boosting of ksoftirqd for CONFIG_RCU_BOOST=y kernels.
Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Tested-by: "Alex,Shi" <alex.shi@intel.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2011-06-14 13:26:25 +08:00
|
|
|
|
}
|
|
|
|
|
|
2012-05-30 18:21:48 +08:00
|
|
|
|
/*
|
|
|
|
|
* Handle any core-RCU processing required by a call_rcu() invocation.
|
|
|
|
|
*/
|
2018-07-04 08:22:34 +08:00
|
|
|
|
static void __call_rcu_core(struct rcu_data *rdp, struct rcu_head *head,
|
|
|
|
|
unsigned long flags)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
{
|
2012-05-23 13:10:24 +08:00
|
|
|
|
/*
|
|
|
|
|
* If called from an extended quiescent state, invoke the RCU
|
|
|
|
|
* core in order to force a re-evaluation of RCU's idleness.
|
|
|
|
|
*/
|
2015-02-25 17:09:46 +08:00
|
|
|
|
if (!rcu_is_watching())
|
2012-05-23 13:10:24 +08:00
|
|
|
|
invoke_rcu_core();
|
|
|
|
|
|
2012-05-26 23:56:01 +08:00
|
|
|
|
/* If interrupts were disabled or CPU offline, don't invoke RCU core. */
|
2012-05-30 18:21:48 +08:00
|
|
|
|
if (irqs_disabled_flags(flags) || cpu_is_offline(smp_processor_id()))
|
2011-04-08 13:47:23 +08:00
|
|
|
|
return;
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
2009-10-15 01:15:55 +08:00
|
|
|
|
/*
|
|
|
|
|
* Force the grace period if too many callbacks or too long waiting.
|
2018-11-29 08:57:54 +08:00
|
|
|
|
* Enforce hysteresis, and don't invoke rcu_force_quiescent_state()
|
2009-10-15 01:15:55 +08:00
|
|
|
|
* if some other CPU has recently done so. Also, don't bother
|
2018-11-29 08:57:54 +08:00
|
|
|
|
* invoking rcu_force_quiescent_state() if the newly enqueued callback
|
2009-10-15 01:15:55 +08:00
|
|
|
|
* is the only one waiting for a grace period to complete.
|
|
|
|
|
*/
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
if (unlikely(rcu_segcblist_n_cbs(&rdp->cblist) >
|
|
|
|
|
rdp->qlen_last_fqs_check + qhimark)) {
|
2010-12-15 09:36:02 +08:00
|
|
|
|
|
|
|
|
|
/* Are we ignoring a completed grace period? */
|
2018-07-04 08:22:34 +08:00
|
|
|
|
note_gp_changes(rdp);
|
2010-12-15 09:36:02 +08:00
|
|
|
|
|
|
|
|
|
/* Start a new grace period if one not already started. */
|
2018-07-04 08:22:34 +08:00
|
|
|
|
if (!rcu_gp_in_progress()) {
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_accelerate_cbs_unlocked(rdp->mynode, rdp);
|
2010-12-15 09:36:02 +08:00
|
|
|
|
} else {
|
|
|
|
|
/* Give the grace period a kick. */
|
2019-04-11 08:01:39 +08:00
|
|
|
|
rdp->blimit = DEFAULT_MAX_RCU_BLIMIT;
|
2021-07-20 21:16:27 +08:00
|
|
|
|
if (READ_ONCE(rcu_state.n_force_qs) == rdp->n_force_qs_snap &&
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
rcu_segcblist_first_pend_cb(&rdp->cblist) != head)
|
2018-11-29 08:57:54 +08:00
|
|
|
|
rcu_force_quiescent_state();
|
2021-07-20 21:16:27 +08:00
|
|
|
|
rdp->n_force_qs_snap = READ_ONCE(rcu_state.n_force_qs);
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
rdp->qlen_last_fqs_check = rcu_segcblist_n_cbs(&rdp->cblist);
|
2010-12-15 09:36:02 +08:00
|
|
|
|
}
|
2012-06-23 08:06:26 +08:00
|
|
|
|
}
|
2012-05-30 18:21:48 +08:00
|
|
|
|
}
|
|
|
|
|
|
2013-04-24 04:20:57 +08:00
|
|
|
|
/*
|
|
|
|
|
* RCU callback function to leak a callback.
|
|
|
|
|
*/
|
|
|
|
|
static void rcu_leak_callback(struct rcu_head *rhp)
|
|
|
|
|
{
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-20 12:35:53 +08:00
|
|
|
|
/*
|
2019-10-31 02:56:10 +08:00
|
|
|
|
* Check and if necessary update the leaf rcu_node structure's
|
|
|
|
|
* ->cbovldmask bit corresponding to the current CPU based on that CPU's
|
|
|
|
|
* number of queued RCU callbacks. The caller must hold the leaf rcu_node
|
|
|
|
|
* structure's ->lock.
|
2012-08-20 12:35:53 +08:00
|
|
|
|
*/
|
2019-10-31 02:56:10 +08:00
|
|
|
|
static void check_cb_ovld_locked(struct rcu_data *rdp, struct rcu_node *rnp)
|
|
|
|
|
{
|
|
|
|
|
raw_lockdep_assert_held_rcu_node(rnp);
|
|
|
|
|
if (qovld_calc <= 0)
|
|
|
|
|
return; // Early boot and wildcard value set.
|
|
|
|
|
if (rcu_segcblist_n_cbs(&rdp->cblist) >= qovld_calc)
|
|
|
|
|
WRITE_ONCE(rnp->cbovldmask, rnp->cbovldmask | rdp->grpmask);
|
|
|
|
|
else
|
|
|
|
|
WRITE_ONCE(rnp->cbovldmask, rnp->cbovldmask & ~rdp->grpmask);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Check and if necessary update the leaf rcu_node structure's
|
|
|
|
|
* ->cbovldmask bit corresponding to the current CPU based on that CPU's
|
|
|
|
|
* number of queued RCU callbacks. No locks need be held, but the
|
|
|
|
|
* caller must have disabled interrupts.
|
|
|
|
|
*
|
|
|
|
|
* Note that this function ignores the possibility that there are a lot
|
|
|
|
|
* of callbacks all of which have already seen the end of their respective
|
|
|
|
|
* grace periods. This omission is due to the need for no-CBs CPUs to
|
|
|
|
|
* be holding ->nocb_lock to do this check, which is too heavy for a
|
|
|
|
|
* common-case operation.
|
2012-08-20 12:35:53 +08:00
|
|
|
|
*/
|
2019-10-31 02:56:10 +08:00
|
|
|
|
static void check_cb_ovld(struct rcu_data *rdp)
|
|
|
|
|
{
|
|
|
|
|
struct rcu_node *const rnp = rdp->mynode;
|
|
|
|
|
|
|
|
|
|
if (qovld_calc <= 0 ||
|
|
|
|
|
((rcu_segcblist_n_cbs(&rdp->cblist) >= qovld_calc) ==
|
|
|
|
|
!!(READ_ONCE(rnp->cbovldmask) & rdp->grpmask)))
|
|
|
|
|
return; // Early boot wildcard value or already set correctly.
|
|
|
|
|
raw_spin_lock_rcu_node(rnp);
|
|
|
|
|
check_cb_ovld_locked(rdp, rnp);
|
|
|
|
|
raw_spin_unlock_rcu_node(rnp);
|
|
|
|
|
}
|
|
|
|
|
|
2022-10-17 00:22:54 +08:00
|
|
|
|
static void
|
2023-01-12 08:52:23 +08:00
|
|
|
|
__call_rcu_common(struct rcu_head *head, rcu_callback_t func, bool lazy_in)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
{
|
2020-12-09 05:45:49 +08:00
|
|
|
|
static atomic_t doublefrees;
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
unsigned long flags;
|
2023-01-12 08:52:23 +08:00
|
|
|
|
bool lazy;
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
struct rcu_data *rdp;
|
2019-05-16 00:56:40 +08:00
|
|
|
|
bool was_alldone;
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
2016-08-23 21:51:47 +08:00
|
|
|
|
/* Misaligned rcu_head! */
|
|
|
|
|
WARN_ON_ONCE((unsigned long)head & (sizeof(void *) - 1));
|
|
|
|
|
|
2013-04-24 04:20:57 +08:00
|
|
|
|
if (debug_rcu_head_queue(head)) {
|
2017-05-04 02:38:55 +08:00
|
|
|
|
/*
|
|
|
|
|
* Probable double call_rcu(), so leak the callback.
|
|
|
|
|
* Use rcu:rcu_callback trace event to find the previous
|
2021-12-19 01:30:33 +08:00
|
|
|
|
* time callback was passed to call_rcu().
|
2017-05-04 02:38:55 +08:00
|
|
|
|
*/
|
2020-12-09 05:45:49 +08:00
|
|
|
|
if (atomic_inc_return(&doublefrees) < 4) {
|
|
|
|
|
pr_err("%s(): Double-freed CB %p->%pS()!!! ", __func__, head, head->func);
|
|
|
|
|
mem_dump_obj(head);
|
|
|
|
|
}
|
2015-03-04 06:57:58 +08:00
|
|
|
|
WRITE_ONCE(head->func, rcu_leak_callback);
|
2013-04-24 04:20:57 +08:00
|
|
|
|
return;
|
|
|
|
|
}
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
head->func = func;
|
|
|
|
|
head->next = NULL;
|
2021-11-16 07:23:02 +08:00
|
|
|
|
kasan_record_aux_stack_noalloc(head);
|
2021-12-26 08:52:04 +08:00
|
|
|
|
local_irq_save(flags);
|
2018-07-04 06:37:16 +08:00
|
|
|
|
rdp = this_cpu_ptr(&rcu_data);
|
2023-01-12 08:52:23 +08:00
|
|
|
|
lazy = lazy_in && !rcu_async_should_hurry();
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
|
|
|
|
/* Add the callback to our list. */
|
2019-05-16 00:56:40 +08:00
|
|
|
|
if (unlikely(!rcu_segcblist_is_enabled(&rdp->cblist))) {
|
|
|
|
|
// This can trigger due to call_rcu() from offline CPU:
|
|
|
|
|
WARN_ON_ONCE(rcu_scheduler_active != RCU_SCHEDULER_INACTIVE);
|
2015-01-20 12:39:20 +08:00
|
|
|
|
WARN_ON_ONCE(!rcu_is_watching());
|
2019-05-16 00:56:40 +08:00
|
|
|
|
// Very early boot, before rcu_init(). Initialize if needed
|
|
|
|
|
// and then drop through to queue the callback.
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
if (rcu_segcblist_empty(&rdp->cblist))
|
|
|
|
|
rcu_segcblist_init(&rdp->cblist);
|
2012-08-04 04:16:15 +08:00
|
|
|
|
}
|
2019-08-31 00:36:32 +08:00
|
|
|
|
|
2019-10-31 02:56:10 +08:00
|
|
|
|
check_cb_ovld(rdp);
|
2022-10-17 00:22:54 +08:00
|
|
|
|
if (rcu_nocb_try_bypass(rdp, head, &was_alldone, flags, lazy))
|
rcu/nocb: Add bypass callback queueing
Use of the rcu_data structure's segmented ->cblist for no-CBs CPUs
takes advantage of unrelated grace periods, thus reducing the memory
footprint in the face of floods of call_rcu() invocations. However,
the ->cblist field is a more-complex rcu_segcblist structure which must
be protected via locking. Even though there are only three entities
which can acquire this lock (the CPU invoking call_rcu(), the no-CBs
grace-period kthread, and the no-CBs callbacks kthread), the contention
on this lock is excessive under heavy stress.
This commit therefore greatly reduces contention by provisioning
an rcu_cblist structure field named ->nocb_bypass within the
rcu_data structure. Each no-CBs CPU is permitted only a limited
number of enqueues onto the ->cblist per jiffy, controlled by a new
nocb_nobypass_lim_per_jiffy kernel boot parameter that defaults to
about 16 enqueues per millisecond (16 * 1000 / HZ). When that limit is
exceeded, the CPU instead enqueues onto the new ->nocb_bypass.
The ->nocb_bypass is flushed into the ->cblist every jiffy or when
the number of callbacks on ->nocb_bypass exceeds qhimark, whichever
happens first. During call_rcu() floods, this flushing is carried out
by the CPU during the course of its call_rcu() invocations. However,
a CPU could simply stop invoking call_rcu() at any time. The no-CBs
grace-period kthread therefore carries out less-aggressive flushing
(every few jiffies or when the number of callbacks on ->nocb_bypass
exceeds (2 * qhimark), whichever comes first). This means that the
no-CBs grace-period kthread cannot be permitted to do unbounded waits
while there are callbacks on ->nocb_bypass. A ->nocb_bypass_timer is
used to provide the needed wakeups.
[ paulmck: Apply Coverity feedback reported by Colin Ian King. ]
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
2019-07-03 07:03:33 +08:00
|
|
|
|
return; // Enqueued onto ->nocb_bypass, so just leave.
|
2020-02-11 23:29:02 +08:00
|
|
|
|
// If no-CBs CPU gets here, rcu_nocb_try_bypass() acquired ->nocb_lock.
|
2019-08-31 00:36:32 +08:00
|
|
|
|
rcu_segcblist_enqueue(&rdp->cblist, head);
|
2020-05-26 05:47:55 +08:00
|
|
|
|
if (__is_kvfree_rcu_offset((unsigned long)func))
|
|
|
|
|
trace_rcu_kvfree_callback(rcu_state.name, head,
|
2018-07-06 06:54:02 +08:00
|
|
|
|
(unsigned long)func,
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
rcu_segcblist_n_cbs(&rdp->cblist));
|
rcu: Add grace-period, quiescent-state, and call_rcu trace events
Add trace events to record grace-period start and end, quiescent states,
CPUs noticing grace-period start and end, grace-period initialization,
call_rcu() invocation, tasks blocking in RCU read-side critical sections,
tasks exiting those same critical sections, force_quiescent_state()
detection of dyntick-idle and offline CPUs, CPUs entering and leaving
dyntick-idle mode (except from NMIs), CPUs coming online and going
offline, and CPUs being kicked for staying in dyntick-idle mode for too
long (as in many weeks, even on 32-bit systems).
Signed-off-by: Paul E. McKenney <paul.mckenney@linaro.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
rcu: Add the rcu flavor to callback trace events
The earlier trace events for registering RCU callbacks and for invoking
them did not include the RCU flavor (rcu_bh, rcu_preempt, or rcu_sched).
This commit adds the RCU flavor to those trace events.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2011-06-25 21:36:56 +08:00
|
|
|
|
else
|
2018-07-06 06:54:02 +08:00
|
|
|
|
trace_rcu_callback(rcu_state.name, head,
|
srcu: Abstract multi-tail callback list handling
RCU has only one multi-tail callback list, which is implemented via
the nxtlist, nxttail, nxtcompleted, qlen_lazy, and qlen fields in the
rcu_data structure, and whose operations are open-code throughout the
Tree RCU implementation. This has been more or less OK in the past,
but upcoming callback-list optimizations in SRCU could really use
a multi-tail callback list there as well.
This commit therefore abstracts the multi-tail callback list handling
into a new kernel/rcu/rcu_segcblist.h file, and uses this new API.
The simple head-and-tail pointer callback list is also abstracted and
applied everywhere except for the NOCB callback-offload lists. (Yes,
the plan is to apply them there as well, but this commit is already
bigger than would be good.)
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-02-09 04:36:42 +08:00
|
|
|
|
rcu_segcblist_n_cbs(&rdp->cblist));
|
rcu: Add grace-period, quiescent-state, and call_rcu trace events
Add trace events to record grace-period start and end, quiescent states,
CPUs noticing grace-period start and end, grace-period initialization,
call_rcu() invocation, tasks blocking in RCU read-side critical sections,
tasks exiting those same critical sections, force_quiescent_state()
detection of dyntick-idle and offline CPUs, CPUs entering and leaving
dyntick-idle mode (except from NMIs), CPUs coming online and going
offline, and CPUs being kicked for staying in dyntick-idle mode for too
long (as in many weeks, even on 32-bit systems).
Signed-off-by: Paul E. McKenney <paul.mckenney@linaro.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
rcu: Add the rcu flavor to callback trace events
The earlier trace events for registering RCU callbacks and for invoking
them did not include the RCU flavor (rcu_bh, rcu_preempt, or rcu_sched).
This commit adds the RCU flavor to those trace events.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2011-06-25 21:36:56 +08:00
|
|
|
|
|
2020-11-15 03:31:32 +08:00
|
|
|
|
trace_rcu_segcb_stats(&rdp->cblist, TPS("SegCBQueued"));
|
|
|
|
|
|
2012-05-30 18:21:48 +08:00
|
|
|
|
/* Go handle any RCU core processing required. */
|
2020-11-12 08:51:21 +08:00
|
|
|
|
if (unlikely(rcu_rdp_is_offloaded(rdp))) {
|
2019-05-16 00:56:40 +08:00
|
|
|
|
__call_rcu_nocb_wake(rdp, was_alldone, flags); /* unlocks */
|
|
|
|
|
} else {
|
|
|
|
|
__call_rcu_core(rdp, head, flags);
|
|
|
|
|
local_irq_restore(flags);
|
|
|
|
|
}
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
|
|
|
|
|
2022-10-17 00:22:54 +08:00
|
|
|
|
#ifdef CONFIG_RCU_LAZY
|
|
|
|
|
/**
|
|
|
|
|
* call_rcu_hurry() - Queue RCU callback for invocation after grace period, and
|
|
|
|
|
* flush all lazy callbacks (including the new one) to the main ->cblist while
|
|
|
|
|
* doing so.
|
|
|
|
|
*
|
|
|
|
|
* @head: structure to be used for queueing the RCU updates.
|
|
|
|
|
* @func: actual callback function to be invoked after the grace period
|
|
|
|
|
*
|
|
|
|
|
* The callback function will be invoked some time after a full grace
|
|
|
|
|
* period elapses, in other words after all pre-existing RCU read-side
|
|
|
|
|
* critical sections have completed.
|
|
|
|
|
*
|
|
|
|
|
* Use this API instead of call_rcu() if you don't want the callback to be
|
|
|
|
|
* invoked after very long periods of time, which can happen on systems without
|
|
|
|
|
* memory pressure and on systems which are lightly loaded or mostly idle.
|
|
|
|
|
* This function will cause callbacks to be invoked sooner than later at the
|
|
|
|
|
* expense of extra power. Other than that, this function is identical to, and
|
|
|
|
|
* reuses call_rcu()'s logic. Refer to call_rcu() for more details about memory
|
|
|
|
|
* ordering and other functionality.
|
|
|
|
|
*/
|
|
|
|
|
void call_rcu_hurry(struct rcu_head *head, rcu_callback_t func)
|
|
|
|
|
{
|
|
|
|
|
return __call_rcu_common(head, func, false);
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(call_rcu_hurry);
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* call_rcu() - Queue an RCU callback for invocation after a grace period.
|
|
|
|
|
* By default the callbacks are 'lazy' and are kept hidden from the main
|
|
|
|
|
* ->cblist to prevent starting of grace periods too soon.
|
|
|
|
|
* If you desire grace periods to start very soon, use call_rcu_hurry().
|
|
|
|
|
*
|
|
|
|
|
* @head: structure to be used for queueing the RCU updates.
|
|
|
|
|
* @func: actual callback function to be invoked after the grace period
|
|
|
|
|
*
|
|
|
|
|
* The callback function will be invoked some time after a full grace
|
|
|
|
|
* period elapses, in other words after all pre-existing RCU read-side
|
|
|
|
|
* critical sections have completed. However, the callback function
|
|
|
|
|
* might well execute concurrently with RCU read-side critical sections
|
|
|
|
|
* that started after call_rcu() was invoked.
|
|
|
|
|
*
|
|
|
|
|
* RCU read-side critical sections are delimited by rcu_read_lock()
|
|
|
|
|
* and rcu_read_unlock(), and may be nested. In addition, but only in
|
|
|
|
|
* v5.0 and later, regions of code across which interrupts, preemption,
|
|
|
|
|
* or softirqs have been disabled also serve as RCU read-side critical
|
|
|
|
|
* sections. This includes hardware interrupt handlers, softirq handlers,
|
|
|
|
|
* and NMI handlers.
|
|
|
|
|
*
|
|
|
|
|
* Note that all CPUs must agree that the grace period extended beyond
|
|
|
|
|
* all pre-existing RCU read-side critical section. On systems with more
|
|
|
|
|
* than one CPU, this means that when "func()" is invoked, each CPU is
|
|
|
|
|
* guaranteed to have executed a full memory barrier since the end of its
|
|
|
|
|
* last RCU read-side critical section whose beginning preceded the call
|
|
|
|
|
* to call_rcu(). It also means that each CPU executing an RCU read-side
|
|
|
|
|
* critical section that continues beyond the start of "func()" must have
|
|
|
|
|
* executed a memory barrier after the call_rcu() but before the beginning
|
|
|
|
|
* of that RCU read-side critical section. Note that these guarantees
|
|
|
|
|
* include CPUs that are offline, idle, or executing in user mode, as
|
|
|
|
|
* well as CPUs that are executing in the kernel.
|
|
|
|
|
*
|
|
|
|
|
* Furthermore, if CPU A invoked call_rcu() and CPU B invoked the
|
|
|
|
|
* resulting RCU callback function "func()", then both CPU A and CPU B are
|
|
|
|
|
* guaranteed to execute a full memory barrier during the time interval
|
|
|
|
|
* between the call to call_rcu() and the invocation of "func()" -- even
|
|
|
|
|
* if CPU A and CPU B are the same CPU (but again only if the system has
|
|
|
|
|
* more than one CPU).
|
|
|
|
|
*
|
|
|
|
|
* Implementation of these memory-ordering guarantees is described here:
|
|
|
|
|
* Documentation/RCU/Design/Memory-Ordering/Tree-RCU-Memory-Ordering.rst.
|
|
|
|
|
*/
|
|
|
|
|
void call_rcu(struct rcu_head *head, rcu_callback_t func)
|
|
|
|
|
{
|
|
|
|
|
return __call_rcu_common(head, func, IS_ENABLED(CONFIG_RCU_LAZY));
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(call_rcu);
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
|
|
|
|
|
/* Maximum number of jiffies to wait before draining a batch. */
|
2022-07-01 00:33:35 +08:00
|
|
|
|
#define KFREE_DRAIN_JIFFIES (5 * HZ)
|
2019-09-20 05:58:26 +08:00
|
|
|
|
#define KFREE_N_BATCHES 2
|
2020-05-26 05:47:53 +08:00
|
|
|
|
#define FREE_N_CHANNELS 2
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
|
|
|
|
|
/**
|
2020-05-26 05:47:53 +08:00
|
|
|
|
* struct kvfree_rcu_bulk_data - single block to store kvfree_rcu() pointers
|
2022-11-29 23:58:19 +08:00
|
|
|
|
* @list: List node. All blocks are linked between each other
|
rcu/kvfree: Use a polled API to speedup a reclaim process
Currently all objects placed into a batch wait for a full grace period
to elapse after that batch is ready to send to RCU. However, this
can unnecessarily delay freeing of the first objects that were added
to the batch. After all, several RCU grace periods might have elapsed
since those objects were added, and if so, there is no point in further
deferring their freeing.
This commit therefore adds per-page grace-period snapshots which are
obtained from get_state_synchronize_rcu(). When the batch is ready
to be passed to call_rcu(), each page's snapshot is checked by passing
it to poll_state_synchronize_rcu(). If a given page's RCU grace period
has already elapsed, its objects are freed immediately by kvfree_rcu_bulk().
Otherwise, these objects are freed after a call to synchronize_rcu().
This approach requires that the pages be traversed in reverse order,
that is, the oldest ones first.
Test example:
kvm.sh --memory 10G --torture rcuscale --allcpus --duration 1 \
--kconfig CONFIG_NR_CPUS=64 \
--kconfig CONFIG_RCU_NOCB_CPU=y \
--kconfig CONFIG_RCU_NOCB_CPU_DEFAULT_ALL=y \
--kconfig CONFIG_RCU_LAZY=n \
--bootargs "rcuscale.kfree_rcu_test=1 rcuscale.kfree_nthreads=16 \
rcuscale.holdoff=20 rcuscale.kfree_loops=10000 \
torture.disable_onoff_at_boot" --trust-make
Before this commit:
Total time taken by all kfree'ers: 8535693700 ns, loops: 10000, batches: 1188, memory footprint: 2248MB
Total time taken by all kfree'ers: 8466933582 ns, loops: 10000, batches: 1157, memory footprint: 2820MB
Total time taken by all kfree'ers: 5375602446 ns, loops: 10000, batches: 1130, memory footprint: 6502MB
Total time taken by all kfree'ers: 7523283832 ns, loops: 10000, batches: 1006, memory footprint: 3343MB
Total time taken by all kfree'ers: 6459171956 ns, loops: 10000, batches: 1150, memory footprint: 6549MB
After this commit:
Total time taken by all kfree'ers: 8560060176 ns, loops: 10000, batches: 1787, memory footprint: 61MB
Total time taken by all kfree'ers: 8573885501 ns, loops: 10000, batches: 1777, memory footprint: 93MB
Total time taken by all kfree'ers: 8320000202 ns, loops: 10000, batches: 1727, memory footprint: 66MB
Total time taken by all kfree'ers: 8552718794 ns, loops: 10000, batches: 1790, memory footprint: 75MB
Total time taken by all kfree'ers: 8601368792 ns, loops: 10000, batches: 1724, memory footprint: 62MB
The reduction in memory footprint is well in excess of an order of
magnitude.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2022-11-29 23:58:22 +08:00
|
|
|
|
* @gp_snap: Snapshot of RCU state for objects placed to this bulk
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
* @nr_records: Number of active pointers in the array
|
2020-05-26 05:47:53 +08:00
|
|
|
|
* @records: Array of the kvfree_rcu() pointers
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
*/
|
2020-05-26 05:47:53 +08:00
|
|
|
|
struct kvfree_rcu_bulk_data {
|
2022-11-29 23:58:19 +08:00
|
|
|
|
struct list_head list;
|
2023-04-04 07:49:14 +08:00
|
|
|
|
struct rcu_gp_oldstate gp_snap;
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
unsigned long nr_records;
|
2020-05-26 05:47:49 +08:00
|
|
|
|
void *records[];
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
};
|
|
|
|
|
|
2020-05-26 05:47:49 +08:00
|
|
|
|
/*
|
|
|
|
|
* This macro defines how many entries the "records" array
|
|
|
|
|
* will contain. It is based on the fact that the size of
|
2020-05-26 05:47:53 +08:00
|
|
|
|
* kvfree_rcu_bulk_data structure becomes exactly one page.
|
2020-05-26 05:47:49 +08:00
|
|
|
|
*/
|
2020-05-26 05:47:53 +08:00
|
|
|
|
#define KVFREE_BULK_MAX_ENTR \
|
|
|
|
|
((PAGE_SIZE - sizeof(struct kvfree_rcu_bulk_data)) / sizeof(void *))
|
2020-05-26 05:47:49 +08:00
|
|
|
|
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
/**
|
2019-09-20 05:58:26 +08:00
|
|
|
|
* struct kfree_rcu_cpu_work - single batch of kfree_rcu() requests
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
* @rcu_work: Let queue_rcu_work() invoke workqueue handler after grace period
|
2019-09-20 05:58:26 +08:00
|
|
|
|
* @head_free: List of kfree_rcu() objects waiting for a grace period
|
2023-04-04 22:13:00 +08:00
|
|
|
|
* @head_free_gp_snap: Grace-period snapshot to check for attempted premature frees.
|
2022-11-29 23:58:19 +08:00
|
|
|
|
* @bulk_head_free: Bulk-List of kvfree_rcu() objects waiting for a grace period
|
2019-09-20 05:58:26 +08:00
|
|
|
|
* @krcp: Pointer to @kfree_rcu_cpu structure
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
struct kfree_rcu_cpu_work {
|
|
|
|
|
struct rcu_work rcu_work;
|
|
|
|
|
struct rcu_head *head_free;
|
2023-04-04 22:13:00 +08:00
|
|
|
|
struct rcu_gp_oldstate head_free_gp_snap;
|
2022-11-29 23:58:19 +08:00
|
|
|
|
struct list_head bulk_head_free[FREE_N_CHANNELS];
|
2019-09-20 05:58:26 +08:00
|
|
|
|
struct kfree_rcu_cpu *krcp;
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* struct kfree_rcu_cpu - batch up kfree_rcu() requests for RCU grace period
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
* @head: List of kfree_rcu() objects not yet waiting for a grace period
|
2022-12-14 20:06:30 +08:00
|
|
|
|
* @head_gp_snap: Snapshot of RCU state for objects placed to "@head"
|
2022-11-29 23:58:19 +08:00
|
|
|
|
* @bulk_head: Bulk-List of kvfree_rcu() objects not yet waiting for a grace period
|
2019-09-20 05:58:26 +08:00
|
|
|
|
* @krw_arr: Array of batches of kfree_rcu() objects waiting for a grace period
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
* @lock: Synchronize access to this structure
|
|
|
|
|
* @monitor_work: Promote @head to @head_free after KFREE_DRAIN_JIFFIES
|
2020-05-26 05:47:51 +08:00
|
|
|
|
* @initialized: The @rcu_work fields have been initialized
|
2022-12-14 20:06:29 +08:00
|
|
|
|
* @head_count: Number of objects in rcu_head singular list
|
|
|
|
|
* @bulk_count: Number of objects in bulk-list
|
2020-09-10 16:20:34 +08:00
|
|
|
|
* @bkvcache:
|
|
|
|
|
* A simple cache list that contains objects for reuse purpose.
|
|
|
|
|
* In order to save some per-cpu space the list is singular.
|
|
|
|
|
* Even though it is lockless an access has to be protected by the
|
|
|
|
|
* per-cpu lock.
|
2020-10-30 00:50:04 +08:00
|
|
|
|
* @page_cache_work: A work to refill the cache when it is empty
|
2021-04-16 01:19:56 +08:00
|
|
|
|
* @backoff_page_cache_fill: Delay cache refills
|
2020-10-30 00:50:04 +08:00
|
|
|
|
* @work_in_progress: Indicates that page_cache_work is running
|
|
|
|
|
* @hrtimer: A hrtimer for scheduling a page_cache_work
|
2020-09-10 16:20:34 +08:00
|
|
|
|
* @nr_bkv_objs: number of allocated objects at @bkvcache.
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
*
|
|
|
|
|
* This is a per-CPU structure. The reason that it is not included in
|
|
|
|
|
* the rcu_data structure is to permit this code to be extracted from
|
|
|
|
|
* the RCU files. Such extraction could allow further optimization of
|
|
|
|
|
* the interactions with the slab allocators.
|
|
|
|
|
*/
|
|
|
|
|
struct kfree_rcu_cpu {
|
2022-12-14 20:06:29 +08:00
|
|
|
|
// Objects queued on a linked list
|
|
|
|
|
// through their rcu_head structures.
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
struct rcu_head *head;
|
2022-12-14 20:06:30 +08:00
|
|
|
|
unsigned long head_gp_snap;
|
2022-12-14 20:06:29 +08:00
|
|
|
|
atomic_t head_count;
|
|
|
|
|
|
|
|
|
|
// Objects queued on a bulk-list.
|
2022-11-29 23:58:19 +08:00
|
|
|
|
struct list_head bulk_head[FREE_N_CHANNELS];
|
2022-12-14 20:06:29 +08:00
|
|
|
|
atomic_t bulk_count[FREE_N_CHANNELS];
|
|
|
|
|
|
2019-09-20 05:58:26 +08:00
|
|
|
|
struct kfree_rcu_cpu_work krw_arr[KFREE_N_BATCHES];
|
2020-05-26 05:47:45 +08:00
|
|
|
|
raw_spinlock_t lock;
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
struct delayed_work monitor_work;
|
|
|
|
|
bool initialized;
|
2020-10-30 00:50:04 +08:00
|
|
|
|
|
2021-04-16 01:19:56 +08:00
|
|
|
|
struct delayed_work page_cache_work;
|
|
|
|
|
atomic_t backoff_page_cache_fill;
|
2020-10-30 00:50:04 +08:00
|
|
|
|
atomic_t work_in_progress;
|
|
|
|
|
struct hrtimer hrtimer;
|
|
|
|
|
|
2020-05-26 05:47:52 +08:00
|
|
|
|
struct llist_head bkvcache;
|
|
|
|
|
int nr_bkv_objs;
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
};
|
|
|
|
|
|
2020-05-26 05:47:51 +08:00
|
|
|
|
static DEFINE_PER_CPU(struct kfree_rcu_cpu, krc) = {
|
|
|
|
|
.lock = __RAW_SPIN_LOCK_UNLOCKED(krc.lock),
|
|
|
|
|
};
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
static __always_inline void
|
2020-05-26 05:47:53 +08:00
|
|
|
|
debug_rcu_bhead_unqueue(struct kvfree_rcu_bulk_data *bhead)
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
{
|
|
|
|
|
#ifdef CONFIG_DEBUG_OBJECTS_RCU_HEAD
|
2020-05-26 05:47:48 +08:00
|
|
|
|
int i;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < bhead->nr_records; i++)
|
|
|
|
|
debug_rcu_head_unqueue((struct rcu_head *)(bhead->records[i]));
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
#endif
|
|
|
|
|
}
|
|
|
|
|
|
2020-05-26 05:47:50 +08:00
|
|
|
|
static inline struct kfree_rcu_cpu *
|
|
|
|
|
krc_this_cpu_lock(unsigned long *flags)
|
|
|
|
|
{
|
|
|
|
|
struct kfree_rcu_cpu *krcp;
|
|
|
|
|
|
|
|
|
|
local_irq_save(*flags); // For safely calling this_cpu_ptr().
|
|
|
|
|
krcp = this_cpu_ptr(&krc);
|
2020-05-26 05:47:51 +08:00
|
|
|
|
raw_spin_lock(&krcp->lock);
|
2020-05-26 05:47:50 +08:00
|
|
|
|
|
|
|
|
|
return krcp;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline void
|
|
|
|
|
krc_this_cpu_unlock(struct kfree_rcu_cpu *krcp, unsigned long flags)
|
|
|
|
|
{
|
2021-01-21 05:38:08 +08:00
|
|
|
|
raw_spin_unlock_irqrestore(&krcp->lock, flags);
|
2020-05-26 05:47:50 +08:00
|
|
|
|
}
|
|
|
|
|
|
2020-05-26 05:47:53 +08:00
|
|
|
|
static inline struct kvfree_rcu_bulk_data *
|
2020-05-26 05:47:52 +08:00
|
|
|
|
get_cached_bnode(struct kfree_rcu_cpu *krcp)
|
|
|
|
|
{
|
|
|
|
|
if (!krcp->nr_bkv_objs)
|
|
|
|
|
return NULL;
|
|
|
|
|
|
2021-04-16 01:19:57 +08:00
|
|
|
|
WRITE_ONCE(krcp->nr_bkv_objs, krcp->nr_bkv_objs - 1);
|
2020-05-26 05:47:53 +08:00
|
|
|
|
return (struct kvfree_rcu_bulk_data *)
|
2020-05-26 05:47:52 +08:00
|
|
|
|
llist_del_first(&krcp->bkvcache);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline bool
|
|
|
|
|
put_cached_bnode(struct kfree_rcu_cpu *krcp,
|
2020-05-26 05:47:53 +08:00
|
|
|
|
struct kvfree_rcu_bulk_data *bnode)
|
2020-05-26 05:47:52 +08:00
|
|
|
|
{
|
|
|
|
|
// Check the limit.
|
|
|
|
|
if (krcp->nr_bkv_objs >= rcu_min_cached_objs)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
llist_add((struct llist_node *) bnode, &krcp->bkvcache);
|
2021-04-16 01:19:57 +08:00
|
|
|
|
WRITE_ONCE(krcp->nr_bkv_objs, krcp->nr_bkv_objs + 1);
|
2020-05-26 05:47:52 +08:00
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2021-04-16 01:19:56 +08:00
|
|
|
|
static int
|
|
|
|
|
drain_page_cache(struct kfree_rcu_cpu *krcp)
|
|
|
|
|
{
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
struct llist_node *page_list, *pos, *n;
|
|
|
|
|
int freed = 0;
|
2020-05-26 05:47:52 +08:00
|
|
|
|
|
2023-04-18 20:27:02 +08:00
|
|
|
|
if (!rcu_min_cached_objs)
|
|
|
|
|
return 0;
|
|
|
|
|
|
2021-04-16 01:19:56 +08:00
|
|
|
|
raw_spin_lock_irqsave(&krcp->lock, flags);
|
|
|
|
|
page_list = llist_del_all(&krcp->bkvcache);
|
2021-04-16 01:19:57 +08:00
|
|
|
|
WRITE_ONCE(krcp->nr_bkv_objs, 0);
|
2021-04-16 01:19:56 +08:00
|
|
|
|
raw_spin_unlock_irqrestore(&krcp->lock, flags);
|
2020-05-26 05:47:52 +08:00
|
|
|
|
|
2021-04-16 01:19:56 +08:00
|
|
|
|
llist_for_each_safe(pos, n, page_list) {
|
|
|
|
|
free_page((unsigned long)pos);
|
|
|
|
|
freed++;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return freed;
|
2020-05-26 05:47:52 +08:00
|
|
|
|
}
|
|
|
|
|
|
2022-11-29 23:58:20 +08:00
|
|
|
|
static void
|
|
|
|
|
kvfree_rcu_bulk(struct kfree_rcu_cpu *krcp,
|
|
|
|
|
struct kvfree_rcu_bulk_data *bnode, int idx)
|
|
|
|
|
{
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
int i;
|
|
|
|
|
|
2023-04-04 07:49:14 +08:00
|
|
|
|
if (!WARN_ON_ONCE(!poll_state_synchronize_rcu_full(&bnode->gp_snap))) {
|
2023-04-05 10:13:59 +08:00
|
|
|
|
debug_rcu_bhead_unqueue(bnode);
|
2023-04-04 07:49:14 +08:00
|
|
|
|
rcu_lock_acquire(&rcu_callback_map);
|
|
|
|
|
if (idx == 0) { // kmalloc() / kfree().
|
|
|
|
|
trace_rcu_invoke_kfree_bulk_callback(
|
|
|
|
|
rcu_state.name, bnode->nr_records,
|
|
|
|
|
bnode->records);
|
|
|
|
|
|
|
|
|
|
kfree_bulk(bnode->nr_records, bnode->records);
|
|
|
|
|
} else { // vmalloc() / vfree().
|
|
|
|
|
for (i = 0; i < bnode->nr_records; i++) {
|
|
|
|
|
trace_rcu_invoke_kvfree_callback(
|
|
|
|
|
rcu_state.name, bnode->records[i], 0);
|
|
|
|
|
|
|
|
|
|
vfree(bnode->records[i]);
|
|
|
|
|
}
|
2022-11-29 23:58:20 +08:00
|
|
|
|
}
|
2023-04-04 07:49:14 +08:00
|
|
|
|
rcu_lock_release(&rcu_callback_map);
|
2022-11-29 23:58:20 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
raw_spin_lock_irqsave(&krcp->lock, flags);
|
|
|
|
|
if (put_cached_bnode(krcp, bnode))
|
|
|
|
|
bnode = NULL;
|
|
|
|
|
raw_spin_unlock_irqrestore(&krcp->lock, flags);
|
|
|
|
|
|
|
|
|
|
if (bnode)
|
|
|
|
|
free_page((unsigned long) bnode);
|
|
|
|
|
|
|
|
|
|
cond_resched_tasks_rcu_qs();
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void
|
|
|
|
|
kvfree_rcu_list(struct rcu_head *head)
|
|
|
|
|
{
|
|
|
|
|
struct rcu_head *next;
|
|
|
|
|
|
|
|
|
|
for (; head; head = next) {
|
|
|
|
|
void *ptr = (void *) head->func;
|
|
|
|
|
unsigned long offset = (void *) head - ptr;
|
|
|
|
|
|
|
|
|
|
next = head->next;
|
|
|
|
|
debug_rcu_head_unqueue((struct rcu_head *)ptr);
|
|
|
|
|
rcu_lock_acquire(&rcu_callback_map);
|
|
|
|
|
trace_rcu_invoke_kvfree_callback(rcu_state.name, head, offset);
|
|
|
|
|
|
|
|
|
|
if (!WARN_ON_ONCE(!__is_kvfree_rcu_offset(offset)))
|
|
|
|
|
kvfree(ptr);
|
|
|
|
|
|
|
|
|
|
rcu_lock_release(&rcu_callback_map);
|
|
|
|
|
cond_resched_tasks_rcu_qs();
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2014-03-19 02:48:48 +08:00
|
|
|
|
/*
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
* This function is invoked in workqueue context after a grace period.
|
2022-11-29 23:58:19 +08:00
|
|
|
|
* It frees all the objects queued on ->bulk_head_free or ->head_free.
|
2014-03-19 02:48:48 +08:00
|
|
|
|
*/
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
static void kfree_rcu_work(struct work_struct *work)
|
|
|
|
|
{
|
|
|
|
|
unsigned long flags;
|
2022-11-29 23:58:19 +08:00
|
|
|
|
struct kvfree_rcu_bulk_data *bnode, *n;
|
|
|
|
|
struct list_head bulk_head[FREE_N_CHANNELS];
|
2022-11-29 23:58:20 +08:00
|
|
|
|
struct rcu_head *head;
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
struct kfree_rcu_cpu *krcp;
|
2019-09-20 05:58:26 +08:00
|
|
|
|
struct kfree_rcu_cpu_work *krwp;
|
2023-04-04 22:13:00 +08:00
|
|
|
|
struct rcu_gp_oldstate head_gp_snap;
|
2022-11-29 23:58:20 +08:00
|
|
|
|
int i;
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
|
2019-09-20 05:58:26 +08:00
|
|
|
|
krwp = container_of(to_rcu_work(work),
|
rcu/kvfree: Use a polled API to speedup a reclaim process
Currently all objects placed into a batch wait for a full grace period
to elapse after that batch is ready to send to RCU. However, this
can unnecessarily delay freeing of the first objects that were added
to the batch. After all, several RCU grace periods might have elapsed
since those objects were added, and if so, there is no point in further
deferring their freeing.
This commit therefore adds per-page grace-period snapshots which are
obtained from get_state_synchronize_rcu(). When the batch is ready
to be passed to call_rcu(), each page's snapshot is checked by passing
it to poll_state_synchronize_rcu(). If a given page's RCU grace period
has already elapsed, its objects are freed immediately by kvfree_rcu_bulk().
Otherwise, these objects are freed after a call to synchronize_rcu().
This approach requires that the pages be traversed in reverse order,
that is, the oldest ones first.
Test example:
kvm.sh --memory 10G --torture rcuscale --allcpus --duration 1 \
--kconfig CONFIG_NR_CPUS=64 \
--kconfig CONFIG_RCU_NOCB_CPU=y \
--kconfig CONFIG_RCU_NOCB_CPU_DEFAULT_ALL=y \
--kconfig CONFIG_RCU_LAZY=n \
--bootargs "rcuscale.kfree_rcu_test=1 rcuscale.kfree_nthreads=16 \
rcuscale.holdoff=20 rcuscale.kfree_loops=10000 \
torture.disable_onoff_at_boot" --trust-make
Before this commit:
Total time taken by all kfree'ers: 8535693700 ns, loops: 10000, batches: 1188, memory footprint: 2248MB
Total time taken by all kfree'ers: 8466933582 ns, loops: 10000, batches: 1157, memory footprint: 2820MB
Total time taken by all kfree'ers: 5375602446 ns, loops: 10000, batches: 1130, memory footprint: 6502MB
Total time taken by all kfree'ers: 7523283832 ns, loops: 10000, batches: 1006, memory footprint: 3343MB
Total time taken by all kfree'ers: 6459171956 ns, loops: 10000, batches: 1150, memory footprint: 6549MB
After this commit:
Total time taken by all kfree'ers: 8560060176 ns, loops: 10000, batches: 1787, memory footprint: 61MB
Total time taken by all kfree'ers: 8573885501 ns, loops: 10000, batches: 1777, memory footprint: 93MB
Total time taken by all kfree'ers: 8320000202 ns, loops: 10000, batches: 1727, memory footprint: 66MB
Total time taken by all kfree'ers: 8552718794 ns, loops: 10000, batches: 1790, memory footprint: 75MB
Total time taken by all kfree'ers: 8601368792 ns, loops: 10000, batches: 1724, memory footprint: 62MB
The reduction in memory footprint is well in excess of an order of
magnitude.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2022-11-29 23:58:22 +08:00
|
|
|
|
struct kfree_rcu_cpu_work, rcu_work);
|
2019-09-20 05:58:26 +08:00
|
|
|
|
krcp = krwp->krcp;
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
|
2020-05-26 05:47:45 +08:00
|
|
|
|
raw_spin_lock_irqsave(&krcp->lock, flags);
|
2020-05-26 05:47:53 +08:00
|
|
|
|
// Channels 1 and 2.
|
2022-11-29 23:58:19 +08:00
|
|
|
|
for (i = 0; i < FREE_N_CHANNELS; i++)
|
|
|
|
|
list_replace_init(&krwp->bulk_head_free[i], &bulk_head[i]);
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
|
2020-05-26 05:47:53 +08:00
|
|
|
|
// Channel 3.
|
2019-09-20 05:58:26 +08:00
|
|
|
|
head = krwp->head_free;
|
|
|
|
|
krwp->head_free = NULL;
|
2023-04-04 22:13:00 +08:00
|
|
|
|
head_gp_snap = krwp->head_free_gp_snap;
|
2020-05-26 05:47:45 +08:00
|
|
|
|
raw_spin_unlock_irqrestore(&krcp->lock, flags);
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
|
2021-03-31 04:47:42 +08:00
|
|
|
|
// Handle the first two channels.
|
2020-05-26 05:47:53 +08:00
|
|
|
|
for (i = 0; i < FREE_N_CHANNELS; i++) {
|
rcu/kvfree: Use a polled API to speedup a reclaim process
Currently all objects placed into a batch wait for a full grace period
to elapse after that batch is ready to send to RCU. However, this
can unnecessarily delay freeing of the first objects that were added
to the batch. After all, several RCU grace periods might have elapsed
since those objects were added, and if so, there is no point in further
deferring their freeing.
This commit therefore adds per-page grace-period snapshots which are
obtained from get_state_synchronize_rcu(). When the batch is ready
to be passed to call_rcu(), each page's snapshot is checked by passing
it to poll_state_synchronize_rcu(). If a given page's RCU grace period
has already elapsed, its objects are freed immediately by kvfree_rcu_bulk().
Otherwise, these objects are freed after a call to synchronize_rcu().
This approach requires that the pages be traversed in reverse order,
that is, the oldest ones first.
Test example:
kvm.sh --memory 10G --torture rcuscale --allcpus --duration 1 \
--kconfig CONFIG_NR_CPUS=64 \
--kconfig CONFIG_RCU_NOCB_CPU=y \
--kconfig CONFIG_RCU_NOCB_CPU_DEFAULT_ALL=y \
--kconfig CONFIG_RCU_LAZY=n \
--bootargs "rcuscale.kfree_rcu_test=1 rcuscale.kfree_nthreads=16 \
rcuscale.holdoff=20 rcuscale.kfree_loops=10000 \
torture.disable_onoff_at_boot" --trust-make
Before this commit:
Total time taken by all kfree'ers: 8535693700 ns, loops: 10000, batches: 1188, memory footprint: 2248MB
Total time taken by all kfree'ers: 8466933582 ns, loops: 10000, batches: 1157, memory footprint: 2820MB
Total time taken by all kfree'ers: 5375602446 ns, loops: 10000, batches: 1130, memory footprint: 6502MB
Total time taken by all kfree'ers: 7523283832 ns, loops: 10000, batches: 1006, memory footprint: 3343MB
Total time taken by all kfree'ers: 6459171956 ns, loops: 10000, batches: 1150, memory footprint: 6549MB
After this commit:
Total time taken by all kfree'ers: 8560060176 ns, loops: 10000, batches: 1787, memory footprint: 61MB
Total time taken by all kfree'ers: 8573885501 ns, loops: 10000, batches: 1777, memory footprint: 93MB
Total time taken by all kfree'ers: 8320000202 ns, loops: 10000, batches: 1727, memory footprint: 66MB
Total time taken by all kfree'ers: 8552718794 ns, loops: 10000, batches: 1790, memory footprint: 75MB
Total time taken by all kfree'ers: 8601368792 ns, loops: 10000, batches: 1724, memory footprint: 62MB
The reduction in memory footprint is well in excess of an order of
magnitude.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2022-11-29 23:58:22 +08:00
|
|
|
|
// Start from the tail page, so a GP is likely passed for it.
|
2022-11-29 23:58:20 +08:00
|
|
|
|
list_for_each_entry_safe(bnode, n, &bulk_head[i], list)
|
|
|
|
|
kvfree_rcu_bulk(krcp, bnode, i);
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
2021-04-28 21:44:22 +08:00
|
|
|
|
* This is used when the "bulk" path can not be used for the
|
|
|
|
|
* double-argument of kvfree_rcu(). This happens when the
|
|
|
|
|
* page-cache is empty, which means that objects are instead
|
|
|
|
|
* queued on a linked list through their rcu_head structures.
|
|
|
|
|
* This list is named "Channel 3".
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
*/
|
2023-04-04 22:13:00 +08:00
|
|
|
|
if (head && !WARN_ON_ONCE(!poll_state_synchronize_rcu_full(&head_gp_snap)))
|
|
|
|
|
kvfree_rcu_list(head);
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
}
|
|
|
|
|
|
2022-06-02 16:06:43 +08:00
|
|
|
|
static bool
|
|
|
|
|
need_offload_krc(struct kfree_rcu_cpu *krcp)
|
|
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < FREE_N_CHANNELS; i++)
|
2022-11-29 23:58:19 +08:00
|
|
|
|
if (!list_empty(&krcp->bulk_head[i]))
|
2022-06-02 16:06:43 +08:00
|
|
|
|
return true;
|
|
|
|
|
|
2022-12-02 21:18:37 +08:00
|
|
|
|
return !!READ_ONCE(krcp->head);
|
2022-06-02 16:06:43 +08:00
|
|
|
|
}
|
|
|
|
|
|
rcu/kvfree: Avoid freeing new kfree_rcu() memory after old grace period
Memory passed to kvfree_rcu() that is to be freed is tracked by a
per-CPU kfree_rcu_cpu structure, which in turn contains pointers
to kvfree_rcu_bulk_data structures that contain pointers to memory
that has not yet been handed to RCU, along with an kfree_rcu_cpu_work
structure that tracks the memory that has already been handed to RCU.
These structures track three categories of memory: (1) Memory for
kfree(), (2) Memory for kvfree(), and (3) Memory for both that arrived
during an OOM episode. The first two categories are tracked in a
cache-friendly manner involving a dynamically allocated page of pointers
(the aforementioned kvfree_rcu_bulk_data structures), while the third
uses a simple (but decidedly cache-unfriendly) linked list through the
rcu_head structures in each block of memory.
On a given CPU, these three categories are handled as a unit, with that
CPU's kfree_rcu_cpu_work structure having one pointer for each of the
three categories. Clearly, new memory for a given category cannot be
placed in the corresponding kfree_rcu_cpu_work structure until any old
memory has had its grace period elapse and thus has been removed. And
the kfree_rcu_monitor() function does in fact check for this.
Except that the kfree_rcu_monitor() function checks these pointers one
at a time. This means that if the previous kfree_rcu() memory passed
to RCU had only category 1 and the current one has only category 2, the
kfree_rcu_monitor() function will send that current category-2 memory
along immediately. This can result in memory being freed too soon,
that is, out from under unsuspecting RCU readers.
To see this, consider the following sequence of events, in which:
o Task A on CPU 0 calls rcu_read_lock(), then uses "from_cset",
then is preempted.
o CPU 1 calls kfree_rcu(cset, rcu_head) in order to free "from_cset"
after a later grace period. Except that "from_cset" is freed
right after the previous grace period ended, so that "from_cset"
is immediately freed. Task A resumes and references "from_cset"'s
member, after which nothing good happens.
In full detail:
CPU 0 CPU 1
---------------------- ----------------------
count_memcg_event_mm()
|rcu_read_lock() <---
|mem_cgroup_from_task()
|// css_set_ptr is the "from_cset" mentioned on CPU 1
|css_set_ptr = rcu_dereference((task)->cgroups)
|// Hard irq comes, current task is scheduled out.
cgroup_attach_task()
|cgroup_migrate()
|cgroup_migrate_execute()
|css_set_move_task(task, from_cset, to_cset, true)
|cgroup_move_task(task, to_cset)
|rcu_assign_pointer(.., to_cset)
|...
|cgroup_migrate_finish()
|put_css_set_locked(from_cset)
|from_cset->refcount return 0
|kfree_rcu(cset, rcu_head) // free from_cset after new gp
|add_ptr_to_bulk_krc_lock()
|schedule_delayed_work(&krcp->monitor_work, ..)
kfree_rcu_monitor()
|krcp->bulk_head[0]'s work attached to krwp->bulk_head_free[]
|queue_rcu_work(system_wq, &krwp->rcu_work)
|if rwork->rcu.work is not in WORK_STRUCT_PENDING_BIT state,
|call_rcu(&rwork->rcu, rcu_work_rcufn) <--- request new gp
// There is a perious call_rcu(.., rcu_work_rcufn)
// gp end, rcu_work_rcufn() is called.
rcu_work_rcufn()
|__queue_work(.., rwork->wq, &rwork->work);
|kfree_rcu_work()
|krwp->bulk_head_free[0] bulk is freed before new gp end!!!
|The "from_cset" is freed before new gp end.
// the task resumes some time later.
|css_set_ptr->subsys[(subsys_id) <--- Caused kernel crash, because css_set_ptr is freed.
This commit therefore causes kfree_rcu_monitor() to refrain from moving
kfree_rcu() memory to the kfree_rcu_cpu_work structure until the RCU
grace period has completed for all three categories.
v2: Use helper function instead of inserted code block at kfree_rcu_monitor().
Fixes: 34c881745549 ("rcu: Support kfree_bulk() interface in kfree_rcu()")
Fixes: 5f3c8d620447 ("rcu/tree: Maintain separate array for vmalloc ptrs")
Reported-by: Mukesh Ojha <quic_mojha@quicinc.com>
Signed-off-by: Ziwei Dai <ziwei.dai@unisoc.com>
Reviewed-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2023-03-31 20:42:09 +08:00
|
|
|
|
static bool
|
|
|
|
|
need_wait_for_krwp_work(struct kfree_rcu_cpu_work *krwp)
|
|
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < FREE_N_CHANNELS; i++)
|
|
|
|
|
if (!list_empty(&krwp->bulk_head_free[i]))
|
|
|
|
|
return true;
|
|
|
|
|
|
|
|
|
|
return !!krwp->head_free;
|
|
|
|
|
}
|
|
|
|
|
|
2022-12-14 20:06:29 +08:00
|
|
|
|
static int krc_count(struct kfree_rcu_cpu *krcp)
|
|
|
|
|
{
|
|
|
|
|
int sum = atomic_read(&krcp->head_count);
|
|
|
|
|
int i;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < FREE_N_CHANNELS; i++)
|
|
|
|
|
sum += atomic_read(&krcp->bulk_count[i]);
|
|
|
|
|
|
|
|
|
|
return sum;
|
2022-06-02 16:06:43 +08:00
|
|
|
|
}
|
|
|
|
|
|
2022-07-01 00:33:35 +08:00
|
|
|
|
static void
|
|
|
|
|
schedule_delayed_monitor_work(struct kfree_rcu_cpu *krcp)
|
|
|
|
|
{
|
|
|
|
|
long delay, delay_left;
|
|
|
|
|
|
2022-12-14 20:06:29 +08:00
|
|
|
|
delay = krc_count(krcp) >= KVFREE_BULK_MAX_ENTR ? 1:KFREE_DRAIN_JIFFIES;
|
2022-07-01 00:33:35 +08:00
|
|
|
|
if (delayed_work_pending(&krcp->monitor_work)) {
|
|
|
|
|
delay_left = krcp->monitor_work.timer.expires - jiffies;
|
|
|
|
|
if (delay < delay_left)
|
|
|
|
|
mod_delayed_work(system_wq, &krcp->monitor_work, delay);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
queue_delayed_work(system_wq, &krcp->monitor_work, delay);
|
|
|
|
|
}
|
|
|
|
|
|
2022-12-14 20:06:30 +08:00
|
|
|
|
static void
|
|
|
|
|
kvfree_rcu_drain_ready(struct kfree_rcu_cpu *krcp)
|
|
|
|
|
{
|
|
|
|
|
struct list_head bulk_ready[FREE_N_CHANNELS];
|
|
|
|
|
struct kvfree_rcu_bulk_data *bnode, *n;
|
|
|
|
|
struct rcu_head *head_ready = NULL;
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
int i;
|
|
|
|
|
|
|
|
|
|
raw_spin_lock_irqsave(&krcp->lock, flags);
|
|
|
|
|
for (i = 0; i < FREE_N_CHANNELS; i++) {
|
|
|
|
|
INIT_LIST_HEAD(&bulk_ready[i]);
|
|
|
|
|
|
|
|
|
|
list_for_each_entry_safe_reverse(bnode, n, &krcp->bulk_head[i], list) {
|
2023-04-04 07:49:14 +08:00
|
|
|
|
if (!poll_state_synchronize_rcu_full(&bnode->gp_snap))
|
2022-12-14 20:06:30 +08:00
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
atomic_sub(bnode->nr_records, &krcp->bulk_count[i]);
|
|
|
|
|
list_move(&bnode->list, &bulk_ready[i]);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (krcp->head && poll_state_synchronize_rcu(krcp->head_gp_snap)) {
|
|
|
|
|
head_ready = krcp->head;
|
|
|
|
|
atomic_set(&krcp->head_count, 0);
|
|
|
|
|
WRITE_ONCE(krcp->head, NULL);
|
|
|
|
|
}
|
|
|
|
|
raw_spin_unlock_irqrestore(&krcp->lock, flags);
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < FREE_N_CHANNELS; i++) {
|
|
|
|
|
list_for_each_entry_safe(bnode, n, &bulk_ready[i], list)
|
|
|
|
|
kvfree_rcu_bulk(krcp, bnode, i);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (head_ready)
|
|
|
|
|
kvfree_rcu_list(head_ready);
|
|
|
|
|
}
|
|
|
|
|
|
2014-03-19 02:48:48 +08:00
|
|
|
|
/*
|
2021-04-21 19:22:52 +08:00
|
|
|
|
* This function is invoked after the KFREE_DRAIN_JIFFIES timeout.
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
*/
|
2021-04-21 19:22:52 +08:00
|
|
|
|
static void kfree_rcu_monitor(struct work_struct *work)
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
{
|
2021-04-21 19:22:52 +08:00
|
|
|
|
struct kfree_rcu_cpu *krcp = container_of(work,
|
|
|
|
|
struct kfree_rcu_cpu, monitor_work.work);
|
|
|
|
|
unsigned long flags;
|
2020-05-26 05:47:53 +08:00
|
|
|
|
int i, j;
|
2019-09-20 05:58:26 +08:00
|
|
|
|
|
2022-12-14 20:06:30 +08:00
|
|
|
|
// Drain ready for reclaim.
|
|
|
|
|
kvfree_rcu_drain_ready(krcp);
|
|
|
|
|
|
2021-04-21 19:22:52 +08:00
|
|
|
|
raw_spin_lock_irqsave(&krcp->lock, flags);
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
|
2021-04-21 19:22:52 +08:00
|
|
|
|
// Attempt to start a new batch.
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
for (i = 0; i < KFREE_N_BATCHES; i++) {
|
2021-04-21 19:22:52 +08:00
|
|
|
|
struct kfree_rcu_cpu_work *krwp = &(krcp->krw_arr[i]);
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
|
rcu/kvfree: Avoid freeing new kfree_rcu() memory after old grace period
Memory passed to kvfree_rcu() that is to be freed is tracked by a
per-CPU kfree_rcu_cpu structure, which in turn contains pointers
to kvfree_rcu_bulk_data structures that contain pointers to memory
that has not yet been handed to RCU, along with an kfree_rcu_cpu_work
structure that tracks the memory that has already been handed to RCU.
These structures track three categories of memory: (1) Memory for
kfree(), (2) Memory for kvfree(), and (3) Memory for both that arrived
during an OOM episode. The first two categories are tracked in a
cache-friendly manner involving a dynamically allocated page of pointers
(the aforementioned kvfree_rcu_bulk_data structures), while the third
uses a simple (but decidedly cache-unfriendly) linked list through the
rcu_head structures in each block of memory.
On a given CPU, these three categories are handled as a unit, with that
CPU's kfree_rcu_cpu_work structure having one pointer for each of the
three categories. Clearly, new memory for a given category cannot be
placed in the corresponding kfree_rcu_cpu_work structure until any old
memory has had its grace period elapse and thus has been removed. And
the kfree_rcu_monitor() function does in fact check for this.
Except that the kfree_rcu_monitor() function checks these pointers one
at a time. This means that if the previous kfree_rcu() memory passed
to RCU had only category 1 and the current one has only category 2, the
kfree_rcu_monitor() function will send that current category-2 memory
along immediately. This can result in memory being freed too soon,
that is, out from under unsuspecting RCU readers.
To see this, consider the following sequence of events, in which:
o Task A on CPU 0 calls rcu_read_lock(), then uses "from_cset",
then is preempted.
o CPU 1 calls kfree_rcu(cset, rcu_head) in order to free "from_cset"
after a later grace period. Except that "from_cset" is freed
right after the previous grace period ended, so that "from_cset"
is immediately freed. Task A resumes and references "from_cset"'s
member, after which nothing good happens.
In full detail:
CPU 0 CPU 1
---------------------- ----------------------
count_memcg_event_mm()
|rcu_read_lock() <---
|mem_cgroup_from_task()
|// css_set_ptr is the "from_cset" mentioned on CPU 1
|css_set_ptr = rcu_dereference((task)->cgroups)
|// Hard irq comes, current task is scheduled out.
cgroup_attach_task()
|cgroup_migrate()
|cgroup_migrate_execute()
|css_set_move_task(task, from_cset, to_cset, true)
|cgroup_move_task(task, to_cset)
|rcu_assign_pointer(.., to_cset)
|...
|cgroup_migrate_finish()
|put_css_set_locked(from_cset)
|from_cset->refcount return 0
|kfree_rcu(cset, rcu_head) // free from_cset after new gp
|add_ptr_to_bulk_krc_lock()
|schedule_delayed_work(&krcp->monitor_work, ..)
kfree_rcu_monitor()
|krcp->bulk_head[0]'s work attached to krwp->bulk_head_free[]
|queue_rcu_work(system_wq, &krwp->rcu_work)
|if rwork->rcu.work is not in WORK_STRUCT_PENDING_BIT state,
|call_rcu(&rwork->rcu, rcu_work_rcufn) <--- request new gp
// There is a perious call_rcu(.., rcu_work_rcufn)
// gp end, rcu_work_rcufn() is called.
rcu_work_rcufn()
|__queue_work(.., rwork->wq, &rwork->work);
|kfree_rcu_work()
|krwp->bulk_head_free[0] bulk is freed before new gp end!!!
|The "from_cset" is freed before new gp end.
// the task resumes some time later.
|css_set_ptr->subsys[(subsys_id) <--- Caused kernel crash, because css_set_ptr is freed.
This commit therefore causes kfree_rcu_monitor() to refrain from moving
kfree_rcu() memory to the kfree_rcu_cpu_work structure until the RCU
grace period has completed for all three categories.
v2: Use helper function instead of inserted code block at kfree_rcu_monitor().
Fixes: 34c881745549 ("rcu: Support kfree_bulk() interface in kfree_rcu()")
Fixes: 5f3c8d620447 ("rcu/tree: Maintain separate array for vmalloc ptrs")
Reported-by: Mukesh Ojha <quic_mojha@quicinc.com>
Signed-off-by: Ziwei Dai <ziwei.dai@unisoc.com>
Reviewed-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2023-03-31 20:42:09 +08:00
|
|
|
|
// Try to detach bulk_head or head and attach it, only when
|
|
|
|
|
// all channels are free. Any channel is not free means at krwp
|
|
|
|
|
// there is on-going rcu work to handle krwp's free business.
|
|
|
|
|
if (need_wait_for_krwp_work(krwp))
|
|
|
|
|
continue;
|
2022-11-29 23:58:19 +08:00
|
|
|
|
|
rcu/kvfree: Avoid freeing new kfree_rcu() memory after old grace period
Memory passed to kvfree_rcu() that is to be freed is tracked by a
per-CPU kfree_rcu_cpu structure, which in turn contains pointers
to kvfree_rcu_bulk_data structures that contain pointers to memory
that has not yet been handed to RCU, along with an kfree_rcu_cpu_work
structure that tracks the memory that has already been handed to RCU.
These structures track three categories of memory: (1) Memory for
kfree(), (2) Memory for kvfree(), and (3) Memory for both that arrived
during an OOM episode. The first two categories are tracked in a
cache-friendly manner involving a dynamically allocated page of pointers
(the aforementioned kvfree_rcu_bulk_data structures), while the third
uses a simple (but decidedly cache-unfriendly) linked list through the
rcu_head structures in each block of memory.
On a given CPU, these three categories are handled as a unit, with that
CPU's kfree_rcu_cpu_work structure having one pointer for each of the
three categories. Clearly, new memory for a given category cannot be
placed in the corresponding kfree_rcu_cpu_work structure until any old
memory has had its grace period elapse and thus has been removed. And
the kfree_rcu_monitor() function does in fact check for this.
Except that the kfree_rcu_monitor() function checks these pointers one
at a time. This means that if the previous kfree_rcu() memory passed
to RCU had only category 1 and the current one has only category 2, the
kfree_rcu_monitor() function will send that current category-2 memory
along immediately. This can result in memory being freed too soon,
that is, out from under unsuspecting RCU readers.
To see this, consider the following sequence of events, in which:
o Task A on CPU 0 calls rcu_read_lock(), then uses "from_cset",
then is preempted.
o CPU 1 calls kfree_rcu(cset, rcu_head) in order to free "from_cset"
after a later grace period. Except that "from_cset" is freed
right after the previous grace period ended, so that "from_cset"
is immediately freed. Task A resumes and references "from_cset"'s
member, after which nothing good happens.
In full detail:
CPU 0 CPU 1
---------------------- ----------------------
count_memcg_event_mm()
|rcu_read_lock() <---
|mem_cgroup_from_task()
|// css_set_ptr is the "from_cset" mentioned on CPU 1
|css_set_ptr = rcu_dereference((task)->cgroups)
|// Hard irq comes, current task is scheduled out.
cgroup_attach_task()
|cgroup_migrate()
|cgroup_migrate_execute()
|css_set_move_task(task, from_cset, to_cset, true)
|cgroup_move_task(task, to_cset)
|rcu_assign_pointer(.., to_cset)
|...
|cgroup_migrate_finish()
|put_css_set_locked(from_cset)
|from_cset->refcount return 0
|kfree_rcu(cset, rcu_head) // free from_cset after new gp
|add_ptr_to_bulk_krc_lock()
|schedule_delayed_work(&krcp->monitor_work, ..)
kfree_rcu_monitor()
|krcp->bulk_head[0]'s work attached to krwp->bulk_head_free[]
|queue_rcu_work(system_wq, &krwp->rcu_work)
|if rwork->rcu.work is not in WORK_STRUCT_PENDING_BIT state,
|call_rcu(&rwork->rcu, rcu_work_rcufn) <--- request new gp
// There is a perious call_rcu(.., rcu_work_rcufn)
// gp end, rcu_work_rcufn() is called.
rcu_work_rcufn()
|__queue_work(.., rwork->wq, &rwork->work);
|kfree_rcu_work()
|krwp->bulk_head_free[0] bulk is freed before new gp end!!!
|The "from_cset" is freed before new gp end.
// the task resumes some time later.
|css_set_ptr->subsys[(subsys_id) <--- Caused kernel crash, because css_set_ptr is freed.
This commit therefore causes kfree_rcu_monitor() to refrain from moving
kfree_rcu() memory to the kfree_rcu_cpu_work structure until the RCU
grace period has completed for all three categories.
v2: Use helper function instead of inserted code block at kfree_rcu_monitor().
Fixes: 34c881745549 ("rcu: Support kfree_bulk() interface in kfree_rcu()")
Fixes: 5f3c8d620447 ("rcu/tree: Maintain separate array for vmalloc ptrs")
Reported-by: Mukesh Ojha <quic_mojha@quicinc.com>
Signed-off-by: Ziwei Dai <ziwei.dai@unisoc.com>
Reviewed-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2023-03-31 20:42:09 +08:00
|
|
|
|
// kvfree_rcu_drain_ready() might handle this krcp, if so give up.
|
|
|
|
|
if (need_offload_krc(krcp)) {
|
2021-04-28 21:44:22 +08:00
|
|
|
|
// Channel 1 corresponds to the SLAB-pointer bulk path.
|
|
|
|
|
// Channel 2 corresponds to vmalloc-pointer bulk path.
|
2020-05-26 05:47:53 +08:00
|
|
|
|
for (j = 0; j < FREE_N_CHANNELS; j++) {
|
2022-12-14 20:06:29 +08:00
|
|
|
|
if (list_empty(&krwp->bulk_head_free[j])) {
|
|
|
|
|
atomic_set(&krcp->bulk_count[j], 0);
|
2022-12-14 20:06:30 +08:00
|
|
|
|
list_replace_init(&krcp->bulk_head[j],
|
|
|
|
|
&krwp->bulk_head_free[j]);
|
2020-05-26 05:47:53 +08:00
|
|
|
|
}
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
}
|
|
|
|
|
|
2021-04-28 21:44:22 +08:00
|
|
|
|
// Channel 3 corresponds to both SLAB and vmalloc
|
|
|
|
|
// objects queued on the linked list.
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
if (!krwp->head_free) {
|
|
|
|
|
krwp->head_free = krcp->head;
|
2023-04-04 22:13:00 +08:00
|
|
|
|
get_state_synchronize_rcu_full(&krwp->head_free_gp_snap);
|
2022-12-14 20:06:29 +08:00
|
|
|
|
atomic_set(&krcp->head_count, 0);
|
2022-12-14 20:06:30 +08:00
|
|
|
|
WRITE_ONCE(krcp->head, NULL);
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
}
|
|
|
|
|
|
2021-04-21 19:22:52 +08:00
|
|
|
|
// One work is per one batch, so there are three
|
|
|
|
|
// "free channels", the batch can handle. It can
|
|
|
|
|
// be that the work is in the pending state when
|
|
|
|
|
// channels have been detached following by each
|
|
|
|
|
// other.
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
queue_rcu_work(system_wq, &krwp->rcu_work);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2022-11-29 23:58:21 +08:00
|
|
|
|
raw_spin_unlock_irqrestore(&krcp->lock, flags);
|
|
|
|
|
|
2021-04-21 19:22:52 +08:00
|
|
|
|
// If there is nothing to detach, it means that our job is
|
|
|
|
|
// successfully done here. In case of having at least one
|
|
|
|
|
// of the channels that is still busy we should rearm the
|
|
|
|
|
// work to repeat an attempt. Because previous batches are
|
|
|
|
|
// still in progress.
|
2022-06-02 16:06:43 +08:00
|
|
|
|
if (need_offload_krc(krcp))
|
2022-07-01 00:33:35 +08:00
|
|
|
|
schedule_delayed_monitor_work(krcp);
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
}
|
|
|
|
|
|
2020-10-30 00:50:04 +08:00
|
|
|
|
static enum hrtimer_restart
|
|
|
|
|
schedule_page_work_fn(struct hrtimer *t)
|
|
|
|
|
{
|
|
|
|
|
struct kfree_rcu_cpu *krcp =
|
|
|
|
|
container_of(t, struct kfree_rcu_cpu, hrtimer);
|
|
|
|
|
|
2021-04-16 01:19:56 +08:00
|
|
|
|
queue_delayed_work(system_highpri_wq, &krcp->page_cache_work, 0);
|
2020-10-30 00:50:04 +08:00
|
|
|
|
return HRTIMER_NORESTART;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void fill_page_cache_func(struct work_struct *work)
|
|
|
|
|
{
|
|
|
|
|
struct kvfree_rcu_bulk_data *bnode;
|
|
|
|
|
struct kfree_rcu_cpu *krcp =
|
|
|
|
|
container_of(work, struct kfree_rcu_cpu,
|
2021-04-16 01:19:56 +08:00
|
|
|
|
page_cache_work.work);
|
2020-10-30 00:50:04 +08:00
|
|
|
|
unsigned long flags;
|
2021-04-16 01:19:56 +08:00
|
|
|
|
int nr_pages;
|
2020-10-30 00:50:04 +08:00
|
|
|
|
bool pushed;
|
|
|
|
|
int i;
|
|
|
|
|
|
2021-04-16 01:19:56 +08:00
|
|
|
|
nr_pages = atomic_read(&krcp->backoff_page_cache_fill) ?
|
|
|
|
|
1 : rcu_min_cached_objs;
|
|
|
|
|
|
2023-04-12 22:31:27 +08:00
|
|
|
|
for (i = READ_ONCE(krcp->nr_bkv_objs); i < nr_pages; i++) {
|
2020-10-30 00:50:04 +08:00
|
|
|
|
bnode = (struct kvfree_rcu_bulk_data *)
|
kvfree_rcu: Use same set of GFP flags as does single-argument
Running an rcuscale stress-suite can lead to "Out of memory" of a
system. This can happen under high memory pressure with a small amount
of physical memory.
For example, a KVM test configuration with 64 CPUs and 512 megabytes
can result in OOM when running rcuscale with below parameters:
../kvm.sh --torture rcuscale --allcpus --duration 10 --kconfig CONFIG_NR_CPUS=64 \
--bootargs "rcuscale.kfree_rcu_test=1 rcuscale.kfree_nthreads=16 rcuscale.holdoff=20 \
rcuscale.kfree_loops=10000 torture.disable_onoff_at_boot" --trust-make
<snip>
[ 12.054448] kworker/1:1H invoked oom-killer: gfp_mask=0x2cc0(GFP_KERNEL|__GFP_NOWARN), order=0, oom_score_adj=0
[ 12.055303] CPU: 1 PID: 377 Comm: kworker/1:1H Not tainted 5.11.0-rc3+ #510
[ 12.055416] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.12.0-1 04/01/2014
[ 12.056485] Workqueue: events_highpri fill_page_cache_func
[ 12.056485] Call Trace:
[ 12.056485] dump_stack+0x57/0x6a
[ 12.056485] dump_header+0x4c/0x30a
[ 12.056485] ? del_timer_sync+0x20/0x30
[ 12.056485] out_of_memory.cold.47+0xa/0x7e
[ 12.056485] __alloc_pages_slowpath.constprop.123+0x82f/0xc00
[ 12.056485] __alloc_pages_nodemask+0x289/0x2c0
[ 12.056485] __get_free_pages+0x8/0x30
[ 12.056485] fill_page_cache_func+0x39/0xb0
[ 12.056485] process_one_work+0x1ed/0x3b0
[ 12.056485] ? process_one_work+0x3b0/0x3b0
[ 12.060485] worker_thread+0x28/0x3c0
[ 12.060485] ? process_one_work+0x3b0/0x3b0
[ 12.060485] kthread+0x138/0x160
[ 12.060485] ? kthread_park+0x80/0x80
[ 12.060485] ret_from_fork+0x22/0x30
[ 12.062156] Mem-Info:
[ 12.062350] active_anon:0 inactive_anon:0 isolated_anon:0
[ 12.062350] active_file:0 inactive_file:0 isolated_file:0
[ 12.062350] unevictable:0 dirty:0 writeback:0
[ 12.062350] slab_reclaimable:2797 slab_unreclaimable:80920
[ 12.062350] mapped:1 shmem:2 pagetables:8 bounce:0
[ 12.062350] free:10488 free_pcp:1227 free_cma:0
...
[ 12.101610] Out of memory and no killable processes...
[ 12.102042] Kernel panic - not syncing: System is deadlocked on memory
[ 12.102583] CPU: 1 PID: 377 Comm: kworker/1:1H Not tainted 5.11.0-rc3+ #510
[ 12.102600] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.12.0-1 04/01/2014
<snip>
Because kvfree_rcu() has a fallback path, memory allocation failure is
not the end of the world. Furthermore, the added overhead of aggressive
GFP settings must be balanced against the overhead of the fallback path,
which is a cache miss for double-argument kvfree_rcu() and a call to
synchronize_rcu() for single-argument kvfree_rcu(). The current choice
of GFP_KERNEL|__GFP_NOWARN can result in longer latencies than a call
to synchronize_rcu(), so less-tenacious GFP flags would be helpful.
Here is the tradeoff that must be balanced:
a) Minimize use of the fallback path,
b) Avoid pushing the system into OOM,
c) Bound allocation latency to that of synchronize_rcu(), and
d) Leave the emergency reserves to use cases lacking fallbacks.
This commit therefore changes GFP flags from GFP_KERNEL|__GFP_NOWARN to
GFP_KERNEL|__GFP_NORETRY|__GFP_NOMEMALLOC|__GFP_NOWARN. This combination
leaves the emergency reserves alone and can initiate reclaim, but will
not invoke the OOM killer.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-01-30 04:05:05 +08:00
|
|
|
|
__get_free_page(GFP_KERNEL | __GFP_NORETRY | __GFP_NOMEMALLOC | __GFP_NOWARN);
|
2020-10-30 00:50:04 +08:00
|
|
|
|
|
rcu: Back off upon fill_page_cache_func() allocation failure
The fill_page_cache_func() function allocates couple of pages to store
kvfree_rcu_bulk_data structures. This is a lightweight (GFP_NORETRY)
allocation which can fail under memory pressure. The function will,
however keep retrying even when the previous attempt has failed.
This retrying is in theory correct, but in practice the allocation is
invoked from workqueue context, which means that if the memory reclaim
gets stuck, these retries can hog the worker for quite some time.
Although the workqueues subsystem automatically adjusts concurrency, such
adjustment is not guaranteed to happen until the worker context sleeps.
And the fill_page_cache_func() function's retry loop is not guaranteed
to sleep (see the should_reclaim_retry() function).
And we have seen this function cause workqueue lockups:
kernel: BUG: workqueue lockup - pool cpus=93 node=1 flags=0x1 nice=0 stuck for 32s!
[...]
kernel: pool 74: cpus=37 node=0 flags=0x1 nice=0 hung=32s workers=2 manager: 2146
kernel: pwq 498: cpus=249 node=1 flags=0x1 nice=0 active=4/256 refcnt=5
kernel: in-flight: 1917:fill_page_cache_func
kernel: pending: dbs_work_handler, free_work, kfree_rcu_monitor
Originally, we thought that the root cause of this lockup was several
retries with direct reclaim, but this is not yet confirmed. Furthermore,
we have seen similar lockups without any heavy memory pressure. This
suggests that there are other factors contributing to these lockups.
However, it is not really clear that endless retries are desireable.
So let's make the fill_page_cache_func() function back off after
allocation failure.
Cc: Uladzislau Rezki (Sony) <urezki@gmail.com>
Cc: "Paul E. McKenney" <paulmck@kernel.org>
Cc: Frederic Weisbecker <frederic@kernel.org>
Cc: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: Joel Fernandes <joel@joelfernandes.org>
Signed-off-by: Michal Hocko <mhocko@suse.com>
Reviewed-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2022-06-22 19:47:11 +08:00
|
|
|
|
if (!bnode)
|
|
|
|
|
break;
|
2020-10-30 00:50:04 +08:00
|
|
|
|
|
rcu: Back off upon fill_page_cache_func() allocation failure
The fill_page_cache_func() function allocates couple of pages to store
kvfree_rcu_bulk_data structures. This is a lightweight (GFP_NORETRY)
allocation which can fail under memory pressure. The function will,
however keep retrying even when the previous attempt has failed.
This retrying is in theory correct, but in practice the allocation is
invoked from workqueue context, which means that if the memory reclaim
gets stuck, these retries can hog the worker for quite some time.
Although the workqueues subsystem automatically adjusts concurrency, such
adjustment is not guaranteed to happen until the worker context sleeps.
And the fill_page_cache_func() function's retry loop is not guaranteed
to sleep (see the should_reclaim_retry() function).
And we have seen this function cause workqueue lockups:
kernel: BUG: workqueue lockup - pool cpus=93 node=1 flags=0x1 nice=0 stuck for 32s!
[...]
kernel: pool 74: cpus=37 node=0 flags=0x1 nice=0 hung=32s workers=2 manager: 2146
kernel: pwq 498: cpus=249 node=1 flags=0x1 nice=0 active=4/256 refcnt=5
kernel: in-flight: 1917:fill_page_cache_func
kernel: pending: dbs_work_handler, free_work, kfree_rcu_monitor
Originally, we thought that the root cause of this lockup was several
retries with direct reclaim, but this is not yet confirmed. Furthermore,
we have seen similar lockups without any heavy memory pressure. This
suggests that there are other factors contributing to these lockups.
However, it is not really clear that endless retries are desireable.
So let's make the fill_page_cache_func() function back off after
allocation failure.
Cc: Uladzislau Rezki (Sony) <urezki@gmail.com>
Cc: "Paul E. McKenney" <paulmck@kernel.org>
Cc: Frederic Weisbecker <frederic@kernel.org>
Cc: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: Joel Fernandes <joel@joelfernandes.org>
Signed-off-by: Michal Hocko <mhocko@suse.com>
Reviewed-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2022-06-22 19:47:11 +08:00
|
|
|
|
raw_spin_lock_irqsave(&krcp->lock, flags);
|
|
|
|
|
pushed = put_cached_bnode(krcp, bnode);
|
|
|
|
|
raw_spin_unlock_irqrestore(&krcp->lock, flags);
|
|
|
|
|
|
|
|
|
|
if (!pushed) {
|
|
|
|
|
free_page((unsigned long) bnode);
|
|
|
|
|
break;
|
2020-10-30 00:50:04 +08:00
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
atomic_set(&krcp->work_in_progress, 0);
|
2021-04-16 01:19:56 +08:00
|
|
|
|
atomic_set(&krcp->backoff_page_cache_fill, 0);
|
2020-10-30 00:50:04 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void
|
|
|
|
|
run_page_cache_worker(struct kfree_rcu_cpu *krcp)
|
|
|
|
|
{
|
2023-04-11 21:13:41 +08:00
|
|
|
|
// If cache disabled, bail out.
|
|
|
|
|
if (!rcu_min_cached_objs)
|
|
|
|
|
return;
|
|
|
|
|
|
2020-10-30 00:50:04 +08:00
|
|
|
|
if (rcu_scheduler_active == RCU_SCHEDULER_RUNNING &&
|
|
|
|
|
!atomic_xchg(&krcp->work_in_progress, 1)) {
|
2021-04-16 01:19:56 +08:00
|
|
|
|
if (atomic_read(&krcp->backoff_page_cache_fill)) {
|
|
|
|
|
queue_delayed_work(system_wq,
|
|
|
|
|
&krcp->page_cache_work,
|
|
|
|
|
msecs_to_jiffies(rcu_delay_page_cache_fill_msec));
|
|
|
|
|
} else {
|
|
|
|
|
hrtimer_init(&krcp->hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
|
|
|
|
|
krcp->hrtimer.function = schedule_page_work_fn;
|
|
|
|
|
hrtimer_start(&krcp->hrtimer, 0, HRTIMER_MODE_REL);
|
|
|
|
|
}
|
2020-10-30 00:50:04 +08:00
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2021-01-21 00:21:46 +08:00
|
|
|
|
// Record ptr in a page managed by krcp, with the pre-krc_this_cpu_lock()
|
|
|
|
|
// state specified by flags. If can_alloc is true, the caller must
|
|
|
|
|
// be schedulable and not be holding any locks or mutexes that might be
|
|
|
|
|
// acquired by the memory allocator or anything that it might invoke.
|
|
|
|
|
// Returns true if ptr was successfully recorded, else the caller must
|
|
|
|
|
// use a fallback.
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
static inline bool
|
2021-01-21 00:21:46 +08:00
|
|
|
|
add_ptr_to_bulk_krc_lock(struct kfree_rcu_cpu **krcp,
|
|
|
|
|
unsigned long *flags, void *ptr, bool can_alloc)
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
{
|
2020-05-26 05:47:53 +08:00
|
|
|
|
struct kvfree_rcu_bulk_data *bnode;
|
|
|
|
|
int idx;
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
|
2021-01-21 00:21:46 +08:00
|
|
|
|
*krcp = krc_this_cpu_lock(flags);
|
|
|
|
|
if (unlikely(!(*krcp)->initialized))
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
return false;
|
|
|
|
|
|
2020-05-26 05:47:53 +08:00
|
|
|
|
idx = !!is_vmalloc_addr(ptr);
|
2022-11-29 23:58:19 +08:00
|
|
|
|
bnode = list_first_entry_or_null(&(*krcp)->bulk_head[idx],
|
|
|
|
|
struct kvfree_rcu_bulk_data, list);
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
|
|
|
|
|
/* Check if a new block is required. */
|
2022-11-29 23:58:19 +08:00
|
|
|
|
if (!bnode || bnode->nr_records == KVFREE_BULK_MAX_ENTR) {
|
2021-01-21 00:21:46 +08:00
|
|
|
|
bnode = get_cached_bnode(*krcp);
|
|
|
|
|
if (!bnode && can_alloc) {
|
|
|
|
|
krc_this_cpu_unlock(*krcp, *flags);
|
2021-01-30 00:16:03 +08:00
|
|
|
|
|
|
|
|
|
// __GFP_NORETRY - allows a light-weight direct reclaim
|
|
|
|
|
// what is OK from minimizing of fallback hitting point of
|
|
|
|
|
// view. Apart of that it forbids any OOM invoking what is
|
|
|
|
|
// also beneficial since we are about to release memory soon.
|
|
|
|
|
//
|
|
|
|
|
// __GFP_NOMEMALLOC - prevents from consuming of all the
|
|
|
|
|
// memory reserves. Please note we have a fallback path.
|
|
|
|
|
//
|
|
|
|
|
// __GFP_NOWARN - it is supposed that an allocation can
|
|
|
|
|
// be failed under low memory or high memory pressure
|
|
|
|
|
// scenarios.
|
2021-01-21 00:21:46 +08:00
|
|
|
|
bnode = (struct kvfree_rcu_bulk_data *)
|
2021-01-30 00:16:03 +08:00
|
|
|
|
__get_free_page(GFP_KERNEL | __GFP_NORETRY | __GFP_NOMEMALLOC | __GFP_NOWARN);
|
2023-04-08 22:25:30 +08:00
|
|
|
|
raw_spin_lock_irqsave(&(*krcp)->lock, *flags);
|
2021-01-21 00:21:46 +08:00
|
|
|
|
}
|
|
|
|
|
|
2020-10-30 00:50:04 +08:00
|
|
|
|
if (!bnode)
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
return false;
|
|
|
|
|
|
2022-11-29 23:58:19 +08:00
|
|
|
|
// Initialize the new block and attach it.
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
bnode->nr_records = 0;
|
2022-11-29 23:58:19 +08:00
|
|
|
|
list_add(&bnode->list, &(*krcp)->bulk_head[idx]);
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
}
|
|
|
|
|
|
rcu/kvfree: Use a polled API to speedup a reclaim process
Currently all objects placed into a batch wait for a full grace period
to elapse after that batch is ready to send to RCU. However, this
can unnecessarily delay freeing of the first objects that were added
to the batch. After all, several RCU grace periods might have elapsed
since those objects were added, and if so, there is no point in further
deferring their freeing.
This commit therefore adds per-page grace-period snapshots which are
obtained from get_state_synchronize_rcu(). When the batch is ready
to be passed to call_rcu(), each page's snapshot is checked by passing
it to poll_state_synchronize_rcu(). If a given page's RCU grace period
has already elapsed, its objects are freed immediately by kvfree_rcu_bulk().
Otherwise, these objects are freed after a call to synchronize_rcu().
This approach requires that the pages be traversed in reverse order,
that is, the oldest ones first.
Test example:
kvm.sh --memory 10G --torture rcuscale --allcpus --duration 1 \
--kconfig CONFIG_NR_CPUS=64 \
--kconfig CONFIG_RCU_NOCB_CPU=y \
--kconfig CONFIG_RCU_NOCB_CPU_DEFAULT_ALL=y \
--kconfig CONFIG_RCU_LAZY=n \
--bootargs "rcuscale.kfree_rcu_test=1 rcuscale.kfree_nthreads=16 \
rcuscale.holdoff=20 rcuscale.kfree_loops=10000 \
torture.disable_onoff_at_boot" --trust-make
Before this commit:
Total time taken by all kfree'ers: 8535693700 ns, loops: 10000, batches: 1188, memory footprint: 2248MB
Total time taken by all kfree'ers: 8466933582 ns, loops: 10000, batches: 1157, memory footprint: 2820MB
Total time taken by all kfree'ers: 5375602446 ns, loops: 10000, batches: 1130, memory footprint: 6502MB
Total time taken by all kfree'ers: 7523283832 ns, loops: 10000, batches: 1006, memory footprint: 3343MB
Total time taken by all kfree'ers: 6459171956 ns, loops: 10000, batches: 1150, memory footprint: 6549MB
After this commit:
Total time taken by all kfree'ers: 8560060176 ns, loops: 10000, batches: 1787, memory footprint: 61MB
Total time taken by all kfree'ers: 8573885501 ns, loops: 10000, batches: 1777, memory footprint: 93MB
Total time taken by all kfree'ers: 8320000202 ns, loops: 10000, batches: 1727, memory footprint: 66MB
Total time taken by all kfree'ers: 8552718794 ns, loops: 10000, batches: 1790, memory footprint: 75MB
Total time taken by all kfree'ers: 8601368792 ns, loops: 10000, batches: 1724, memory footprint: 62MB
The reduction in memory footprint is well in excess of an order of
magnitude.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2022-11-29 23:58:22 +08:00
|
|
|
|
// Finally insert and update the GP for this page.
|
2022-11-29 23:58:19 +08:00
|
|
|
|
bnode->records[bnode->nr_records++] = ptr;
|
2023-04-04 07:49:14 +08:00
|
|
|
|
get_state_synchronize_rcu_full(&bnode->gp_snap);
|
2022-12-14 20:06:29 +08:00
|
|
|
|
atomic_inc(&(*krcp)->bulk_count[idx]);
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
/*
|
2021-03-31 04:47:42 +08:00
|
|
|
|
* Queue a request for lazy invocation of the appropriate free routine
|
|
|
|
|
* after a grace period. Please note that three paths are maintained,
|
|
|
|
|
* two for the common case using arrays of pointers and a third one that
|
|
|
|
|
* is used only when the main paths cannot be used, for example, due to
|
|
|
|
|
* memory pressure.
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
*
|
2020-05-26 05:47:55 +08:00
|
|
|
|
* Each kvfree_call_rcu() request is added to a batch. The batch will be drained
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
* every KFREE_DRAIN_JIFFIES number of jiffies. All the objects in the batch will
|
|
|
|
|
* be free'd in workqueue context. This allows us to: batch requests together to
|
2020-05-26 05:47:53 +08:00
|
|
|
|
* reduce the number of grace periods during heavy kfree_rcu()/kvfree_rcu() load.
|
2014-03-19 02:48:48 +08:00
|
|
|
|
*/
|
2022-10-25 22:46:12 +08:00
|
|
|
|
void kvfree_call_rcu(struct rcu_head *head, void *ptr)
|
2014-03-19 02:48:48 +08:00
|
|
|
|
{
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
unsigned long flags;
|
|
|
|
|
struct kfree_rcu_cpu *krcp;
|
2020-05-26 05:47:58 +08:00
|
|
|
|
bool success;
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
|
2022-10-25 22:46:12 +08:00
|
|
|
|
/*
|
|
|
|
|
* Please note there is a limitation for the head-less
|
|
|
|
|
* variant, that is why there is a clear rule for such
|
|
|
|
|
* objects: it can be used from might_sleep() context
|
|
|
|
|
* only. For other places please embed an rcu_head to
|
|
|
|
|
* your data.
|
|
|
|
|
*/
|
|
|
|
|
if (!head)
|
2020-05-26 05:47:58 +08:00
|
|
|
|
might_sleep();
|
|
|
|
|
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
// Queue the object but don't yet schedule the batch.
|
2020-05-26 05:47:48 +08:00
|
|
|
|
if (debug_rcu_head_queue(ptr)) {
|
2019-09-23 04:03:17 +08:00
|
|
|
|
// Probable double kfree_rcu(), just leak.
|
|
|
|
|
WARN_ONCE(1, "%s(): Double-freed call. rcu_head %p\n",
|
|
|
|
|
__func__, head);
|
2020-05-26 05:47:58 +08:00
|
|
|
|
|
|
|
|
|
// Mark as success and leave.
|
2021-01-21 00:21:46 +08:00
|
|
|
|
return;
|
2019-09-23 04:03:17 +08:00
|
|
|
|
}
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
|
2021-11-16 07:23:02 +08:00
|
|
|
|
kasan_record_aux_stack_noalloc(ptr);
|
2021-01-21 00:21:46 +08:00
|
|
|
|
success = add_ptr_to_bulk_krc_lock(&krcp, &flags, ptr, !head);
|
2020-05-26 05:47:58 +08:00
|
|
|
|
if (!success) {
|
2020-10-30 00:50:04 +08:00
|
|
|
|
run_page_cache_worker(krcp);
|
|
|
|
|
|
2020-05-26 05:47:58 +08:00
|
|
|
|
if (head == NULL)
|
|
|
|
|
// Inline if kvfree_rcu(one_arg) call.
|
|
|
|
|
goto unlock_return;
|
|
|
|
|
|
2022-10-25 22:46:12 +08:00
|
|
|
|
head->func = ptr;
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
head->next = krcp->head;
|
2022-11-29 23:58:21 +08:00
|
|
|
|
WRITE_ONCE(krcp->head, head);
|
2022-12-14 20:06:29 +08:00
|
|
|
|
atomic_inc(&krcp->head_count);
|
2022-12-14 20:06:30 +08:00
|
|
|
|
|
|
|
|
|
// Take a snapshot for this krcp.
|
|
|
|
|
krcp->head_gp_snap = get_state_synchronize_rcu();
|
2020-05-26 05:47:58 +08:00
|
|
|
|
success = true;
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
}
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
|
2023-10-01 01:46:56 +08:00
|
|
|
|
/*
|
|
|
|
|
* The kvfree_rcu() caller considers the pointer freed at this point
|
|
|
|
|
* and likely removes any references to it. Since the actual slab
|
|
|
|
|
* freeing (and kmemleak_free()) is deferred, tell kmemleak to ignore
|
|
|
|
|
* this object (no scanning or false positives reporting).
|
|
|
|
|
*/
|
|
|
|
|
kmemleak_ignore(ptr);
|
|
|
|
|
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
// Set timer to drain after KFREE_DRAIN_JIFFIES.
|
2022-06-02 16:06:43 +08:00
|
|
|
|
if (rcu_scheduler_active == RCU_SCHEDULER_RUNNING)
|
2022-07-01 00:33:35 +08:00
|
|
|
|
schedule_delayed_monitor_work(krcp);
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
|
2019-09-23 04:03:17 +08:00
|
|
|
|
unlock_return:
|
2020-05-26 05:47:50 +08:00
|
|
|
|
krc_this_cpu_unlock(krcp, flags);
|
2020-05-26 05:47:58 +08:00
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Inline kvfree() after synchronize_rcu(). We can do
|
|
|
|
|
* it from might_sleep() context only, so the current
|
|
|
|
|
* CPU can pass the QS state.
|
|
|
|
|
*/
|
|
|
|
|
if (!success) {
|
|
|
|
|
debug_rcu_head_unqueue((struct rcu_head *) ptr);
|
|
|
|
|
synchronize_rcu();
|
|
|
|
|
kvfree(ptr);
|
|
|
|
|
}
|
2014-03-19 02:48:48 +08:00
|
|
|
|
}
|
2020-05-26 05:47:55 +08:00
|
|
|
|
EXPORT_SYMBOL_GPL(kvfree_call_rcu);
|
2014-03-19 02:48:48 +08:00
|
|
|
|
|
2020-03-17 00:32:27 +08:00
|
|
|
|
static unsigned long
|
|
|
|
|
kfree_rcu_shrink_count(struct shrinker *shrink, struct shrink_control *sc)
|
|
|
|
|
{
|
|
|
|
|
int cpu;
|
2020-03-17 00:32:28 +08:00
|
|
|
|
unsigned long count = 0;
|
2020-03-17 00:32:27 +08:00
|
|
|
|
|
|
|
|
|
/* Snapshot count of all CPUs */
|
2020-08-14 14:45:57 +08:00
|
|
|
|
for_each_possible_cpu(cpu) {
|
2020-03-17 00:32:27 +08:00
|
|
|
|
struct kfree_rcu_cpu *krcp = per_cpu_ptr(&krc, cpu);
|
|
|
|
|
|
2022-12-14 20:06:29 +08:00
|
|
|
|
count += krc_count(krcp);
|
2021-04-16 01:19:57 +08:00
|
|
|
|
count += READ_ONCE(krcp->nr_bkv_objs);
|
2021-04-16 01:19:56 +08:00
|
|
|
|
atomic_set(&krcp->backoff_page_cache_fill, 1);
|
2020-03-17 00:32:27 +08:00
|
|
|
|
}
|
|
|
|
|
|
2022-06-23 06:51:02 +08:00
|
|
|
|
return count == 0 ? SHRINK_EMPTY : count;
|
2020-03-17 00:32:27 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static unsigned long
|
|
|
|
|
kfree_rcu_shrink_scan(struct shrinker *shrink, struct shrink_control *sc)
|
|
|
|
|
{
|
|
|
|
|
int cpu, freed = 0;
|
|
|
|
|
|
2020-08-14 14:45:57 +08:00
|
|
|
|
for_each_possible_cpu(cpu) {
|
2020-03-17 00:32:27 +08:00
|
|
|
|
int count;
|
|
|
|
|
struct kfree_rcu_cpu *krcp = per_cpu_ptr(&krc, cpu);
|
|
|
|
|
|
2022-12-14 20:06:29 +08:00
|
|
|
|
count = krc_count(krcp);
|
2021-04-16 01:19:56 +08:00
|
|
|
|
count += drain_page_cache(krcp);
|
2021-04-16 01:20:00 +08:00
|
|
|
|
kfree_rcu_monitor(&krcp->monitor_work.work);
|
2020-03-17 00:32:27 +08:00
|
|
|
|
|
|
|
|
|
sc->nr_to_scan -= count;
|
|
|
|
|
freed += count;
|
|
|
|
|
|
|
|
|
|
if (sc->nr_to_scan <= 0)
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
2020-06-04 18:23:20 +08:00
|
|
|
|
return freed == 0 ? SHRINK_STOP : freed;
|
2020-03-17 00:32:27 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static struct shrinker kfree_rcu_shrinker = {
|
|
|
|
|
.count_objects = kfree_rcu_shrink_count,
|
|
|
|
|
.scan_objects = kfree_rcu_shrink_scan,
|
|
|
|
|
.batch = 0,
|
|
|
|
|
.seeks = DEFAULT_SEEKS,
|
|
|
|
|
};
|
|
|
|
|
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
void __init kfree_rcu_scheduler_running(void)
|
|
|
|
|
{
|
|
|
|
|
int cpu;
|
|
|
|
|
|
2020-08-14 14:45:57 +08:00
|
|
|
|
for_each_possible_cpu(cpu) {
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
struct kfree_rcu_cpu *krcp = per_cpu_ptr(&krc, cpu);
|
|
|
|
|
|
2022-06-02 16:06:43 +08:00
|
|
|
|
if (need_offload_krc(krcp))
|
2022-07-01 00:33:35 +08:00
|
|
|
|
schedule_delayed_monitor_work(krcp);
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2018-11-30 02:42:06 +08:00
|
|
|
|
/*
|
|
|
|
|
* During early boot, any blocking grace-period wait automatically
|
2022-08-05 07:07:04 +08:00
|
|
|
|
* implies a grace period.
|
2018-11-30 02:42:06 +08:00
|
|
|
|
*
|
2022-08-05 07:07:04 +08:00
|
|
|
|
* Later on, this could in theory be the case for kernels built with
|
|
|
|
|
* CONFIG_SMP=y && CONFIG_PREEMPTION=y running on a single CPU, but this
|
|
|
|
|
* is not a common case. Furthermore, this optimization would cause
|
|
|
|
|
* the rcu_gp_oldstate structure to expand by 50%, so this potential
|
|
|
|
|
* grace-period optimization is ignored once the scheduler is running.
|
2018-11-30 02:42:06 +08:00
|
|
|
|
*/
|
|
|
|
|
static int rcu_blocking_is_gp(void)
|
|
|
|
|
{
|
2022-12-15 11:57:55 +08:00
|
|
|
|
if (rcu_scheduler_active != RCU_SCHEDULER_INACTIVE) {
|
|
|
|
|
might_sleep();
|
2022-08-05 07:07:04 +08:00
|
|
|
|
return false;
|
2022-12-15 11:57:55 +08:00
|
|
|
|
}
|
2022-08-05 07:07:04 +08:00
|
|
|
|
return true;
|
2018-11-30 02:42:06 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* synchronize_rcu - wait until a grace period has elapsed.
|
|
|
|
|
*
|
|
|
|
|
* Control will return to the caller some time after a full grace
|
|
|
|
|
* period has elapsed, in other words after all currently executing RCU
|
|
|
|
|
* read-side critical sections have completed. Note, however, that
|
|
|
|
|
* upon return from synchronize_rcu(), the caller might well be executing
|
|
|
|
|
* concurrently with new RCU read-side critical sections that began while
|
2021-04-30 02:18:01 +08:00
|
|
|
|
* synchronize_rcu() was waiting.
|
|
|
|
|
*
|
|
|
|
|
* RCU read-side critical sections are delimited by rcu_read_lock()
|
|
|
|
|
* and rcu_read_unlock(), and may be nested. In addition, but only in
|
|
|
|
|
* v5.0 and later, regions of code across which interrupts, preemption,
|
|
|
|
|
* or softirqs have been disabled also serve as RCU read-side critical
|
2018-11-30 02:42:06 +08:00
|
|
|
|
* sections. This includes hardware interrupt handlers, softirq handlers,
|
|
|
|
|
* and NMI handlers.
|
|
|
|
|
*
|
|
|
|
|
* Note that this guarantee implies further memory-ordering guarantees.
|
|
|
|
|
* On systems with more than one CPU, when synchronize_rcu() returns,
|
|
|
|
|
* each CPU is guaranteed to have executed a full memory barrier since
|
|
|
|
|
* the end of its last RCU read-side critical section whose beginning
|
|
|
|
|
* preceded the call to synchronize_rcu(). In addition, each CPU having
|
|
|
|
|
* an RCU read-side critical section that extends beyond the return from
|
|
|
|
|
* synchronize_rcu() is guaranteed to have executed a full memory barrier
|
|
|
|
|
* after the beginning of synchronize_rcu() and before the beginning of
|
|
|
|
|
* that RCU read-side critical section. Note that these guarantees include
|
|
|
|
|
* CPUs that are offline, idle, or executing in user mode, as well as CPUs
|
|
|
|
|
* that are executing in the kernel.
|
|
|
|
|
*
|
|
|
|
|
* Furthermore, if CPU A invoked synchronize_rcu(), which returned
|
|
|
|
|
* to its caller on CPU B, then both CPU A and CPU B are guaranteed
|
|
|
|
|
* to have executed a full memory barrier during the execution of
|
|
|
|
|
* synchronize_rcu() -- even if CPU A and CPU B are the same CPU (but
|
|
|
|
|
* again only if the system has more than one CPU).
|
2021-04-17 07:53:16 +08:00
|
|
|
|
*
|
|
|
|
|
* Implementation of these memory-ordering guarantees is described here:
|
|
|
|
|
* Documentation/RCU/Design/Memory-Ordering/Tree-RCU-Memory-Ordering.rst.
|
2018-11-30 02:42:06 +08:00
|
|
|
|
*/
|
|
|
|
|
void synchronize_rcu(void)
|
|
|
|
|
{
|
2022-08-05 08:28:01 +08:00
|
|
|
|
unsigned long flags;
|
|
|
|
|
struct rcu_node *rnp;
|
|
|
|
|
|
2018-11-30 02:42:06 +08:00
|
|
|
|
RCU_LOCKDEP_WARN(lock_is_held(&rcu_bh_lock_map) ||
|
|
|
|
|
lock_is_held(&rcu_lock_map) ||
|
|
|
|
|
lock_is_held(&rcu_sched_lock_map),
|
|
|
|
|
"Illegal synchronize_rcu() in RCU read-side critical section");
|
2022-08-05 08:28:01 +08:00
|
|
|
|
if (!rcu_blocking_is_gp()) {
|
|
|
|
|
if (rcu_gp_is_expedited())
|
|
|
|
|
synchronize_rcu_expedited();
|
|
|
|
|
else
|
2022-10-17 00:22:54 +08:00
|
|
|
|
wait_rcu_gp(call_rcu_hurry);
|
2022-08-05 08:28:01 +08:00
|
|
|
|
return;
|
2022-04-14 08:46:15 +08:00
|
|
|
|
}
|
2022-08-05 08:28:01 +08:00
|
|
|
|
|
|
|
|
|
// Context allows vacuous grace periods.
|
|
|
|
|
// Note well that this code runs with !PREEMPT && !SMP.
|
|
|
|
|
// In addition, all code that advances grace periods runs at
|
|
|
|
|
// process level. Therefore, this normal GP overlaps with other
|
|
|
|
|
// normal GPs only by being fully nested within them, which allows
|
|
|
|
|
// reuse of ->gp_seq_polled_snap.
|
|
|
|
|
rcu_poll_gp_seq_start_unlocked(&rcu_state.gp_seq_polled_snap);
|
|
|
|
|
rcu_poll_gp_seq_end_unlocked(&rcu_state.gp_seq_polled_snap);
|
|
|
|
|
|
2022-08-06 06:42:25 +08:00
|
|
|
|
// Update the normal grace-period counters to record
|
|
|
|
|
// this grace period, but only those used by the boot CPU.
|
|
|
|
|
// The rcu_scheduler_starting() will take care of the rest of
|
|
|
|
|
// these counters.
|
2022-08-05 08:28:01 +08:00
|
|
|
|
local_irq_save(flags);
|
|
|
|
|
WARN_ON_ONCE(num_online_cpus() > 1);
|
|
|
|
|
rcu_state.gp_seq += (1 << RCU_SEQ_CTR_SHIFT);
|
2022-08-06 06:42:25 +08:00
|
|
|
|
for (rnp = this_cpu_ptr(&rcu_data)->mynode; rnp; rnp = rnp->parent)
|
2022-08-05 08:28:01 +08:00
|
|
|
|
rnp->gp_seq_needed = rnp->gp_seq = rcu_state.gp_seq;
|
|
|
|
|
local_irq_restore(flags);
|
2018-11-30 02:42:06 +08:00
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(synchronize_rcu);
|
|
|
|
|
|
2022-07-29 06:37:05 +08:00
|
|
|
|
/**
|
|
|
|
|
* get_completed_synchronize_rcu_full - Return a full pre-completed polled state cookie
|
|
|
|
|
* @rgosp: Place to put state cookie
|
|
|
|
|
*
|
|
|
|
|
* Stores into @rgosp a value that will always be treated by functions
|
|
|
|
|
* like poll_state_synchronize_rcu_full() as a cookie whose grace period
|
|
|
|
|
* has already completed.
|
|
|
|
|
*/
|
|
|
|
|
void get_completed_synchronize_rcu_full(struct rcu_gp_oldstate *rgosp)
|
|
|
|
|
{
|
|
|
|
|
rgosp->rgos_norm = RCU_GET_STATE_COMPLETED;
|
|
|
|
|
rgosp->rgos_exp = RCU_GET_STATE_COMPLETED;
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(get_completed_synchronize_rcu_full);
|
|
|
|
|
|
2014-03-15 07:37:08 +08:00
|
|
|
|
/**
|
|
|
|
|
* get_state_synchronize_rcu - Snapshot current RCU state
|
|
|
|
|
*
|
|
|
|
|
* Returns a cookie that is used by a later call to cond_synchronize_rcu()
|
2021-02-26 08:10:38 +08:00
|
|
|
|
* or poll_state_synchronize_rcu() to determine whether or not a full
|
|
|
|
|
* grace period has elapsed in the meantime.
|
2014-03-15 07:37:08 +08:00
|
|
|
|
*/
|
|
|
|
|
unsigned long get_state_synchronize_rcu(void)
|
|
|
|
|
{
|
|
|
|
|
/*
|
|
|
|
|
* Any prior manipulation of RCU-protected data must happen
|
2018-04-28 06:16:50 +08:00
|
|
|
|
* before the load from ->gp_seq.
|
2014-03-15 07:37:08 +08:00
|
|
|
|
*/
|
|
|
|
|
smp_mb(); /* ^^^ */
|
2022-04-14 08:46:15 +08:00
|
|
|
|
return rcu_seq_snap(&rcu_state.gp_seq_polled);
|
2014-03-15 07:37:08 +08:00
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(get_state_synchronize_rcu);
|
|
|
|
|
|
2021-02-26 08:10:38 +08:00
|
|
|
|
/**
|
2022-07-29 10:58:13 +08:00
|
|
|
|
* get_state_synchronize_rcu_full - Snapshot RCU state, both normal and expedited
|
|
|
|
|
* @rgosp: location to place combined normal/expedited grace-period state
|
2021-02-26 08:10:38 +08:00
|
|
|
|
*
|
2022-07-29 10:58:13 +08:00
|
|
|
|
* Places the normal and expedited grace-period states in @rgosp. This
|
|
|
|
|
* state value can be passed to a later call to cond_synchronize_rcu_full()
|
|
|
|
|
* or poll_state_synchronize_rcu_full() to determine whether or not a
|
|
|
|
|
* grace period (whether normal or expedited) has elapsed in the meantime.
|
|
|
|
|
* The rcu_gp_oldstate structure takes up twice the memory of an unsigned
|
|
|
|
|
* long, but is guaranteed to see all grace periods. In contrast, the
|
|
|
|
|
* combined state occupies less memory, but can sometimes fail to take
|
|
|
|
|
* grace periods into account.
|
2021-02-26 08:10:38 +08:00
|
|
|
|
*
|
2022-07-29 10:58:13 +08:00
|
|
|
|
* This does not guarantee that the needed grace period will actually
|
|
|
|
|
* start.
|
2021-02-26 08:10:38 +08:00
|
|
|
|
*/
|
2022-07-29 10:58:13 +08:00
|
|
|
|
void get_state_synchronize_rcu_full(struct rcu_gp_oldstate *rgosp)
|
|
|
|
|
{
|
|
|
|
|
struct rcu_node *rnp = rcu_get_root();
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Any prior manipulation of RCU-protected data must happen
|
|
|
|
|
* before the loads from ->gp_seq and ->expedited_sequence.
|
|
|
|
|
*/
|
|
|
|
|
smp_mb(); /* ^^^ */
|
|
|
|
|
rgosp->rgos_norm = rcu_seq_snap(&rnp->gp_seq);
|
|
|
|
|
rgosp->rgos_exp = rcu_seq_snap(&rcu_state.expedited_sequence);
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(get_state_synchronize_rcu_full);
|
|
|
|
|
|
2022-08-03 08:04:54 +08:00
|
|
|
|
/*
|
|
|
|
|
* Helper function for start_poll_synchronize_rcu() and
|
|
|
|
|
* start_poll_synchronize_rcu_full().
|
2021-02-26 08:10:38 +08:00
|
|
|
|
*/
|
2022-08-03 08:04:54 +08:00
|
|
|
|
static void start_poll_synchronize_rcu_common(void)
|
2021-02-26 08:10:38 +08:00
|
|
|
|
{
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
bool needwake;
|
|
|
|
|
struct rcu_data *rdp;
|
|
|
|
|
struct rcu_node *rnp;
|
|
|
|
|
|
|
|
|
|
lockdep_assert_irqs_enabled();
|
|
|
|
|
local_irq_save(flags);
|
|
|
|
|
rdp = this_cpu_ptr(&rcu_data);
|
|
|
|
|
rnp = rdp->mynode;
|
|
|
|
|
raw_spin_lock_rcu_node(rnp); // irqs already disabled.
|
2022-04-14 08:46:15 +08:00
|
|
|
|
// Note it is possible for a grace period to have elapsed between
|
|
|
|
|
// the above call to get_state_synchronize_rcu() and the below call
|
|
|
|
|
// to rcu_seq_snap. This is OK, the worst that happens is that we
|
|
|
|
|
// get a grace period that no one needed. These accesses are ordered
|
|
|
|
|
// by smp_mb(), and we are accessing them in the opposite order
|
|
|
|
|
// from which they are updated at grace-period start, as required.
|
|
|
|
|
needwake = rcu_start_this_gp(rnp, rdp, rcu_seq_snap(&rcu_state.gp_seq));
|
2021-02-26 08:10:38 +08:00
|
|
|
|
raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
|
|
|
|
|
if (needwake)
|
|
|
|
|
rcu_gp_kthread_wake();
|
2022-08-03 08:04:54 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* start_poll_synchronize_rcu - Snapshot and start RCU grace period
|
|
|
|
|
*
|
|
|
|
|
* Returns a cookie that is used by a later call to cond_synchronize_rcu()
|
|
|
|
|
* or poll_state_synchronize_rcu() to determine whether or not a full
|
|
|
|
|
* grace period has elapsed in the meantime. If the needed grace period
|
|
|
|
|
* is not already slated to start, notifies RCU core of the need for that
|
|
|
|
|
* grace period.
|
|
|
|
|
*
|
|
|
|
|
* Interrupts must be enabled for the case where it is necessary to awaken
|
|
|
|
|
* the grace-period kthread.
|
|
|
|
|
*/
|
|
|
|
|
unsigned long start_poll_synchronize_rcu(void)
|
|
|
|
|
{
|
|
|
|
|
unsigned long gp_seq = get_state_synchronize_rcu();
|
|
|
|
|
|
|
|
|
|
start_poll_synchronize_rcu_common();
|
2021-02-26 08:10:38 +08:00
|
|
|
|
return gp_seq;
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(start_poll_synchronize_rcu);
|
|
|
|
|
|
|
|
|
|
/**
|
2022-08-03 08:04:54 +08:00
|
|
|
|
* start_poll_synchronize_rcu_full - Take a full snapshot and start RCU grace period
|
|
|
|
|
* @rgosp: value from get_state_synchronize_rcu_full() or start_poll_synchronize_rcu_full()
|
2021-02-26 08:10:38 +08:00
|
|
|
|
*
|
2022-08-03 08:04:54 +08:00
|
|
|
|
* Places the normal and expedited grace-period states in *@rgos. This
|
|
|
|
|
* state value can be passed to a later call to cond_synchronize_rcu_full()
|
|
|
|
|
* or poll_state_synchronize_rcu_full() to determine whether or not a
|
|
|
|
|
* grace period (whether normal or expedited) has elapsed in the meantime.
|
|
|
|
|
* If the needed grace period is not already slated to start, notifies
|
|
|
|
|
* RCU core of the need for that grace period.
|
|
|
|
|
*
|
|
|
|
|
* Interrupts must be enabled for the case where it is necessary to awaken
|
|
|
|
|
* the grace-period kthread.
|
|
|
|
|
*/
|
|
|
|
|
void start_poll_synchronize_rcu_full(struct rcu_gp_oldstate *rgosp)
|
|
|
|
|
{
|
|
|
|
|
get_state_synchronize_rcu_full(rgosp);
|
|
|
|
|
|
|
|
|
|
start_poll_synchronize_rcu_common();
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(start_poll_synchronize_rcu_full);
|
|
|
|
|
|
2021-02-26 08:10:38 +08:00
|
|
|
|
/**
|
2022-07-29 06:37:05 +08:00
|
|
|
|
* poll_state_synchronize_rcu - Has the specified RCU grace period completed?
|
2021-04-17 07:53:16 +08:00
|
|
|
|
* @oldstate: value from get_state_synchronize_rcu() or start_poll_synchronize_rcu()
|
2021-02-26 08:10:38 +08:00
|
|
|
|
*
|
|
|
|
|
* If a full RCU grace period has elapsed since the earlier call from
|
2022-08-04 07:57:47 +08:00
|
|
|
|
* which @oldstate was obtained, return @true, otherwise return @false.
|
2021-03-23 13:29:10 +08:00
|
|
|
|
* If @false is returned, it is the caller's responsibility to invoke this
|
2021-02-26 08:10:38 +08:00
|
|
|
|
* function later on until it does return @true. Alternatively, the caller
|
|
|
|
|
* can explicitly wait for a grace period, for example, by passing @oldstate
|
2022-11-26 00:43:10 +08:00
|
|
|
|
* to either cond_synchronize_rcu() or cond_synchronize_rcu_expedited()
|
|
|
|
|
* on the one hand or by directly invoking either synchronize_rcu() or
|
|
|
|
|
* synchronize_rcu_expedited() on the other.
|
2021-02-26 08:10:38 +08:00
|
|
|
|
*
|
|
|
|
|
* Yes, this function does not take counter wrap into account.
|
|
|
|
|
* But counter wrap is harmless. If the counter wraps, we have waited for
|
2022-03-22 09:41:46 +08:00
|
|
|
|
* more than a billion grace periods (and way more on a 64-bit system!).
|
2022-08-04 07:57:47 +08:00
|
|
|
|
* Those needing to keep old state values for very long time periods
|
2022-07-29 06:37:05 +08:00
|
|
|
|
* (many hours even on 32-bit systems) should check them occasionally and
|
|
|
|
|
* either refresh them or set a flag indicating that the grace period has
|
|
|
|
|
* completed. Alternatively, they can use get_completed_synchronize_rcu()
|
|
|
|
|
* to get a guaranteed-completed grace-period state.
|
2021-04-17 07:53:16 +08:00
|
|
|
|
*
|
2022-11-26 00:43:10 +08:00
|
|
|
|
* In addition, because oldstate compresses the grace-period state for
|
|
|
|
|
* both normal and expedited grace periods into a single unsigned long,
|
|
|
|
|
* it can miss a grace period when synchronize_rcu() runs concurrently
|
|
|
|
|
* with synchronize_rcu_expedited(). If this is unacceptable, please
|
|
|
|
|
* instead use the _full() variant of these polling APIs.
|
|
|
|
|
*
|
2021-04-17 07:53:16 +08:00
|
|
|
|
* This function provides the same memory-ordering guarantees that
|
|
|
|
|
* would be provided by a synchronize_rcu() that was invoked at the call
|
|
|
|
|
* to the function that provided @oldstate, and that returned at the end
|
|
|
|
|
* of this function.
|
2021-02-26 08:10:38 +08:00
|
|
|
|
*/
|
|
|
|
|
bool poll_state_synchronize_rcu(unsigned long oldstate)
|
|
|
|
|
{
|
2022-04-14 06:17:25 +08:00
|
|
|
|
if (oldstate == RCU_GET_STATE_COMPLETED ||
|
2022-04-14 08:46:15 +08:00
|
|
|
|
rcu_seq_done_exact(&rcu_state.gp_seq_polled, oldstate)) {
|
2021-02-26 08:10:38 +08:00
|
|
|
|
smp_mb(); /* Ensure GP ends before subsequent accesses. */
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(poll_state_synchronize_rcu);
|
|
|
|
|
|
2014-03-15 07:37:08 +08:00
|
|
|
|
/**
|
2022-07-29 06:37:05 +08:00
|
|
|
|
* poll_state_synchronize_rcu_full - Has the specified RCU grace period completed?
|
|
|
|
|
* @rgosp: value from get_state_synchronize_rcu_full() or start_poll_synchronize_rcu_full()
|
2014-03-15 07:37:08 +08:00
|
|
|
|
*
|
2022-07-29 06:37:05 +08:00
|
|
|
|
* If a full RCU grace period has elapsed since the earlier call from
|
|
|
|
|
* which *rgosp was obtained, return @true, otherwise return @false.
|
|
|
|
|
* If @false is returned, it is the caller's responsibility to invoke this
|
|
|
|
|
* function later on until it does return @true. Alternatively, the caller
|
|
|
|
|
* can explicitly wait for a grace period, for example, by passing @rgosp
|
|
|
|
|
* to cond_synchronize_rcu() or by directly invoking synchronize_rcu().
|
|
|
|
|
*
|
|
|
|
|
* Yes, this function does not take counter wrap into account.
|
|
|
|
|
* But counter wrap is harmless. If the counter wraps, we have waited
|
|
|
|
|
* for more than a billion grace periods (and way more on a 64-bit
|
|
|
|
|
* system!). Those needing to keep rcu_gp_oldstate values for very
|
|
|
|
|
* long time periods (many hours even on 32-bit systems) should check
|
|
|
|
|
* them occasionally and either refresh them or set a flag indicating
|
|
|
|
|
* that the grace period has completed. Alternatively, they can use
|
|
|
|
|
* get_completed_synchronize_rcu_full() to get a guaranteed-completed
|
|
|
|
|
* grace-period state.
|
2014-03-15 07:37:08 +08:00
|
|
|
|
*
|
2022-07-29 06:37:05 +08:00
|
|
|
|
* This function provides the same memory-ordering guarantees that would
|
|
|
|
|
* be provided by a synchronize_rcu() that was invoked at the call to
|
|
|
|
|
* the function that provided @rgosp, and that returned at the end of this
|
|
|
|
|
* function. And this guarantee requires that the root rcu_node structure's
|
|
|
|
|
* ->gp_seq field be checked instead of that of the rcu_state structure.
|
|
|
|
|
* The problem is that the just-ending grace-period's callbacks can be
|
|
|
|
|
* invoked between the time that the root rcu_node structure's ->gp_seq
|
|
|
|
|
* field is updated and the time that the rcu_state structure's ->gp_seq
|
|
|
|
|
* field is updated. Therefore, if a single synchronize_rcu() is to
|
|
|
|
|
* cause a subsequent poll_state_synchronize_rcu_full() to return @true,
|
|
|
|
|
* then the root rcu_node structure is the one that needs to be polled.
|
|
|
|
|
*/
|
|
|
|
|
bool poll_state_synchronize_rcu_full(struct rcu_gp_oldstate *rgosp)
|
|
|
|
|
{
|
|
|
|
|
struct rcu_node *rnp = rcu_get_root();
|
|
|
|
|
|
|
|
|
|
smp_mb(); // Order against root rcu_node structure grace-period cleanup.
|
|
|
|
|
if (rgosp->rgos_norm == RCU_GET_STATE_COMPLETED ||
|
|
|
|
|
rcu_seq_done_exact(&rnp->gp_seq, rgosp->rgos_norm) ||
|
|
|
|
|
rgosp->rgos_exp == RCU_GET_STATE_COMPLETED ||
|
2022-08-05 08:54:53 +08:00
|
|
|
|
rcu_seq_done_exact(&rcu_state.expedited_sequence, rgosp->rgos_exp)) {
|
2022-07-29 06:37:05 +08:00
|
|
|
|
smp_mb(); /* Ensure GP ends before subsequent accesses. */
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(poll_state_synchronize_rcu_full);
|
|
|
|
|
|
2014-03-15 07:37:08 +08:00
|
|
|
|
/**
|
|
|
|
|
* cond_synchronize_rcu - Conditionally wait for an RCU grace period
|
2022-04-16 01:55:42 +08:00
|
|
|
|
* @oldstate: value from get_state_synchronize_rcu(), start_poll_synchronize_rcu(), or start_poll_synchronize_rcu_expedited()
|
2014-03-15 07:37:08 +08:00
|
|
|
|
*
|
|
|
|
|
* If a full RCU grace period has elapsed since the earlier call to
|
2021-02-26 08:10:38 +08:00
|
|
|
|
* get_state_synchronize_rcu() or start_poll_synchronize_rcu(), just return.
|
|
|
|
|
* Otherwise, invoke synchronize_rcu() to wait for a full grace period.
|
2014-03-15 07:37:08 +08:00
|
|
|
|
*
|
2022-04-16 01:55:42 +08:00
|
|
|
|
* Yes, this function does not take counter wrap into account.
|
|
|
|
|
* But counter wrap is harmless. If the counter wraps, we have waited for
|
2014-03-15 07:37:08 +08:00
|
|
|
|
* more than 2 billion grace periods (and way more on a 64-bit system!),
|
2022-04-16 01:55:42 +08:00
|
|
|
|
* so waiting for a couple of additional grace periods should be just fine.
|
2021-04-17 07:53:16 +08:00
|
|
|
|
*
|
|
|
|
|
* This function provides the same memory-ordering guarantees that
|
|
|
|
|
* would be provided by a synchronize_rcu() that was invoked at the call
|
2022-04-16 01:55:42 +08:00
|
|
|
|
* to the function that provided @oldstate and that returned at the end
|
2021-04-17 07:53:16 +08:00
|
|
|
|
* of this function.
|
2014-03-15 07:37:08 +08:00
|
|
|
|
*/
|
|
|
|
|
void cond_synchronize_rcu(unsigned long oldstate)
|
|
|
|
|
{
|
2021-02-26 08:10:38 +08:00
|
|
|
|
if (!poll_state_synchronize_rcu(oldstate))
|
2014-03-15 07:37:08 +08:00
|
|
|
|
synchronize_rcu();
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(cond_synchronize_rcu);
|
|
|
|
|
|
2022-08-05 04:46:05 +08:00
|
|
|
|
/**
|
|
|
|
|
* cond_synchronize_rcu_full - Conditionally wait for an RCU grace period
|
|
|
|
|
* @rgosp: value from get_state_synchronize_rcu_full(), start_poll_synchronize_rcu_full(), or start_poll_synchronize_rcu_expedited_full()
|
|
|
|
|
*
|
|
|
|
|
* If a full RCU grace period has elapsed since the call to
|
|
|
|
|
* get_state_synchronize_rcu_full(), start_poll_synchronize_rcu_full(),
|
|
|
|
|
* or start_poll_synchronize_rcu_expedited_full() from which @rgosp was
|
|
|
|
|
* obtained, just return. Otherwise, invoke synchronize_rcu() to wait
|
|
|
|
|
* for a full grace period.
|
|
|
|
|
*
|
|
|
|
|
* Yes, this function does not take counter wrap into account.
|
|
|
|
|
* But counter wrap is harmless. If the counter wraps, we have waited for
|
|
|
|
|
* more than 2 billion grace periods (and way more on a 64-bit system!),
|
|
|
|
|
* so waiting for a couple of additional grace periods should be just fine.
|
|
|
|
|
*
|
|
|
|
|
* This function provides the same memory-ordering guarantees that
|
|
|
|
|
* would be provided by a synchronize_rcu() that was invoked at the call
|
|
|
|
|
* to the function that provided @rgosp and that returned at the end of
|
|
|
|
|
* this function.
|
|
|
|
|
*/
|
|
|
|
|
void cond_synchronize_rcu_full(struct rcu_gp_oldstate *rgosp)
|
|
|
|
|
{
|
|
|
|
|
if (!poll_state_synchronize_rcu_full(rgosp))
|
|
|
|
|
synchronize_rcu();
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(cond_synchronize_rcu_full);
|
|
|
|
|
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
/*
|
2018-07-04 08:22:34 +08:00
|
|
|
|
* Check to see if there is any immediate RCU-related work to be done by
|
2018-07-08 09:12:26 +08:00
|
|
|
|
* the current CPU, returning 1 if so and zero otherwise. The checks are
|
|
|
|
|
* in order of increasing expense: checks that can be carried out against
|
|
|
|
|
* CPU-local state are performed first. However, we must check for CPU
|
|
|
|
|
* stalls first, else we might not get a chance.
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
*/
|
2019-09-14 18:39:22 +08:00
|
|
|
|
static int rcu_pending(int user)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
{
|
2019-09-14 05:09:56 +08:00
|
|
|
|
bool gp_in_progress;
|
2018-07-04 08:22:34 +08:00
|
|
|
|
struct rcu_data *rdp = this_cpu_ptr(&rcu_data);
|
2009-11-14 11:51:39 +08:00
|
|
|
|
struct rcu_node *rnp = rdp->mynode;
|
|
|
|
|
|
2020-11-20 02:13:06 +08:00
|
|
|
|
lockdep_assert_irqs_disabled();
|
|
|
|
|
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
/* Check for CPU stalls, if enabled. */
|
2018-07-04 08:22:34 +08:00
|
|
|
|
check_cpu_stall(rdp);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
2019-04-17 05:48:28 +08:00
|
|
|
|
/* Does this CPU need a deferred NOCB wakeup? */
|
2021-02-23 08:10:10 +08:00
|
|
|
|
if (rcu_nocb_need_deferred_wakeup(rdp, RCU_NOCB_WAKE))
|
2019-04-17 05:48:28 +08:00
|
|
|
|
return 1;
|
|
|
|
|
|
2019-09-14 18:39:22 +08:00
|
|
|
|
/* Is this a nohz_full CPU in userspace or idle? (Ignore RCU if so.) */
|
|
|
|
|
if ((user || rcu_is_cpu_rrupt_from_idle()) && rcu_nohz_full_cpu())
|
2013-11-09 01:03:10 +08:00
|
|
|
|
return 0;
|
|
|
|
|
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
/* Is the RCU core waiting for a quiescent state from this CPU? */
|
2019-09-14 05:09:56 +08:00
|
|
|
|
gp_in_progress = rcu_gp_in_progress();
|
|
|
|
|
if (rdp->core_needs_qs && !rdp->cpu_no_qs.b.norm && gp_in_progress)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
return 1;
|
|
|
|
|
|
|
|
|
|
/* Does this CPU have callbacks ready to invoke? */
|
2020-11-12 08:51:21 +08:00
|
|
|
|
if (!rcu_rdp_is_offloaded(rdp) &&
|
rcu/tree: nocb: Avoid raising softirq for offloaded ready-to-execute CBs
Testing showed that rcu_pending() can return 1 when offloaded callbacks
are ready to execute. This invokes RCU core processing, for example,
by raising RCU_SOFTIRQ, eventually resulting in a call to rcu_core().
However, rcu_core() explicitly avoids in any way manipulating offloaded
callbacks, which are instead handled by the rcuog and rcuoc kthreads,
which work independently of rcu_core().
One exception to this independence is that rcu_core() invokes
do_nocb_deferred_wakeup(), however, rcu_pending() also checks
rcu_nocb_need_deferred_wakeup() in order to correctly handle this case,
invoking rcu_core() when needed.
This commit therefore avoids needlessly invoking RCU core processing
by checking rcu_segcblist_ready_cbs() only on non-offloaded CPUs.
This reduces overhead, for example, by reducing softirq activity.
This change passed 30 minute tests of TREE01 through TREE09 each.
On TREE08, there is at most 150us from the time that rcu_pending() chose
not to invoke RCU core processing to the time when the ready callbacks
were invoked by the rcuoc kthread. This provides further evidence that
there is no need to invoke rcu_core() for offloaded callbacks that are
ready to invoke.
Cc: Neeraj Upadhyay <neeraju@codeaurora.org>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Reviewed-by: Neeraj Upadhyay <neeraju@codeaurora.org>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-10-08 04:50:36 +08:00
|
|
|
|
rcu_segcblist_ready_cbs(&rdp->cblist))
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
return 1;
|
|
|
|
|
|
|
|
|
|
/* Has RCU gone idle with this CPU needing another grace period? */
|
2019-09-14 05:09:56 +08:00
|
|
|
|
if (!gp_in_progress && rcu_segcblist_is_enabled(&rdp->cblist) &&
|
2020-11-12 08:51:21 +08:00
|
|
|
|
!rcu_rdp_is_offloaded(rdp) &&
|
2018-04-13 07:29:13 +08:00
|
|
|
|
!rcu_segcblist_restempty(&rdp->cblist, RCU_NEXT_READY_TAIL))
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
return 1;
|
|
|
|
|
|
2018-04-28 07:01:46 +08:00
|
|
|
|
/* Have RCU grace period completed or started? */
|
|
|
|
|
if (rcu_seq_current(&rnp->gp_seq) != rdp->gp_seq ||
|
2018-01-11 04:36:00 +08:00
|
|
|
|
unlikely(READ_ONCE(rdp->gpwrap))) /* outside lock */
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
return 1;
|
|
|
|
|
|
|
|
|
|
/* nothing to do */
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2012-05-24 09:47:05 +08:00
|
|
|
|
/*
|
2018-07-11 09:37:30 +08:00
|
|
|
|
* Helper function for rcu_barrier() tracing. If tracing is disabled,
|
2012-05-24 09:47:05 +08:00
|
|
|
|
* the compiler is expected to optimize this away.
|
|
|
|
|
*/
|
2018-07-11 09:37:30 +08:00
|
|
|
|
static void rcu_barrier_trace(const char *s, int cpu, unsigned long done)
|
2012-05-24 09:47:05 +08:00
|
|
|
|
{
|
2018-07-04 08:22:34 +08:00
|
|
|
|
trace_rcu_barrier(rcu_state.name, s, cpu,
|
|
|
|
|
atomic_read(&rcu_state.barrier_cpu_count), done);
|
2012-05-24 09:47:05 +08:00
|
|
|
|
}
|
|
|
|
|
|
2012-03-02 05:18:08 +08:00
|
|
|
|
/*
|
2018-07-11 09:37:30 +08:00
|
|
|
|
* RCU callback function for rcu_barrier(). If we are last, wake
|
|
|
|
|
* up the task executing rcu_barrier().
|
rcu: Fix rcu_barrier_callback() race condition
The rcu_barrier_callback() function does an atomic_dec_and_test(), and
if it is the last CPU to check in, does the required wakeup. Either way,
it does an event trace. Unfortunately, this is susceptible to the
following sequence of events:
o CPU 0 invokes rcu_barrier_callback(), but atomic_dec_and_test()
says that it is not last. But at this point, CPU 0 is delayed,
perhaps due to an NMI, SMI, or vCPU preemption.
o CPU 1 invokes rcu_barrier_callback(), and atomic_dec_and_test()
says that it is last. So CPU 1 traces completion and does
the needed wakeup.
o The awakened rcu_barrier() function does cleanup and releases
rcu_state.barrier_mutex.
o Another CPU now acquires rcu_state.barrier_mutex and starts
another round of rcu_barrier() processing, including updating
rcu_state.barrier_sequence.
o CPU 0 gets its act back together and does its tracing. Except
that rcu_state.barrier_sequence has already been updated, so
its tracing is incorrect and probably quite confusing.
(Wait! Why did this CPU check in twice for one rcu_barrier()
invocation???)
This commit therefore causes rcu_barrier_callback() to take a
snapshot of the value of rcu_state.barrier_sequence before invoking
atomic_dec_and_test(), thus guaranteeing that the event-trace output
is sensible, even if the timing of the event-trace output might still
be confusing. (Wait! Why did the old rcu_barrier() complete before
all of its CPUs checked in???) But being that this is RCU, only so much
confusion can reasonably be eliminated.
This data race was reported by KCSAN. Not appropriate for backporting
due to failure being unlikely and due to the mild consequences of the
failure, namely a confusing event trace.
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-21 07:43:45 +08:00
|
|
|
|
*
|
|
|
|
|
* Note that the value of rcu_state.barrier_sequence must be captured
|
|
|
|
|
* before the atomic_dec_and_test(). Otherwise, if this CPU is not last,
|
|
|
|
|
* other CPUs might count the value down to zero before this CPU gets
|
|
|
|
|
* around to invoking rcu_barrier_trace(), which might result in bogus
|
|
|
|
|
* data from the next instance of rcu_barrier().
|
2012-03-02 05:18:08 +08:00
|
|
|
|
*/
|
2012-05-29 15:34:56 +08:00
|
|
|
|
static void rcu_barrier_callback(struct rcu_head *rhp)
|
2009-10-07 12:48:16 +08:00
|
|
|
|
{
|
rcu: Fix rcu_barrier_callback() race condition
The rcu_barrier_callback() function does an atomic_dec_and_test(), and
if it is the last CPU to check in, does the required wakeup. Either way,
it does an event trace. Unfortunately, this is susceptible to the
following sequence of events:
o CPU 0 invokes rcu_barrier_callback(), but atomic_dec_and_test()
says that it is not last. But at this point, CPU 0 is delayed,
perhaps due to an NMI, SMI, or vCPU preemption.
o CPU 1 invokes rcu_barrier_callback(), and atomic_dec_and_test()
says that it is last. So CPU 1 traces completion and does
the needed wakeup.
o The awakened rcu_barrier() function does cleanup and releases
rcu_state.barrier_mutex.
o Another CPU now acquires rcu_state.barrier_mutex and starts
another round of rcu_barrier() processing, including updating
rcu_state.barrier_sequence.
o CPU 0 gets its act back together and does its tracing. Except
that rcu_state.barrier_sequence has already been updated, so
its tracing is incorrect and probably quite confusing.
(Wait! Why did this CPU check in twice for one rcu_barrier()
invocation???)
This commit therefore causes rcu_barrier_callback() to take a
snapshot of the value of rcu_state.barrier_sequence before invoking
atomic_dec_and_test(), thus guaranteeing that the event-trace output
is sensible, even if the timing of the event-trace output might still
be confusing. (Wait! Why did the old rcu_barrier() complete before
all of its CPUs checked in???) But being that this is RCU, only so much
confusion can reasonably be eliminated.
This data race was reported by KCSAN. Not appropriate for backporting
due to failure being unlikely and due to the mild consequences of the
failure, namely a confusing event trace.
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-21 07:43:45 +08:00
|
|
|
|
unsigned long __maybe_unused s = rcu_state.barrier_sequence;
|
|
|
|
|
|
2018-07-06 07:26:12 +08:00
|
|
|
|
if (atomic_dec_and_test(&rcu_state.barrier_cpu_count)) {
|
rcu: Fix rcu_barrier_callback() race condition
The rcu_barrier_callback() function does an atomic_dec_and_test(), and
if it is the last CPU to check in, does the required wakeup. Either way,
it does an event trace. Unfortunately, this is susceptible to the
following sequence of events:
o CPU 0 invokes rcu_barrier_callback(), but atomic_dec_and_test()
says that it is not last. But at this point, CPU 0 is delayed,
perhaps due to an NMI, SMI, or vCPU preemption.
o CPU 1 invokes rcu_barrier_callback(), and atomic_dec_and_test()
says that it is last. So CPU 1 traces completion and does
the needed wakeup.
o The awakened rcu_barrier() function does cleanup and releases
rcu_state.barrier_mutex.
o Another CPU now acquires rcu_state.barrier_mutex and starts
another round of rcu_barrier() processing, including updating
rcu_state.barrier_sequence.
o CPU 0 gets its act back together and does its tracing. Except
that rcu_state.barrier_sequence has already been updated, so
its tracing is incorrect and probably quite confusing.
(Wait! Why did this CPU check in twice for one rcu_barrier()
invocation???)
This commit therefore causes rcu_barrier_callback() to take a
snapshot of the value of rcu_state.barrier_sequence before invoking
atomic_dec_and_test(), thus guaranteeing that the event-trace output
is sensible, even if the timing of the event-trace output might still
be confusing. (Wait! Why did the old rcu_barrier() complete before
all of its CPUs checked in???) But being that this is RCU, only so much
confusion can reasonably be eliminated.
This data race was reported by KCSAN. Not appropriate for backporting
due to failure being unlikely and due to the mild consequences of the
failure, namely a confusing event trace.
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-21 07:43:45 +08:00
|
|
|
|
rcu_barrier_trace(TPS("LastCB"), -1, s);
|
2018-07-06 07:26:12 +08:00
|
|
|
|
complete(&rcu_state.barrier_completion);
|
2012-05-24 09:47:05 +08:00
|
|
|
|
} else {
|
rcu: Fix rcu_barrier_callback() race condition
The rcu_barrier_callback() function does an atomic_dec_and_test(), and
if it is the last CPU to check in, does the required wakeup. Either way,
it does an event trace. Unfortunately, this is susceptible to the
following sequence of events:
o CPU 0 invokes rcu_barrier_callback(), but atomic_dec_and_test()
says that it is not last. But at this point, CPU 0 is delayed,
perhaps due to an NMI, SMI, or vCPU preemption.
o CPU 1 invokes rcu_barrier_callback(), and atomic_dec_and_test()
says that it is last. So CPU 1 traces completion and does
the needed wakeup.
o The awakened rcu_barrier() function does cleanup and releases
rcu_state.barrier_mutex.
o Another CPU now acquires rcu_state.barrier_mutex and starts
another round of rcu_barrier() processing, including updating
rcu_state.barrier_sequence.
o CPU 0 gets its act back together and does its tracing. Except
that rcu_state.barrier_sequence has already been updated, so
its tracing is incorrect and probably quite confusing.
(Wait! Why did this CPU check in twice for one rcu_barrier()
invocation???)
This commit therefore causes rcu_barrier_callback() to take a
snapshot of the value of rcu_state.barrier_sequence before invoking
atomic_dec_and_test(), thus guaranteeing that the event-trace output
is sensible, even if the timing of the event-trace output might still
be confusing. (Wait! Why did the old rcu_barrier() complete before
all of its CPUs checked in???) But being that this is RCU, only so much
confusion can reasonably be eliminated.
This data race was reported by KCSAN. Not appropriate for backporting
due to failure being unlikely and due to the mild consequences of the
failure, namely a confusing event trace.
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-21 07:43:45 +08:00
|
|
|
|
rcu_barrier_trace(TPS("CB"), -1, s);
|
2012-05-24 09:47:05 +08:00
|
|
|
|
}
|
2009-10-07 12:48:16 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
2021-12-15 05:15:18 +08:00
|
|
|
|
* If needed, entrain an rcu_barrier() callback on rdp->cblist.
|
2009-10-07 12:48:16 +08:00
|
|
|
|
*/
|
2021-12-15 05:15:18 +08:00
|
|
|
|
static void rcu_barrier_entrain(struct rcu_data *rdp)
|
2009-10-07 12:48:16 +08:00
|
|
|
|
{
|
2021-12-15 05:15:18 +08:00
|
|
|
|
unsigned long gseq = READ_ONCE(rcu_state.barrier_sequence);
|
|
|
|
|
unsigned long lseq = READ_ONCE(rdp->barrier_seq_snap);
|
2022-10-17 00:22:53 +08:00
|
|
|
|
bool wake_nocb = false;
|
|
|
|
|
bool was_alldone = false;
|
2009-10-07 12:48:16 +08:00
|
|
|
|
|
rcu: Make rcu_barrier() no longer block CPU-hotplug operations
This commit removes the cpus_read_lock() and cpus_read_unlock() calls
from rcu_barrier(), thus allowing CPUs to come and go during the course
of rcu_barrier() execution. Posting of the ->barrier_head callbacks does
synchronize with portions of RCU's CPU-hotplug notifiers, but these locks
are held for short time periods on both sides. Thus, full CPU-hotplug
operations could both start and finish during the execution of a given
rcu_barrier() invocation.
Additional synchronization is provided by a global ->barrier_lock.
Since the ->barrier_lock is only used during rcu_barrier() execution and
during onlining/offlining a CPU, the contention for this lock should
be low. It might be tempting to make use of a per-CPU lock just on
general principles, but straightforward attempts to do this have the
problems shown below.
Initial state: 3 CPUs present, CPU 0 and CPU1 do not have
any callback and CPU2 has callbacks.
1. CPU0 calls rcu_barrier().
2. CPU1 starts offlining for CPU2. CPU1 calls
rcutree_migrate_callbacks(). rcu_barrier_entrain() is called
from rcutree_migrate_callbacks(), with CPU2's rdp->barrier_lock.
It does not entrain ->barrier_head for CPU2, as rcu_barrier()
on CPU0 hasn't started the barrier sequence (by calling
rcu_seq_start(&rcu_state.barrier_sequence)) yet.
3. CPU0 starts new barrier sequence. It iterates over
CPU0 and CPU1, after acquiring their per-cpu ->barrier_lock
and finds 0 segcblist length. It updates ->barrier_seq_snap
for CPU0 and CPU1 and continues loop iteration to CPU2.
for_each_possible_cpu(cpu) {
raw_spin_lock_irqsave(&rdp->barrier_lock, flags);
if (!rcu_segcblist_n_cbs(&rdp->cblist)) {
WRITE_ONCE(rdp->barrier_seq_snap, gseq);
raw_spin_unlock_irqrestore(&rdp->barrier_lock, flags);
rcu_barrier_trace(TPS("NQ"), cpu, rcu_state.barrier_sequence);
continue;
}
4. rcutree_migrate_callbacks() completes execution on CPU1.
Segcblist len for CPU2 becomes 0.
5. The loop iteration on CPU0, checks rcu_segcblist_n_cbs(&rdp->cblist)
for CPU2 and completes the loop iteration after setting
->barrier_seq_snap.
6. As there isn't any ->barrier_head callback entrained; at
this point, rcu_barrier() in CPU0 returns.
7. The callbacks, which migrated from CPU2 to CPU1, execute.
Straightforward per-CPU locking is also subject to the following race
condition noted by Boqun Feng:
1. CPU0 calls rcu_barrier(), starting a new barrier sequence by invoking
rcu_seq_start() and init_completion(), but does not yet initialize
rcu_state.barrier_cpu_count.
2. CPU1 starts offlining for CPU2, calling rcutree_migrate_callbacks(),
which in turn calls rcu_barrier_entrain() holding CPU2's.
rdp->barrier_lock. It then entrains ->barrier_head for CPU2
and atomically increments rcu_state.barrier_cpu_count, which is
unfortunately not yet initialized to the value 2.
3. The just-entrained RCU callback is invoked. It atomically
decrements rcu_state.barrier_cpu_count and sees that it is
now zero. This callback therefore invokes complete().
4. CPU0 continues executing rcu_barrier(), but is not blocked
by its call to wait_for_completion(). This results in rcu_barrier()
returning before all pre-existing callbacks have been invoked,
which is a bug.
Therefore, synchronization is provided by rcu_state.barrier_lock,
which is also held across the initialization sequence, especially the
rcu_seq_start() and the atomic_set() that sets rcu_state.barrier_cpu_count
to the value 2. In addition, this lock is held when entraining the
rcu_barrier() callback, when deciding whether or not a CPU has callbacks
that rcu_barrier() must wait on, when setting the ->qsmaskinitnext for
incoming CPUs, and when migrating callbacks from a CPU that is going
offline.
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Co-developed-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-12-15 05:35:17 +08:00
|
|
|
|
lockdep_assert_held(&rcu_state.barrier_lock);
|
2021-12-15 05:15:18 +08:00
|
|
|
|
if (rcu_seq_state(lseq) || !rcu_seq_state(gseq) || rcu_seq_ctr(lseq) != rcu_seq_ctr(gseq))
|
|
|
|
|
return;
|
2018-07-11 09:37:30 +08:00
|
|
|
|
rcu_barrier_trace(TPS("IRQ"), -1, rcu_state.barrier_sequence);
|
2017-04-11 06:40:35 +08:00
|
|
|
|
rdp->barrier_head.func = rcu_barrier_callback;
|
|
|
|
|
debug_rcu_head_queue(&rdp->barrier_head);
|
2019-05-16 00:56:40 +08:00
|
|
|
|
rcu_nocb_lock(rdp);
|
2022-10-17 00:22:53 +08:00
|
|
|
|
/*
|
|
|
|
|
* Flush bypass and wakeup rcuog if we add callbacks to an empty regular
|
|
|
|
|
* queue. This way we don't wait for bypass timer that can reach seconds
|
|
|
|
|
* if it's fully lazy.
|
|
|
|
|
*/
|
|
|
|
|
was_alldone = rcu_rdp_is_offloaded(rdp) && !rcu_segcblist_pend_cbs(&rdp->cblist);
|
2022-10-17 00:22:54 +08:00
|
|
|
|
WARN_ON_ONCE(!rcu_nocb_flush_bypass(rdp, NULL, jiffies, false));
|
2022-10-17 00:22:53 +08:00
|
|
|
|
wake_nocb = was_alldone && rcu_segcblist_pend_cbs(&rdp->cblist);
|
2019-08-31 00:36:32 +08:00
|
|
|
|
if (rcu_segcblist_entrain(&rdp->cblist, &rdp->barrier_head)) {
|
2018-07-06 07:26:12 +08:00
|
|
|
|
atomic_inc(&rcu_state.barrier_cpu_count);
|
2017-04-11 06:40:35 +08:00
|
|
|
|
} else {
|
|
|
|
|
debug_rcu_head_unqueue(&rdp->barrier_head);
|
2021-12-15 05:15:18 +08:00
|
|
|
|
rcu_barrier_trace(TPS("IRQNQ"), -1, rcu_state.barrier_sequence);
|
2017-04-11 06:40:35 +08:00
|
|
|
|
}
|
2019-05-16 00:56:40 +08:00
|
|
|
|
rcu_nocb_unlock(rdp);
|
2022-10-17 00:22:53 +08:00
|
|
|
|
if (wake_nocb)
|
|
|
|
|
wake_nocb_gp(rdp, false);
|
2021-12-15 05:15:18 +08:00
|
|
|
|
smp_store_release(&rdp->barrier_seq_snap, gseq);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Called with preemption disabled, and from cross-cpu IRQ context.
|
|
|
|
|
*/
|
|
|
|
|
static void rcu_barrier_handler(void *cpu_in)
|
|
|
|
|
{
|
|
|
|
|
uintptr_t cpu = (uintptr_t)cpu_in;
|
|
|
|
|
struct rcu_data *rdp = per_cpu_ptr(&rcu_data, cpu);
|
|
|
|
|
|
|
|
|
|
lockdep_assert_irqs_disabled();
|
|
|
|
|
WARN_ON_ONCE(cpu != rdp->cpu);
|
|
|
|
|
WARN_ON_ONCE(cpu != smp_processor_id());
|
rcu: Make rcu_barrier() no longer block CPU-hotplug operations
This commit removes the cpus_read_lock() and cpus_read_unlock() calls
from rcu_barrier(), thus allowing CPUs to come and go during the course
of rcu_barrier() execution. Posting of the ->barrier_head callbacks does
synchronize with portions of RCU's CPU-hotplug notifiers, but these locks
are held for short time periods on both sides. Thus, full CPU-hotplug
operations could both start and finish during the execution of a given
rcu_barrier() invocation.
Additional synchronization is provided by a global ->barrier_lock.
Since the ->barrier_lock is only used during rcu_barrier() execution and
during onlining/offlining a CPU, the contention for this lock should
be low. It might be tempting to make use of a per-CPU lock just on
general principles, but straightforward attempts to do this have the
problems shown below.
Initial state: 3 CPUs present, CPU 0 and CPU1 do not have
any callback and CPU2 has callbacks.
1. CPU0 calls rcu_barrier().
2. CPU1 starts offlining for CPU2. CPU1 calls
rcutree_migrate_callbacks(). rcu_barrier_entrain() is called
from rcutree_migrate_callbacks(), with CPU2's rdp->barrier_lock.
It does not entrain ->barrier_head for CPU2, as rcu_barrier()
on CPU0 hasn't started the barrier sequence (by calling
rcu_seq_start(&rcu_state.barrier_sequence)) yet.
3. CPU0 starts new barrier sequence. It iterates over
CPU0 and CPU1, after acquiring their per-cpu ->barrier_lock
and finds 0 segcblist length. It updates ->barrier_seq_snap
for CPU0 and CPU1 and continues loop iteration to CPU2.
for_each_possible_cpu(cpu) {
raw_spin_lock_irqsave(&rdp->barrier_lock, flags);
if (!rcu_segcblist_n_cbs(&rdp->cblist)) {
WRITE_ONCE(rdp->barrier_seq_snap, gseq);
raw_spin_unlock_irqrestore(&rdp->barrier_lock, flags);
rcu_barrier_trace(TPS("NQ"), cpu, rcu_state.barrier_sequence);
continue;
}
4. rcutree_migrate_callbacks() completes execution on CPU1.
Segcblist len for CPU2 becomes 0.
5. The loop iteration on CPU0, checks rcu_segcblist_n_cbs(&rdp->cblist)
for CPU2 and completes the loop iteration after setting
->barrier_seq_snap.
6. As there isn't any ->barrier_head callback entrained; at
this point, rcu_barrier() in CPU0 returns.
7. The callbacks, which migrated from CPU2 to CPU1, execute.
Straightforward per-CPU locking is also subject to the following race
condition noted by Boqun Feng:
1. CPU0 calls rcu_barrier(), starting a new barrier sequence by invoking
rcu_seq_start() and init_completion(), but does not yet initialize
rcu_state.barrier_cpu_count.
2. CPU1 starts offlining for CPU2, calling rcutree_migrate_callbacks(),
which in turn calls rcu_barrier_entrain() holding CPU2's.
rdp->barrier_lock. It then entrains ->barrier_head for CPU2
and atomically increments rcu_state.barrier_cpu_count, which is
unfortunately not yet initialized to the value 2.
3. The just-entrained RCU callback is invoked. It atomically
decrements rcu_state.barrier_cpu_count and sees that it is
now zero. This callback therefore invokes complete().
4. CPU0 continues executing rcu_barrier(), but is not blocked
by its call to wait_for_completion(). This results in rcu_barrier()
returning before all pre-existing callbacks have been invoked,
which is a bug.
Therefore, synchronization is provided by rcu_state.barrier_lock,
which is also held across the initialization sequence, especially the
rcu_seq_start() and the atomic_set() that sets rcu_state.barrier_cpu_count
to the value 2. In addition, this lock is held when entraining the
rcu_barrier() callback, when deciding whether or not a CPU has callbacks
that rcu_barrier() must wait on, when setting the ->qsmaskinitnext for
incoming CPUs, and when migrating callbacks from a CPU that is going
offline.
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Co-developed-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-12-15 05:35:17 +08:00
|
|
|
|
raw_spin_lock(&rcu_state.barrier_lock);
|
2021-12-15 05:15:18 +08:00
|
|
|
|
rcu_barrier_entrain(rdp);
|
rcu: Make rcu_barrier() no longer block CPU-hotplug operations
This commit removes the cpus_read_lock() and cpus_read_unlock() calls
from rcu_barrier(), thus allowing CPUs to come and go during the course
of rcu_barrier() execution. Posting of the ->barrier_head callbacks does
synchronize with portions of RCU's CPU-hotplug notifiers, but these locks
are held for short time periods on both sides. Thus, full CPU-hotplug
operations could both start and finish during the execution of a given
rcu_barrier() invocation.
Additional synchronization is provided by a global ->barrier_lock.
Since the ->barrier_lock is only used during rcu_barrier() execution and
during onlining/offlining a CPU, the contention for this lock should
be low. It might be tempting to make use of a per-CPU lock just on
general principles, but straightforward attempts to do this have the
problems shown below.
Initial state: 3 CPUs present, CPU 0 and CPU1 do not have
any callback and CPU2 has callbacks.
1. CPU0 calls rcu_barrier().
2. CPU1 starts offlining for CPU2. CPU1 calls
rcutree_migrate_callbacks(). rcu_barrier_entrain() is called
from rcutree_migrate_callbacks(), with CPU2's rdp->barrier_lock.
It does not entrain ->barrier_head for CPU2, as rcu_barrier()
on CPU0 hasn't started the barrier sequence (by calling
rcu_seq_start(&rcu_state.barrier_sequence)) yet.
3. CPU0 starts new barrier sequence. It iterates over
CPU0 and CPU1, after acquiring their per-cpu ->barrier_lock
and finds 0 segcblist length. It updates ->barrier_seq_snap
for CPU0 and CPU1 and continues loop iteration to CPU2.
for_each_possible_cpu(cpu) {
raw_spin_lock_irqsave(&rdp->barrier_lock, flags);
if (!rcu_segcblist_n_cbs(&rdp->cblist)) {
WRITE_ONCE(rdp->barrier_seq_snap, gseq);
raw_spin_unlock_irqrestore(&rdp->barrier_lock, flags);
rcu_barrier_trace(TPS("NQ"), cpu, rcu_state.barrier_sequence);
continue;
}
4. rcutree_migrate_callbacks() completes execution on CPU1.
Segcblist len for CPU2 becomes 0.
5. The loop iteration on CPU0, checks rcu_segcblist_n_cbs(&rdp->cblist)
for CPU2 and completes the loop iteration after setting
->barrier_seq_snap.
6. As there isn't any ->barrier_head callback entrained; at
this point, rcu_barrier() in CPU0 returns.
7. The callbacks, which migrated from CPU2 to CPU1, execute.
Straightforward per-CPU locking is also subject to the following race
condition noted by Boqun Feng:
1. CPU0 calls rcu_barrier(), starting a new barrier sequence by invoking
rcu_seq_start() and init_completion(), but does not yet initialize
rcu_state.barrier_cpu_count.
2. CPU1 starts offlining for CPU2, calling rcutree_migrate_callbacks(),
which in turn calls rcu_barrier_entrain() holding CPU2's.
rdp->barrier_lock. It then entrains ->barrier_head for CPU2
and atomically increments rcu_state.barrier_cpu_count, which is
unfortunately not yet initialized to the value 2.
3. The just-entrained RCU callback is invoked. It atomically
decrements rcu_state.barrier_cpu_count and sees that it is
now zero. This callback therefore invokes complete().
4. CPU0 continues executing rcu_barrier(), but is not blocked
by its call to wait_for_completion(). This results in rcu_barrier()
returning before all pre-existing callbacks have been invoked,
which is a bug.
Therefore, synchronization is provided by rcu_state.barrier_lock,
which is also held across the initialization sequence, especially the
rcu_seq_start() and the atomic_set() that sets rcu_state.barrier_cpu_count
to the value 2. In addition, this lock is held when entraining the
rcu_barrier() callback, when deciding whether or not a CPU has callbacks
that rcu_barrier() must wait on, when setting the ->qsmaskinitnext for
incoming CPUs, and when migrating callbacks from a CPU that is going
offline.
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Co-developed-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-12-15 05:35:17 +08:00
|
|
|
|
raw_spin_unlock(&rcu_state.barrier_lock);
|
2009-10-07 12:48:16 +08:00
|
|
|
|
}
|
|
|
|
|
|
2018-07-11 09:37:30 +08:00
|
|
|
|
/**
|
|
|
|
|
* rcu_barrier - Wait until all in-flight call_rcu() callbacks complete.
|
|
|
|
|
*
|
|
|
|
|
* Note that this primitive does not necessarily wait for an RCU grace period
|
|
|
|
|
* to complete. For example, if there are no RCU callbacks queued anywhere
|
|
|
|
|
* in the system, then rcu_barrier() is within its rights to return
|
|
|
|
|
* immediately, without waiting for anything, much less an RCU grace period.
|
2009-10-07 12:48:16 +08:00
|
|
|
|
*/
|
2018-07-11 09:37:30 +08:00
|
|
|
|
void rcu_barrier(void)
|
2009-10-07 12:48:16 +08:00
|
|
|
|
{
|
2020-02-11 22:17:33 +08:00
|
|
|
|
uintptr_t cpu;
|
2021-12-15 05:15:18 +08:00
|
|
|
|
unsigned long flags;
|
|
|
|
|
unsigned long gseq;
|
2012-03-02 05:18:08 +08:00
|
|
|
|
struct rcu_data *rdp;
|
2018-07-06 07:26:12 +08:00
|
|
|
|
unsigned long s = rcu_seq_snap(&rcu_state.barrier_sequence);
|
2012-03-02 05:18:08 +08:00
|
|
|
|
|
2018-07-11 09:37:30 +08:00
|
|
|
|
rcu_barrier_trace(TPS("Begin"), -1, s);
|
2012-03-02 05:18:08 +08:00
|
|
|
|
|
2009-10-07 12:48:17 +08:00
|
|
|
|
/* Take mutex to serialize concurrent rcu_barrier() requests. */
|
2018-07-06 07:26:12 +08:00
|
|
|
|
mutex_lock(&rcu_state.barrier_mutex);
|
2012-03-02 05:18:08 +08:00
|
|
|
|
|
2015-06-27 02:20:00 +08:00
|
|
|
|
/* Did someone else do our work for us? */
|
2018-07-06 07:26:12 +08:00
|
|
|
|
if (rcu_seq_done(&rcu_state.barrier_sequence, s)) {
|
2021-12-11 08:25:20 +08:00
|
|
|
|
rcu_barrier_trace(TPS("EarlyExit"), -1, rcu_state.barrier_sequence);
|
2012-05-30 05:56:46 +08:00
|
|
|
|
smp_mb(); /* caller's subsequent code after above check. */
|
2018-07-06 07:26:12 +08:00
|
|
|
|
mutex_unlock(&rcu_state.barrier_mutex);
|
2012-05-30 05:56:46 +08:00
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
2015-06-27 02:20:00 +08:00
|
|
|
|
/* Mark the start of the barrier operation. */
|
rcu: Make rcu_barrier() no longer block CPU-hotplug operations
This commit removes the cpus_read_lock() and cpus_read_unlock() calls
from rcu_barrier(), thus allowing CPUs to come and go during the course
of rcu_barrier() execution. Posting of the ->barrier_head callbacks does
synchronize with portions of RCU's CPU-hotplug notifiers, but these locks
are held for short time periods on both sides. Thus, full CPU-hotplug
operations could both start and finish during the execution of a given
rcu_barrier() invocation.
Additional synchronization is provided by a global ->barrier_lock.
Since the ->barrier_lock is only used during rcu_barrier() execution and
during onlining/offlining a CPU, the contention for this lock should
be low. It might be tempting to make use of a per-CPU lock just on
general principles, but straightforward attempts to do this have the
problems shown below.
Initial state: 3 CPUs present, CPU 0 and CPU1 do not have
any callback and CPU2 has callbacks.
1. CPU0 calls rcu_barrier().
2. CPU1 starts offlining for CPU2. CPU1 calls
rcutree_migrate_callbacks(). rcu_barrier_entrain() is called
from rcutree_migrate_callbacks(), with CPU2's rdp->barrier_lock.
It does not entrain ->barrier_head for CPU2, as rcu_barrier()
on CPU0 hasn't started the barrier sequence (by calling
rcu_seq_start(&rcu_state.barrier_sequence)) yet.
3. CPU0 starts new barrier sequence. It iterates over
CPU0 and CPU1, after acquiring their per-cpu ->barrier_lock
and finds 0 segcblist length. It updates ->barrier_seq_snap
for CPU0 and CPU1 and continues loop iteration to CPU2.
for_each_possible_cpu(cpu) {
raw_spin_lock_irqsave(&rdp->barrier_lock, flags);
if (!rcu_segcblist_n_cbs(&rdp->cblist)) {
WRITE_ONCE(rdp->barrier_seq_snap, gseq);
raw_spin_unlock_irqrestore(&rdp->barrier_lock, flags);
rcu_barrier_trace(TPS("NQ"), cpu, rcu_state.barrier_sequence);
continue;
}
4. rcutree_migrate_callbacks() completes execution on CPU1.
Segcblist len for CPU2 becomes 0.
5. The loop iteration on CPU0, checks rcu_segcblist_n_cbs(&rdp->cblist)
for CPU2 and completes the loop iteration after setting
->barrier_seq_snap.
6. As there isn't any ->barrier_head callback entrained; at
this point, rcu_barrier() in CPU0 returns.
7. The callbacks, which migrated from CPU2 to CPU1, execute.
Straightforward per-CPU locking is also subject to the following race
condition noted by Boqun Feng:
1. CPU0 calls rcu_barrier(), starting a new barrier sequence by invoking
rcu_seq_start() and init_completion(), but does not yet initialize
rcu_state.barrier_cpu_count.
2. CPU1 starts offlining for CPU2, calling rcutree_migrate_callbacks(),
which in turn calls rcu_barrier_entrain() holding CPU2's.
rdp->barrier_lock. It then entrains ->barrier_head for CPU2
and atomically increments rcu_state.barrier_cpu_count, which is
unfortunately not yet initialized to the value 2.
3. The just-entrained RCU callback is invoked. It atomically
decrements rcu_state.barrier_cpu_count and sees that it is
now zero. This callback therefore invokes complete().
4. CPU0 continues executing rcu_barrier(), but is not blocked
by its call to wait_for_completion(). This results in rcu_barrier()
returning before all pre-existing callbacks have been invoked,
which is a bug.
Therefore, synchronization is provided by rcu_state.barrier_lock,
which is also held across the initialization sequence, especially the
rcu_seq_start() and the atomic_set() that sets rcu_state.barrier_cpu_count
to the value 2. In addition, this lock is held when entraining the
rcu_barrier() callback, when deciding whether or not a CPU has callbacks
that rcu_barrier() must wait on, when setting the ->qsmaskinitnext for
incoming CPUs, and when migrating callbacks from a CPU that is going
offline.
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Co-developed-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-12-15 05:35:17 +08:00
|
|
|
|
raw_spin_lock_irqsave(&rcu_state.barrier_lock, flags);
|
2018-07-06 07:26:12 +08:00
|
|
|
|
rcu_seq_start(&rcu_state.barrier_sequence);
|
2021-12-15 05:15:18 +08:00
|
|
|
|
gseq = rcu_state.barrier_sequence;
|
2018-07-11 09:37:30 +08:00
|
|
|
|
rcu_barrier_trace(TPS("Inc1"), -1, rcu_state.barrier_sequence);
|
2012-03-02 05:18:08 +08:00
|
|
|
|
|
2009-10-07 12:48:16 +08:00
|
|
|
|
/*
|
2020-02-11 22:17:33 +08:00
|
|
|
|
* Initialize the count to two rather than to zero in order
|
|
|
|
|
* to avoid a too-soon return to zero in case of an immediate
|
|
|
|
|
* invocation of the just-enqueued callback (or preemption of
|
|
|
|
|
* this task). Exclude CPU-hotplug operations to ensure that no
|
|
|
|
|
* offline non-offloaded CPU has callbacks queued.
|
2009-10-07 12:48:16 +08:00
|
|
|
|
*/
|
2018-07-06 07:26:12 +08:00
|
|
|
|
init_completion(&rcu_state.barrier_completion);
|
2020-02-11 22:17:33 +08:00
|
|
|
|
atomic_set(&rcu_state.barrier_cpu_count, 2);
|
rcu: Make rcu_barrier() no longer block CPU-hotplug operations
This commit removes the cpus_read_lock() and cpus_read_unlock() calls
from rcu_barrier(), thus allowing CPUs to come and go during the course
of rcu_barrier() execution. Posting of the ->barrier_head callbacks does
synchronize with portions of RCU's CPU-hotplug notifiers, but these locks
are held for short time periods on both sides. Thus, full CPU-hotplug
operations could both start and finish during the execution of a given
rcu_barrier() invocation.
Additional synchronization is provided by a global ->barrier_lock.
Since the ->barrier_lock is only used during rcu_barrier() execution and
during onlining/offlining a CPU, the contention for this lock should
be low. It might be tempting to make use of a per-CPU lock just on
general principles, but straightforward attempts to do this have the
problems shown below.
Initial state: 3 CPUs present, CPU 0 and CPU1 do not have
any callback and CPU2 has callbacks.
1. CPU0 calls rcu_barrier().
2. CPU1 starts offlining for CPU2. CPU1 calls
rcutree_migrate_callbacks(). rcu_barrier_entrain() is called
from rcutree_migrate_callbacks(), with CPU2's rdp->barrier_lock.
It does not entrain ->barrier_head for CPU2, as rcu_barrier()
on CPU0 hasn't started the barrier sequence (by calling
rcu_seq_start(&rcu_state.barrier_sequence)) yet.
3. CPU0 starts new barrier sequence. It iterates over
CPU0 and CPU1, after acquiring their per-cpu ->barrier_lock
and finds 0 segcblist length. It updates ->barrier_seq_snap
for CPU0 and CPU1 and continues loop iteration to CPU2.
for_each_possible_cpu(cpu) {
raw_spin_lock_irqsave(&rdp->barrier_lock, flags);
if (!rcu_segcblist_n_cbs(&rdp->cblist)) {
WRITE_ONCE(rdp->barrier_seq_snap, gseq);
raw_spin_unlock_irqrestore(&rdp->barrier_lock, flags);
rcu_barrier_trace(TPS("NQ"), cpu, rcu_state.barrier_sequence);
continue;
}
4. rcutree_migrate_callbacks() completes execution on CPU1.
Segcblist len for CPU2 becomes 0.
5. The loop iteration on CPU0, checks rcu_segcblist_n_cbs(&rdp->cblist)
for CPU2 and completes the loop iteration after setting
->barrier_seq_snap.
6. As there isn't any ->barrier_head callback entrained; at
this point, rcu_barrier() in CPU0 returns.
7. The callbacks, which migrated from CPU2 to CPU1, execute.
Straightforward per-CPU locking is also subject to the following race
condition noted by Boqun Feng:
1. CPU0 calls rcu_barrier(), starting a new barrier sequence by invoking
rcu_seq_start() and init_completion(), but does not yet initialize
rcu_state.barrier_cpu_count.
2. CPU1 starts offlining for CPU2, calling rcutree_migrate_callbacks(),
which in turn calls rcu_barrier_entrain() holding CPU2's.
rdp->barrier_lock. It then entrains ->barrier_head for CPU2
and atomically increments rcu_state.barrier_cpu_count, which is
unfortunately not yet initialized to the value 2.
3. The just-entrained RCU callback is invoked. It atomically
decrements rcu_state.barrier_cpu_count and sees that it is
now zero. This callback therefore invokes complete().
4. CPU0 continues executing rcu_barrier(), but is not blocked
by its call to wait_for_completion(). This results in rcu_barrier()
returning before all pre-existing callbacks have been invoked,
which is a bug.
Therefore, synchronization is provided by rcu_state.barrier_lock,
which is also held across the initialization sequence, especially the
rcu_seq_start() and the atomic_set() that sets rcu_state.barrier_cpu_count
to the value 2. In addition, this lock is held when entraining the
rcu_barrier() callback, when deciding whether or not a CPU has callbacks
that rcu_barrier() must wait on, when setting the ->qsmaskinitnext for
incoming CPUs, and when migrating callbacks from a CPU that is going
offline.
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Co-developed-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-12-15 05:35:17 +08:00
|
|
|
|
raw_spin_unlock_irqrestore(&rcu_state.barrier_lock, flags);
|
2012-03-02 05:18:08 +08:00
|
|
|
|
|
|
|
|
|
/*
|
2012-08-03 08:43:50 +08:00
|
|
|
|
* Force each CPU with callbacks to register a new callback.
|
|
|
|
|
* When that callback is invoked, we will know that all of the
|
|
|
|
|
* corresponding CPU's preceding callbacks have been invoked.
|
2012-03-02 05:18:08 +08:00
|
|
|
|
*/
|
2012-08-20 12:35:53 +08:00
|
|
|
|
for_each_possible_cpu(cpu) {
|
2018-07-04 06:37:16 +08:00
|
|
|
|
rdp = per_cpu_ptr(&rcu_data, cpu);
|
2021-12-15 05:15:18 +08:00
|
|
|
|
retry:
|
|
|
|
|
if (smp_load_acquire(&rdp->barrier_seq_snap) == gseq)
|
2019-04-13 06:58:34 +08:00
|
|
|
|
continue;
|
rcu: Make rcu_barrier() no longer block CPU-hotplug operations
This commit removes the cpus_read_lock() and cpus_read_unlock() calls
from rcu_barrier(), thus allowing CPUs to come and go during the course
of rcu_barrier() execution. Posting of the ->barrier_head callbacks does
synchronize with portions of RCU's CPU-hotplug notifiers, but these locks
are held for short time periods on both sides. Thus, full CPU-hotplug
operations could both start and finish during the execution of a given
rcu_barrier() invocation.
Additional synchronization is provided by a global ->barrier_lock.
Since the ->barrier_lock is only used during rcu_barrier() execution and
during onlining/offlining a CPU, the contention for this lock should
be low. It might be tempting to make use of a per-CPU lock just on
general principles, but straightforward attempts to do this have the
problems shown below.
Initial state: 3 CPUs present, CPU 0 and CPU1 do not have
any callback and CPU2 has callbacks.
1. CPU0 calls rcu_barrier().
2. CPU1 starts offlining for CPU2. CPU1 calls
rcutree_migrate_callbacks(). rcu_barrier_entrain() is called
from rcutree_migrate_callbacks(), with CPU2's rdp->barrier_lock.
It does not entrain ->barrier_head for CPU2, as rcu_barrier()
on CPU0 hasn't started the barrier sequence (by calling
rcu_seq_start(&rcu_state.barrier_sequence)) yet.
3. CPU0 starts new barrier sequence. It iterates over
CPU0 and CPU1, after acquiring their per-cpu ->barrier_lock
and finds 0 segcblist length. It updates ->barrier_seq_snap
for CPU0 and CPU1 and continues loop iteration to CPU2.
for_each_possible_cpu(cpu) {
raw_spin_lock_irqsave(&rdp->barrier_lock, flags);
if (!rcu_segcblist_n_cbs(&rdp->cblist)) {
WRITE_ONCE(rdp->barrier_seq_snap, gseq);
raw_spin_unlock_irqrestore(&rdp->barrier_lock, flags);
rcu_barrier_trace(TPS("NQ"), cpu, rcu_state.barrier_sequence);
continue;
}
4. rcutree_migrate_callbacks() completes execution on CPU1.
Segcblist len for CPU2 becomes 0.
5. The loop iteration on CPU0, checks rcu_segcblist_n_cbs(&rdp->cblist)
for CPU2 and completes the loop iteration after setting
->barrier_seq_snap.
6. As there isn't any ->barrier_head callback entrained; at
this point, rcu_barrier() in CPU0 returns.
7. The callbacks, which migrated from CPU2 to CPU1, execute.
Straightforward per-CPU locking is also subject to the following race
condition noted by Boqun Feng:
1. CPU0 calls rcu_barrier(), starting a new barrier sequence by invoking
rcu_seq_start() and init_completion(), but does not yet initialize
rcu_state.barrier_cpu_count.
2. CPU1 starts offlining for CPU2, calling rcutree_migrate_callbacks(),
which in turn calls rcu_barrier_entrain() holding CPU2's.
rdp->barrier_lock. It then entrains ->barrier_head for CPU2
and atomically increments rcu_state.barrier_cpu_count, which is
unfortunately not yet initialized to the value 2.
3. The just-entrained RCU callback is invoked. It atomically
decrements rcu_state.barrier_cpu_count and sees that it is
now zero. This callback therefore invokes complete().
4. CPU0 continues executing rcu_barrier(), but is not blocked
by its call to wait_for_completion(). This results in rcu_barrier()
returning before all pre-existing callbacks have been invoked,
which is a bug.
Therefore, synchronization is provided by rcu_state.barrier_lock,
which is also held across the initialization sequence, especially the
rcu_seq_start() and the atomic_set() that sets rcu_state.barrier_cpu_count
to the value 2. In addition, this lock is held when entraining the
rcu_barrier() callback, when deciding whether or not a CPU has callbacks
that rcu_barrier() must wait on, when setting the ->qsmaskinitnext for
incoming CPUs, and when migrating callbacks from a CPU that is going
offline.
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Co-developed-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-12-15 05:35:17 +08:00
|
|
|
|
raw_spin_lock_irqsave(&rcu_state.barrier_lock, flags);
|
2021-12-11 08:25:20 +08:00
|
|
|
|
if (!rcu_segcblist_n_cbs(&rdp->cblist)) {
|
2021-12-15 05:15:18 +08:00
|
|
|
|
WRITE_ONCE(rdp->barrier_seq_snap, gseq);
|
rcu: Make rcu_barrier() no longer block CPU-hotplug operations
This commit removes the cpus_read_lock() and cpus_read_unlock() calls
from rcu_barrier(), thus allowing CPUs to come and go during the course
of rcu_barrier() execution. Posting of the ->barrier_head callbacks does
synchronize with portions of RCU's CPU-hotplug notifiers, but these locks
are held for short time periods on both sides. Thus, full CPU-hotplug
operations could both start and finish during the execution of a given
rcu_barrier() invocation.
Additional synchronization is provided by a global ->barrier_lock.
Since the ->barrier_lock is only used during rcu_barrier() execution and
during onlining/offlining a CPU, the contention for this lock should
be low. It might be tempting to make use of a per-CPU lock just on
general principles, but straightforward attempts to do this have the
problems shown below.
Initial state: 3 CPUs present, CPU 0 and CPU1 do not have
any callback and CPU2 has callbacks.
1. CPU0 calls rcu_barrier().
2. CPU1 starts offlining for CPU2. CPU1 calls
rcutree_migrate_callbacks(). rcu_barrier_entrain() is called
from rcutree_migrate_callbacks(), with CPU2's rdp->barrier_lock.
It does not entrain ->barrier_head for CPU2, as rcu_barrier()
on CPU0 hasn't started the barrier sequence (by calling
rcu_seq_start(&rcu_state.barrier_sequence)) yet.
3. CPU0 starts new barrier sequence. It iterates over
CPU0 and CPU1, after acquiring their per-cpu ->barrier_lock
and finds 0 segcblist length. It updates ->barrier_seq_snap
for CPU0 and CPU1 and continues loop iteration to CPU2.
for_each_possible_cpu(cpu) {
raw_spin_lock_irqsave(&rdp->barrier_lock, flags);
if (!rcu_segcblist_n_cbs(&rdp->cblist)) {
WRITE_ONCE(rdp->barrier_seq_snap, gseq);
raw_spin_unlock_irqrestore(&rdp->barrier_lock, flags);
rcu_barrier_trace(TPS("NQ"), cpu, rcu_state.barrier_sequence);
continue;
}
4. rcutree_migrate_callbacks() completes execution on CPU1.
Segcblist len for CPU2 becomes 0.
5. The loop iteration on CPU0, checks rcu_segcblist_n_cbs(&rdp->cblist)
for CPU2 and completes the loop iteration after setting
->barrier_seq_snap.
6. As there isn't any ->barrier_head callback entrained; at
this point, rcu_barrier() in CPU0 returns.
7. The callbacks, which migrated from CPU2 to CPU1, execute.
Straightforward per-CPU locking is also subject to the following race
condition noted by Boqun Feng:
1. CPU0 calls rcu_barrier(), starting a new barrier sequence by invoking
rcu_seq_start() and init_completion(), but does not yet initialize
rcu_state.barrier_cpu_count.
2. CPU1 starts offlining for CPU2, calling rcutree_migrate_callbacks(),
which in turn calls rcu_barrier_entrain() holding CPU2's.
rdp->barrier_lock. It then entrains ->barrier_head for CPU2
and atomically increments rcu_state.barrier_cpu_count, which is
unfortunately not yet initialized to the value 2.
3. The just-entrained RCU callback is invoked. It atomically
decrements rcu_state.barrier_cpu_count and sees that it is
now zero. This callback therefore invokes complete().
4. CPU0 continues executing rcu_barrier(), but is not blocked
by its call to wait_for_completion(). This results in rcu_barrier()
returning before all pre-existing callbacks have been invoked,
which is a bug.
Therefore, synchronization is provided by rcu_state.barrier_lock,
which is also held across the initialization sequence, especially the
rcu_seq_start() and the atomic_set() that sets rcu_state.barrier_cpu_count
to the value 2. In addition, this lock is held when entraining the
rcu_barrier() callback, when deciding whether or not a CPU has callbacks
that rcu_barrier() must wait on, when setting the ->qsmaskinitnext for
incoming CPUs, and when migrating callbacks from a CPU that is going
offline.
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Co-developed-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-12-15 05:35:17 +08:00
|
|
|
|
raw_spin_unlock_irqrestore(&rcu_state.barrier_lock, flags);
|
2021-12-11 08:25:20 +08:00
|
|
|
|
rcu_barrier_trace(TPS("NQ"), cpu, rcu_state.barrier_sequence);
|
2019-04-13 06:58:34 +08:00
|
|
|
|
continue;
|
2021-12-11 08:25:20 +08:00
|
|
|
|
}
|
2021-12-15 05:15:18 +08:00
|
|
|
|
if (!rcu_rdp_cpu_online(rdp)) {
|
|
|
|
|
rcu_barrier_entrain(rdp);
|
|
|
|
|
WARN_ON_ONCE(READ_ONCE(rdp->barrier_seq_snap) != gseq);
|
rcu: Make rcu_barrier() no longer block CPU-hotplug operations
This commit removes the cpus_read_lock() and cpus_read_unlock() calls
from rcu_barrier(), thus allowing CPUs to come and go during the course
of rcu_barrier() execution. Posting of the ->barrier_head callbacks does
synchronize with portions of RCU's CPU-hotplug notifiers, but these locks
are held for short time periods on both sides. Thus, full CPU-hotplug
operations could both start and finish during the execution of a given
rcu_barrier() invocation.
Additional synchronization is provided by a global ->barrier_lock.
Since the ->barrier_lock is only used during rcu_barrier() execution and
during onlining/offlining a CPU, the contention for this lock should
be low. It might be tempting to make use of a per-CPU lock just on
general principles, but straightforward attempts to do this have the
problems shown below.
Initial state: 3 CPUs present, CPU 0 and CPU1 do not have
any callback and CPU2 has callbacks.
1. CPU0 calls rcu_barrier().
2. CPU1 starts offlining for CPU2. CPU1 calls
rcutree_migrate_callbacks(). rcu_barrier_entrain() is called
from rcutree_migrate_callbacks(), with CPU2's rdp->barrier_lock.
It does not entrain ->barrier_head for CPU2, as rcu_barrier()
on CPU0 hasn't started the barrier sequence (by calling
rcu_seq_start(&rcu_state.barrier_sequence)) yet.
3. CPU0 starts new barrier sequence. It iterates over
CPU0 and CPU1, after acquiring their per-cpu ->barrier_lock
and finds 0 segcblist length. It updates ->barrier_seq_snap
for CPU0 and CPU1 and continues loop iteration to CPU2.
for_each_possible_cpu(cpu) {
raw_spin_lock_irqsave(&rdp->barrier_lock, flags);
if (!rcu_segcblist_n_cbs(&rdp->cblist)) {
WRITE_ONCE(rdp->barrier_seq_snap, gseq);
raw_spin_unlock_irqrestore(&rdp->barrier_lock, flags);
rcu_barrier_trace(TPS("NQ"), cpu, rcu_state.barrier_sequence);
continue;
}
4. rcutree_migrate_callbacks() completes execution on CPU1.
Segcblist len for CPU2 becomes 0.
5. The loop iteration on CPU0, checks rcu_segcblist_n_cbs(&rdp->cblist)
for CPU2 and completes the loop iteration after setting
->barrier_seq_snap.
6. As there isn't any ->barrier_head callback entrained; at
this point, rcu_barrier() in CPU0 returns.
7. The callbacks, which migrated from CPU2 to CPU1, execute.
Straightforward per-CPU locking is also subject to the following race
condition noted by Boqun Feng:
1. CPU0 calls rcu_barrier(), starting a new barrier sequence by invoking
rcu_seq_start() and init_completion(), but does not yet initialize
rcu_state.barrier_cpu_count.
2. CPU1 starts offlining for CPU2, calling rcutree_migrate_callbacks(),
which in turn calls rcu_barrier_entrain() holding CPU2's.
rdp->barrier_lock. It then entrains ->barrier_head for CPU2
and atomically increments rcu_state.barrier_cpu_count, which is
unfortunately not yet initialized to the value 2.
3. The just-entrained RCU callback is invoked. It atomically
decrements rcu_state.barrier_cpu_count and sees that it is
now zero. This callback therefore invokes complete().
4. CPU0 continues executing rcu_barrier(), but is not blocked
by its call to wait_for_completion(). This results in rcu_barrier()
returning before all pre-existing callbacks have been invoked,
which is a bug.
Therefore, synchronization is provided by rcu_state.barrier_lock,
which is also held across the initialization sequence, especially the
rcu_seq_start() and the atomic_set() that sets rcu_state.barrier_cpu_count
to the value 2. In addition, this lock is held when entraining the
rcu_barrier() callback, when deciding whether or not a CPU has callbacks
that rcu_barrier() must wait on, when setting the ->qsmaskinitnext for
incoming CPUs, and when migrating callbacks from a CPU that is going
offline.
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Co-developed-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-12-15 05:35:17 +08:00
|
|
|
|
raw_spin_unlock_irqrestore(&rcu_state.barrier_lock, flags);
|
2021-12-11 08:25:20 +08:00
|
|
|
|
rcu_barrier_trace(TPS("OfflineNoCBQ"), cpu, rcu_state.barrier_sequence);
|
2021-12-15 05:15:18 +08:00
|
|
|
|
continue;
|
2012-03-02 05:18:08 +08:00
|
|
|
|
}
|
rcu: Make rcu_barrier() no longer block CPU-hotplug operations
This commit removes the cpus_read_lock() and cpus_read_unlock() calls
from rcu_barrier(), thus allowing CPUs to come and go during the course
of rcu_barrier() execution. Posting of the ->barrier_head callbacks does
synchronize with portions of RCU's CPU-hotplug notifiers, but these locks
are held for short time periods on both sides. Thus, full CPU-hotplug
operations could both start and finish during the execution of a given
rcu_barrier() invocation.
Additional synchronization is provided by a global ->barrier_lock.
Since the ->barrier_lock is only used during rcu_barrier() execution and
during onlining/offlining a CPU, the contention for this lock should
be low. It might be tempting to make use of a per-CPU lock just on
general principles, but straightforward attempts to do this have the
problems shown below.
Initial state: 3 CPUs present, CPU 0 and CPU1 do not have
any callback and CPU2 has callbacks.
1. CPU0 calls rcu_barrier().
2. CPU1 starts offlining for CPU2. CPU1 calls
rcutree_migrate_callbacks(). rcu_barrier_entrain() is called
from rcutree_migrate_callbacks(), with CPU2's rdp->barrier_lock.
It does not entrain ->barrier_head for CPU2, as rcu_barrier()
on CPU0 hasn't started the barrier sequence (by calling
rcu_seq_start(&rcu_state.barrier_sequence)) yet.
3. CPU0 starts new barrier sequence. It iterates over
CPU0 and CPU1, after acquiring their per-cpu ->barrier_lock
and finds 0 segcblist length. It updates ->barrier_seq_snap
for CPU0 and CPU1 and continues loop iteration to CPU2.
for_each_possible_cpu(cpu) {
raw_spin_lock_irqsave(&rdp->barrier_lock, flags);
if (!rcu_segcblist_n_cbs(&rdp->cblist)) {
WRITE_ONCE(rdp->barrier_seq_snap, gseq);
raw_spin_unlock_irqrestore(&rdp->barrier_lock, flags);
rcu_barrier_trace(TPS("NQ"), cpu, rcu_state.barrier_sequence);
continue;
}
4. rcutree_migrate_callbacks() completes execution on CPU1.
Segcblist len for CPU2 becomes 0.
5. The loop iteration on CPU0, checks rcu_segcblist_n_cbs(&rdp->cblist)
for CPU2 and completes the loop iteration after setting
->barrier_seq_snap.
6. As there isn't any ->barrier_head callback entrained; at
this point, rcu_barrier() in CPU0 returns.
7. The callbacks, which migrated from CPU2 to CPU1, execute.
Straightforward per-CPU locking is also subject to the following race
condition noted by Boqun Feng:
1. CPU0 calls rcu_barrier(), starting a new barrier sequence by invoking
rcu_seq_start() and init_completion(), but does not yet initialize
rcu_state.barrier_cpu_count.
2. CPU1 starts offlining for CPU2, calling rcutree_migrate_callbacks(),
which in turn calls rcu_barrier_entrain() holding CPU2's.
rdp->barrier_lock. It then entrains ->barrier_head for CPU2
and atomically increments rcu_state.barrier_cpu_count, which is
unfortunately not yet initialized to the value 2.
3. The just-entrained RCU callback is invoked. It atomically
decrements rcu_state.barrier_cpu_count and sees that it is
now zero. This callback therefore invokes complete().
4. CPU0 continues executing rcu_barrier(), but is not blocked
by its call to wait_for_completion(). This results in rcu_barrier()
returning before all pre-existing callbacks have been invoked,
which is a bug.
Therefore, synchronization is provided by rcu_state.barrier_lock,
which is also held across the initialization sequence, especially the
rcu_seq_start() and the atomic_set() that sets rcu_state.barrier_cpu_count
to the value 2. In addition, this lock is held when entraining the
rcu_barrier() callback, when deciding whether or not a CPU has callbacks
that rcu_barrier() must wait on, when setting the ->qsmaskinitnext for
incoming CPUs, and when migrating callbacks from a CPU that is going
offline.
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Co-developed-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-12-15 05:35:17 +08:00
|
|
|
|
raw_spin_unlock_irqrestore(&rcu_state.barrier_lock, flags);
|
2021-12-15 05:15:18 +08:00
|
|
|
|
if (smp_call_function_single(cpu, rcu_barrier_handler, (void *)cpu, 1)) {
|
|
|
|
|
schedule_timeout_uninterruptible(1);
|
|
|
|
|
goto retry;
|
2012-03-02 05:18:08 +08:00
|
|
|
|
}
|
2021-12-15 05:15:18 +08:00
|
|
|
|
WARN_ON_ONCE(READ_ONCE(rdp->barrier_seq_snap) != gseq);
|
|
|
|
|
rcu_barrier_trace(TPS("OnlineQ"), cpu, rcu_state.barrier_sequence);
|
2012-03-02 05:18:08 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Now that we have an rcu_barrier_callback() callback on each
|
|
|
|
|
* CPU, and thus each counted, remove the initial count.
|
|
|
|
|
*/
|
2020-02-11 22:17:33 +08:00
|
|
|
|
if (atomic_sub_and_test(2, &rcu_state.barrier_cpu_count))
|
2018-07-06 07:26:12 +08:00
|
|
|
|
complete(&rcu_state.barrier_completion);
|
2012-03-02 05:18:08 +08:00
|
|
|
|
|
|
|
|
|
/* Wait for all rcu_barrier_callback() callbacks to be invoked. */
|
2018-07-06 07:26:12 +08:00
|
|
|
|
wait_for_completion(&rcu_state.barrier_completion);
|
2012-03-02 05:18:08 +08:00
|
|
|
|
|
2015-06-27 02:20:00 +08:00
|
|
|
|
/* Mark the end of the barrier operation. */
|
2018-07-11 09:37:30 +08:00
|
|
|
|
rcu_barrier_trace(TPS("Inc2"), -1, rcu_state.barrier_sequence);
|
2018-07-06 07:26:12 +08:00
|
|
|
|
rcu_seq_end(&rcu_state.barrier_sequence);
|
2021-12-15 05:15:18 +08:00
|
|
|
|
gseq = rcu_state.barrier_sequence;
|
|
|
|
|
for_each_possible_cpu(cpu) {
|
|
|
|
|
rdp = per_cpu_ptr(&rcu_data, cpu);
|
|
|
|
|
|
|
|
|
|
WRITE_ONCE(rdp->barrier_seq_snap, gseq);
|
|
|
|
|
}
|
2015-06-27 02:20:00 +08:00
|
|
|
|
|
2012-03-02 05:18:08 +08:00
|
|
|
|
/* Other rcu_barrier() invocations can now safely proceed. */
|
2018-07-06 07:26:12 +08:00
|
|
|
|
mutex_unlock(&rcu_state.barrier_mutex);
|
2009-10-07 12:48:16 +08:00
|
|
|
|
}
|
2018-07-03 05:30:37 +08:00
|
|
|
|
EXPORT_SYMBOL_GPL(rcu_barrier);
|
2009-10-07 12:48:16 +08:00
|
|
|
|
|
2022-11-07 08:33:38 +08:00
|
|
|
|
/*
|
|
|
|
|
* Compute the mask of online CPUs for the specified rcu_node structure.
|
|
|
|
|
* This will not be stable unless the rcu_node structure's ->lock is
|
|
|
|
|
* held, but the bit corresponding to the current CPU will be stable
|
|
|
|
|
* in most contexts.
|
|
|
|
|
*/
|
|
|
|
|
static unsigned long rcu_rnp_online_cpus(struct rcu_node *rnp)
|
|
|
|
|
{
|
|
|
|
|
return READ_ONCE(rnp->qsmaskinitnext);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Is the CPU corresponding to the specified rcu_data structure online
|
|
|
|
|
* from RCU's perspective? This perspective is given by that structure's
|
|
|
|
|
* ->qsmaskinitnext field rather than by the global cpu_online_mask.
|
|
|
|
|
*/
|
|
|
|
|
static bool rcu_rdp_cpu_online(struct rcu_data *rdp)
|
|
|
|
|
{
|
|
|
|
|
return !!(rdp->grpmask & rcu_rnp_online_cpus(rdp->mynode));
|
|
|
|
|
}
|
|
|
|
|
|
2023-10-27 22:40:47 +08:00
|
|
|
|
bool rcu_cpu_online(int cpu)
|
|
|
|
|
{
|
|
|
|
|
struct rcu_data *rdp = per_cpu_ptr(&rcu_data, cpu);
|
|
|
|
|
|
|
|
|
|
return rcu_rdp_cpu_online(rdp);
|
|
|
|
|
}
|
|
|
|
|
|
2022-11-07 08:33:38 +08:00
|
|
|
|
#if defined(CONFIG_PROVE_RCU) && defined(CONFIG_HOTPLUG_CPU)
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Is the current CPU online as far as RCU is concerned?
|
|
|
|
|
*
|
|
|
|
|
* Disable preemption to avoid false positives that could otherwise
|
|
|
|
|
* happen due to the current CPU number being sampled, this task being
|
|
|
|
|
* preempted, its old CPU being taken offline, resuming on some other CPU,
|
|
|
|
|
* then determining that its old CPU is now offline.
|
|
|
|
|
*
|
|
|
|
|
* Disable checking if in an NMI handler because we cannot safely
|
|
|
|
|
* report errors from NMI handlers anyway. In addition, it is OK to use
|
|
|
|
|
* RCU on an offline processor during initial boot, hence the check for
|
|
|
|
|
* rcu_scheduler_fully_active.
|
|
|
|
|
*/
|
|
|
|
|
bool rcu_lockdep_current_cpu_online(void)
|
|
|
|
|
{
|
|
|
|
|
struct rcu_data *rdp;
|
|
|
|
|
bool ret = false;
|
|
|
|
|
|
|
|
|
|
if (in_nmi() || !rcu_scheduler_fully_active)
|
|
|
|
|
return true;
|
|
|
|
|
preempt_disable_notrace();
|
|
|
|
|
rdp = this_cpu_ptr(&rcu_data);
|
|
|
|
|
/*
|
|
|
|
|
* Strictly, we care here about the case where the current CPU is
|
|
|
|
|
* in rcu_cpu_starting() and thus has an excuse for rdp->grpmask
|
|
|
|
|
* not being up to date. So arch_spin_is_locked() might have a
|
|
|
|
|
* false positive if it's held by some *other* CPU, but that's
|
|
|
|
|
* OK because that just means a false *negative* on the warning.
|
|
|
|
|
*/
|
|
|
|
|
if (rcu_rdp_cpu_online(rdp) || arch_spin_is_locked(&rcu_state.ofl_lock))
|
|
|
|
|
ret = true;
|
|
|
|
|
preempt_enable_notrace();
|
|
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(rcu_lockdep_current_cpu_online);
|
|
|
|
|
|
|
|
|
|
#endif /* #if defined(CONFIG_PROVE_RCU) && defined(CONFIG_HOTPLUG_CPU) */
|
|
|
|
|
|
|
|
|
|
// Has rcu_init() been invoked? This is used (for example) to determine
|
|
|
|
|
// whether spinlocks may be acquired safely.
|
|
|
|
|
static bool rcu_init_invoked(void)
|
|
|
|
|
{
|
|
|
|
|
return !!rcu_state.n_online_cpus;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Near the end of the offline process. Trace the fact that this CPU
|
|
|
|
|
* is going offline.
|
|
|
|
|
*/
|
|
|
|
|
int rcutree_dying_cpu(unsigned int cpu)
|
|
|
|
|
{
|
|
|
|
|
bool blkd;
|
|
|
|
|
struct rcu_data *rdp = per_cpu_ptr(&rcu_data, cpu);
|
|
|
|
|
struct rcu_node *rnp = rdp->mynode;
|
|
|
|
|
|
|
|
|
|
if (!IS_ENABLED(CONFIG_HOTPLUG_CPU))
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
blkd = !!(READ_ONCE(rnp->qsmask) & rdp->grpmask);
|
|
|
|
|
trace_rcu_grace_period(rcu_state.name, READ_ONCE(rnp->gp_seq),
|
|
|
|
|
blkd ? TPS("cpuofl-bgp") : TPS("cpuofl"));
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* All CPUs for the specified rcu_node structure have gone offline,
|
|
|
|
|
* and all tasks that were preempted within an RCU read-side critical
|
|
|
|
|
* section while running on one of those CPUs have since exited their RCU
|
|
|
|
|
* read-side critical section. Some other CPU is reporting this fact with
|
|
|
|
|
* the specified rcu_node structure's ->lock held and interrupts disabled.
|
|
|
|
|
* This function therefore goes up the tree of rcu_node structures,
|
|
|
|
|
* clearing the corresponding bits in the ->qsmaskinit fields. Note that
|
|
|
|
|
* the leaf rcu_node structure's ->qsmaskinit field has already been
|
|
|
|
|
* updated.
|
|
|
|
|
*
|
|
|
|
|
* This function does check that the specified rcu_node structure has
|
|
|
|
|
* all CPUs offline and no blocked tasks, so it is OK to invoke it
|
|
|
|
|
* prematurely. That said, invoking it after the fact will cost you
|
|
|
|
|
* a needless lock acquisition. So once it has done its work, don't
|
|
|
|
|
* invoke it again.
|
|
|
|
|
*/
|
|
|
|
|
static void rcu_cleanup_dead_rnp(struct rcu_node *rnp_leaf)
|
|
|
|
|
{
|
|
|
|
|
long mask;
|
|
|
|
|
struct rcu_node *rnp = rnp_leaf;
|
|
|
|
|
|
|
|
|
|
raw_lockdep_assert_held_rcu_node(rnp_leaf);
|
|
|
|
|
if (!IS_ENABLED(CONFIG_HOTPLUG_CPU) ||
|
|
|
|
|
WARN_ON_ONCE(rnp_leaf->qsmaskinit) ||
|
|
|
|
|
WARN_ON_ONCE(rcu_preempt_has_tasks(rnp_leaf)))
|
|
|
|
|
return;
|
|
|
|
|
for (;;) {
|
|
|
|
|
mask = rnp->grpmask;
|
|
|
|
|
rnp = rnp->parent;
|
|
|
|
|
if (!rnp)
|
|
|
|
|
break;
|
|
|
|
|
raw_spin_lock_rcu_node(rnp); /* irqs already disabled. */
|
|
|
|
|
rnp->qsmaskinit &= ~mask;
|
|
|
|
|
/* Between grace periods, so better already be zero! */
|
|
|
|
|
WARN_ON_ONCE(rnp->qsmask);
|
|
|
|
|
if (rnp->qsmaskinit) {
|
|
|
|
|
raw_spin_unlock_rcu_node(rnp);
|
|
|
|
|
/* irqs remain disabled. */
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
raw_spin_unlock_rcu_node(rnp); /* irqs remain disabled. */
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* The CPU has been completely removed, and some other CPU is reporting
|
|
|
|
|
* this fact from process context. Do the remainder of the cleanup.
|
|
|
|
|
* There can only be one CPU hotplug operation at a time, so no need for
|
|
|
|
|
* explicit locking.
|
|
|
|
|
*/
|
|
|
|
|
int rcutree_dead_cpu(unsigned int cpu)
|
|
|
|
|
{
|
|
|
|
|
if (!IS_ENABLED(CONFIG_HOTPLUG_CPU))
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
WRITE_ONCE(rcu_state.n_online_cpus, rcu_state.n_online_cpus - 1);
|
|
|
|
|
// Stop-machine done, so allow nohz_full to disable tick.
|
|
|
|
|
tick_dep_clear(TICK_DEP_BIT_RCU);
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
rcu: Process offlining and onlining only at grace-period start
Races between CPU hotplug and grace periods can be difficult to resolve,
so the ->onoff_mutex is used to exclude the two events. Unfortunately,
this means that it is impossible for an outgoing CPU to perform the
last bits of its offlining from its last pass through the idle loop,
because sleeplocks cannot be acquired in that context.
This commit avoids these problems by buffering online and offline events
in a new ->qsmaskinitnext field in the leaf rcu_node structures. When a
grace period starts, the events accumulated in this mask are applied to
the ->qsmaskinit field, and, if needed, up the rcu_node tree. The special
case of all CPUs corresponding to a given leaf rcu_node structure being
offline while there are still elements in that structure's ->blkd_tasks
list is handled using a new ->wait_blkd_tasks field. In this case,
propagating the offline bits up the tree is deferred until the beginning
of the grace period after all of the tasks have exited their RCU read-side
critical sections and removed themselves from the list, at which point
the ->wait_blkd_tasks flag is cleared. If one of that leaf rcu_node
structure's CPUs comes back online before the list empties, then the
->wait_blkd_tasks flag is simply cleared.
This of course means that RCU's notion of which CPUs are offline can be
out of date. This is OK because RCU need only wait on CPUs that were
online at the time that the grace period started. In addition, RCU's
force-quiescent-state actions will handle the case where a CPU goes
offline after the grace period starts.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2015-01-24 13:52:37 +08:00
|
|
|
|
/*
|
|
|
|
|
* Propagate ->qsinitmask bits up the rcu_node tree to account for the
|
|
|
|
|
* first CPU in a given leaf rcu_node structure coming online. The caller
|
2021-03-23 13:29:10 +08:00
|
|
|
|
* must hold the corresponding leaf rcu_node ->lock with interrupts
|
rcu: Process offlining and onlining only at grace-period start
Races between CPU hotplug and grace periods can be difficult to resolve,
so the ->onoff_mutex is used to exclude the two events. Unfortunately,
this means that it is impossible for an outgoing CPU to perform the
last bits of its offlining from its last pass through the idle loop,
because sleeplocks cannot be acquired in that context.
This commit avoids these problems by buffering online and offline events
in a new ->qsmaskinitnext field in the leaf rcu_node structures. When a
grace period starts, the events accumulated in this mask are applied to
the ->qsmaskinit field, and, if needed, up the rcu_node tree. The special
case of all CPUs corresponding to a given leaf rcu_node structure being
offline while there are still elements in that structure's ->blkd_tasks
list is handled using a new ->wait_blkd_tasks field. In this case,
propagating the offline bits up the tree is deferred until the beginning
of the grace period after all of the tasks have exited their RCU read-side
critical sections and removed themselves from the list, at which point
the ->wait_blkd_tasks flag is cleared. If one of that leaf rcu_node
structure's CPUs comes back online before the list empties, then the
->wait_blkd_tasks flag is simply cleared.
This of course means that RCU's notion of which CPUs are offline can be
out of date. This is OK because RCU need only wait on CPUs that were
online at the time that the grace period started. In addition, RCU's
force-quiescent-state actions will handle the case where a CPU goes
offline after the grace period starts.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2015-01-24 13:52:37 +08:00
|
|
|
|
* disabled.
|
|
|
|
|
*/
|
|
|
|
|
static void rcu_init_new_rnp(struct rcu_node *rnp_leaf)
|
|
|
|
|
{
|
|
|
|
|
long mask;
|
2018-05-03 05:46:43 +08:00
|
|
|
|
long oldmask;
|
rcu: Process offlining and onlining only at grace-period start
Races between CPU hotplug and grace periods can be difficult to resolve,
so the ->onoff_mutex is used to exclude the two events. Unfortunately,
this means that it is impossible for an outgoing CPU to perform the
last bits of its offlining from its last pass through the idle loop,
because sleeplocks cannot be acquired in that context.
This commit avoids these problems by buffering online and offline events
in a new ->qsmaskinitnext field in the leaf rcu_node structures. When a
grace period starts, the events accumulated in this mask are applied to
the ->qsmaskinit field, and, if needed, up the rcu_node tree. The special
case of all CPUs corresponding to a given leaf rcu_node structure being
offline while there are still elements in that structure's ->blkd_tasks
list is handled using a new ->wait_blkd_tasks field. In this case,
propagating the offline bits up the tree is deferred until the beginning
of the grace period after all of the tasks have exited their RCU read-side
critical sections and removed themselves from the list, at which point
the ->wait_blkd_tasks flag is cleared. If one of that leaf rcu_node
structure's CPUs comes back online before the list empties, then the
->wait_blkd_tasks flag is simply cleared.
This of course means that RCU's notion of which CPUs are offline can be
out of date. This is OK because RCU need only wait on CPUs that were
online at the time that the grace period started. In addition, RCU's
force-quiescent-state actions will handle the case where a CPU goes
offline after the grace period starts.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2015-01-24 13:52:37 +08:00
|
|
|
|
struct rcu_node *rnp = rnp_leaf;
|
|
|
|
|
|
2018-05-03 05:46:43 +08:00
|
|
|
|
raw_lockdep_assert_held_rcu_node(rnp_leaf);
|
rcu: Clean up handling of tasks blocked across full-rcu_node offline
Commit 0aa04b055e71 ("rcu: Process offlining and onlining only at
grace-period start") deferred handling of CPU-hotplug events until the
start of the next grace period, but consider the following sequence
of events:
1. A task is preempted within an RCU-preempt read-side critical
section.
2. The CPU that this task was running on goes offline, along with all
other CPUs sharing the corresponding leaf rcu_node structure.
3. The task resumes execution.
4. One of those CPUs comes back online before a new grace period starts.
In step 2, the code in the next rcu_gp_init() invocation will (correctly)
defer removing the leaf rcu_node structure from the upper-level bitmasks,
and will (correctly) set that structure's ->wait_blkd_tasks field. During
the ensuing interval, RCU will (correctly) track the tasks preempted on
that structure because they must block any subsequent grace period.
In step 3, the code in rcu_read_unlock_special() will (correctly) remove
the task from the leaf rcu_node structure. From this point forward, RCU
need not pay attention to this structure, at least not until one of the
corresponding CPUs comes back online.
In step 4, the code in the next rcu_gp_init() invocation will
(incorrectly) invoke rcu_init_new_rnp(). This is incorrect because
the corresponding rcu_cleanup_dead_rnp() was never invoked. This is
nevertheless harmless because the upper-level bits are still set.
So, no harm, no foul, right?
At least, all is well until a little further into rcu_gp_init()
invocation, which will notice that there are no longer any tasks blocked
on the leaf rcu_node structure, conclude that there is no longer anything
left over from step 2's offline operation, and will therefore invoke
rcu_cleanup_dead_rnp(). But this invocation of rcu_cleanup_dead_rnp()
is for the beginning of the earlier offline interval, and the previous
invocation of rcu_init_new_rnp() is for the end of that same interval.
That is right, they are invoked out of order.
That cannot be good, can it?
It turns out that this is not a (correctness!) problem because
rcu_cleanup_dead_rnp() checks to see if any of the corresponding CPUs
are online, and refuses to do anything if so. In other words, in the
case where rcu_init_new_rnp() and rcu_cleanup_dead_rnp() execute out of
order, they both have no effect.
But this is at best an accident waiting to happen.
This commit therefore adds logic to rcu_gp_init() so that
rcu_init_new_rnp() and rcu_cleanup_dead_rnp() are always invoked in
order, and so that neither are invoked at all in cases where RCU had to
pay attention to the leaf rcu_node structure during the entire time that
all corresponding CPUs were offline.
And, while in the area, this commit reduces confusion by using formal
parameters rather than local variables that just happen to have the same
value at that particular point in the code.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-05-03 03:49:21 +08:00
|
|
|
|
WARN_ON_ONCE(rnp->wait_blkd_tasks);
|
rcu: Process offlining and onlining only at grace-period start
Races between CPU hotplug and grace periods can be difficult to resolve,
so the ->onoff_mutex is used to exclude the two events. Unfortunately,
this means that it is impossible for an outgoing CPU to perform the
last bits of its offlining from its last pass through the idle loop,
because sleeplocks cannot be acquired in that context.
This commit avoids these problems by buffering online and offline events
in a new ->qsmaskinitnext field in the leaf rcu_node structures. When a
grace period starts, the events accumulated in this mask are applied to
the ->qsmaskinit field, and, if needed, up the rcu_node tree. The special
case of all CPUs corresponding to a given leaf rcu_node structure being
offline while there are still elements in that structure's ->blkd_tasks
list is handled using a new ->wait_blkd_tasks field. In this case,
propagating the offline bits up the tree is deferred until the beginning
of the grace period after all of the tasks have exited their RCU read-side
critical sections and removed themselves from the list, at which point
the ->wait_blkd_tasks flag is cleared. If one of that leaf rcu_node
structure's CPUs comes back online before the list empties, then the
->wait_blkd_tasks flag is simply cleared.
This of course means that RCU's notion of which CPUs are offline can be
out of date. This is OK because RCU need only wait on CPUs that were
online at the time that the grace period started. In addition, RCU's
force-quiescent-state actions will handle the case where a CPU goes
offline after the grace period starts.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2015-01-24 13:52:37 +08:00
|
|
|
|
for (;;) {
|
|
|
|
|
mask = rnp->grpmask;
|
|
|
|
|
rnp = rnp->parent;
|
|
|
|
|
if (rnp == NULL)
|
|
|
|
|
return;
|
2015-10-09 06:36:54 +08:00
|
|
|
|
raw_spin_lock_rcu_node(rnp); /* Interrupts already disabled. */
|
2018-05-03 05:46:43 +08:00
|
|
|
|
oldmask = rnp->qsmaskinit;
|
rcu: Process offlining and onlining only at grace-period start
Races between CPU hotplug and grace periods can be difficult to resolve,
so the ->onoff_mutex is used to exclude the two events. Unfortunately,
this means that it is impossible for an outgoing CPU to perform the
last bits of its offlining from its last pass through the idle loop,
because sleeplocks cannot be acquired in that context.
This commit avoids these problems by buffering online and offline events
in a new ->qsmaskinitnext field in the leaf rcu_node structures. When a
grace period starts, the events accumulated in this mask are applied to
the ->qsmaskinit field, and, if needed, up the rcu_node tree. The special
case of all CPUs corresponding to a given leaf rcu_node structure being
offline while there are still elements in that structure's ->blkd_tasks
list is handled using a new ->wait_blkd_tasks field. In this case,
propagating the offline bits up the tree is deferred until the beginning
of the grace period after all of the tasks have exited their RCU read-side
critical sections and removed themselves from the list, at which point
the ->wait_blkd_tasks flag is cleared. If one of that leaf rcu_node
structure's CPUs comes back online before the list empties, then the
->wait_blkd_tasks flag is simply cleared.
This of course means that RCU's notion of which CPUs are offline can be
out of date. This is OK because RCU need only wait on CPUs that were
online at the time that the grace period started. In addition, RCU's
force-quiescent-state actions will handle the case where a CPU goes
offline after the grace period starts.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2015-01-24 13:52:37 +08:00
|
|
|
|
rnp->qsmaskinit |= mask;
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_unlock_rcu_node(rnp); /* Interrupts remain disabled. */
|
2018-05-03 05:46:43 +08:00
|
|
|
|
if (oldmask)
|
|
|
|
|
return;
|
rcu: Process offlining and onlining only at grace-period start
Races between CPU hotplug and grace periods can be difficult to resolve,
so the ->onoff_mutex is used to exclude the two events. Unfortunately,
this means that it is impossible for an outgoing CPU to perform the
last bits of its offlining from its last pass through the idle loop,
because sleeplocks cannot be acquired in that context.
This commit avoids these problems by buffering online and offline events
in a new ->qsmaskinitnext field in the leaf rcu_node structures. When a
grace period starts, the events accumulated in this mask are applied to
the ->qsmaskinit field, and, if needed, up the rcu_node tree. The special
case of all CPUs corresponding to a given leaf rcu_node structure being
offline while there are still elements in that structure's ->blkd_tasks
list is handled using a new ->wait_blkd_tasks field. In this case,
propagating the offline bits up the tree is deferred until the beginning
of the grace period after all of the tasks have exited their RCU read-side
critical sections and removed themselves from the list, at which point
the ->wait_blkd_tasks flag is cleared. If one of that leaf rcu_node
structure's CPUs comes back online before the list empties, then the
->wait_blkd_tasks flag is simply cleared.
This of course means that RCU's notion of which CPUs are offline can be
out of date. This is OK because RCU need only wait on CPUs that were
online at the time that the grace period started. In addition, RCU's
force-quiescent-state actions will handle the case where a CPU goes
offline after the grace period starts.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2015-01-24 13:52:37 +08:00
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
/*
|
2009-08-16 00:53:46 +08:00
|
|
|
|
* Do boot-time initialization of a CPU's per-CPU RCU data.
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
*/
|
2009-08-16 00:53:46 +08:00
|
|
|
|
static void __init
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_boot_init_percpu_data(int cpu)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
{
|
2022-06-08 22:40:30 +08:00
|
|
|
|
struct context_tracking *ct = this_cpu_ptr(&context_tracking);
|
2018-07-04 06:37:16 +08:00
|
|
|
|
struct rcu_data *rdp = per_cpu_ptr(&rcu_data, cpu);
|
2009-08-16 00:53:46 +08:00
|
|
|
|
|
|
|
|
|
/* Set up local state, ensuring consistent view of global state. */
|
rcu: Correctly handle sparse possible cpus
In many cases in the RCU tree code, we iterate over the set of cpus for
a leaf node described by rcu_node::grplo and rcu_node::grphi, checking
per-cpu data for each cpu in this range. However, if the set of possible
cpus is sparse, some cpus described in this range are not possible, and
thus no per-cpu region will have been allocated (or initialised) for
them by the generic percpu code.
Erroneous accesses to a per-cpu area for these !possible cpus may fault
or may hit other data depending on the addressed generated when the
erroneous per cpu offset is applied. In practice, both cases have been
observed on arm64 hardware (the former being silent, but detectable with
additional patches).
To avoid issues resulting from this, we must iterate over the set of
*possible* cpus for a given leaf node. This patch add a new helper,
for_each_leaf_node_possible_cpu, to enable this. As iteration is often
intertwined with rcu_node local bitmask manipulation, a new
leaf_node_cpu_bit helper is added to make this simpler and more
consistent. The RCU tree code is made to use both of these where
appropriate.
Without this patch, running reboot at a shell can result in an oops
like:
[ 3369.075979] Unable to handle kernel paging request at virtual address ffffff8008b21b4c
[ 3369.083881] pgd = ffffffc3ecdda000
[ 3369.087270] [ffffff8008b21b4c] *pgd=00000083eca48003, *pud=00000083eca48003, *pmd=0000000000000000
[ 3369.096222] Internal error: Oops: 96000007 [#1] PREEMPT SMP
[ 3369.101781] Modules linked in:
[ 3369.104825] CPU: 2 PID: 1817 Comm: NetworkManager Tainted: G W 4.6.0+ #3
[ 3369.121239] task: ffffffc0fa13e000 ti: ffffffc3eb940000 task.ti: ffffffc3eb940000
[ 3369.128708] PC is at sync_rcu_exp_select_cpus+0x188/0x510
[ 3369.134094] LR is at sync_rcu_exp_select_cpus+0x104/0x510
[ 3369.139479] pc : [<ffffff80081109a8>] lr : [<ffffff8008110924>] pstate: 200001c5
[ 3369.146860] sp : ffffffc3eb9435a0
[ 3369.150162] x29: ffffffc3eb9435a0 x28: ffffff8008be4f88
[ 3369.155465] x27: ffffff8008b66c80 x26: ffffffc3eceb2600
[ 3369.160767] x25: 0000000000000001 x24: ffffff8008be4f88
[ 3369.166070] x23: ffffff8008b51c3c x22: ffffff8008b66c80
[ 3369.171371] x21: 0000000000000001 x20: ffffff8008b21b40
[ 3369.176673] x19: ffffff8008b66c80 x18: 0000000000000000
[ 3369.181975] x17: 0000007fa951a010 x16: ffffff80086a30f0
[ 3369.187278] x15: 0000007fa9505590 x14: 0000000000000000
[ 3369.192580] x13: ffffff8008b51000 x12: ffffffc3eb940000
[ 3369.197882] x11: 0000000000000006 x10: ffffff8008b51b78
[ 3369.203184] x9 : 0000000000000001 x8 : ffffff8008be4000
[ 3369.208486] x7 : ffffff8008b21b40 x6 : 0000000000001003
[ 3369.213788] x5 : 0000000000000000 x4 : ffffff8008b27280
[ 3369.219090] x3 : ffffff8008b21b4c x2 : 0000000000000001
[ 3369.224406] x1 : 0000000000000001 x0 : 0000000000000140
...
[ 3369.972257] [<ffffff80081109a8>] sync_rcu_exp_select_cpus+0x188/0x510
[ 3369.978685] [<ffffff80081128b4>] synchronize_rcu_expedited+0x64/0xa8
[ 3369.985026] [<ffffff80086b987c>] synchronize_net+0x24/0x30
[ 3369.990499] [<ffffff80086ddb54>] dev_deactivate_many+0x28c/0x298
[ 3369.996493] [<ffffff80086b6bb8>] __dev_close_many+0x60/0xd0
[ 3370.002052] [<ffffff80086b6d48>] __dev_close+0x28/0x40
[ 3370.007178] [<ffffff80086bf62c>] __dev_change_flags+0x8c/0x158
[ 3370.012999] [<ffffff80086bf718>] dev_change_flags+0x20/0x60
[ 3370.018558] [<ffffff80086cf7f0>] do_setlink+0x288/0x918
[ 3370.023771] [<ffffff80086d0798>] rtnl_newlink+0x398/0x6a8
[ 3370.029158] [<ffffff80086cee84>] rtnetlink_rcv_msg+0xe4/0x220
[ 3370.034891] [<ffffff80086e274c>] netlink_rcv_skb+0xc4/0xf8
[ 3370.040364] [<ffffff80086ced8c>] rtnetlink_rcv+0x2c/0x40
[ 3370.045663] [<ffffff80086e1fe8>] netlink_unicast+0x160/0x238
[ 3370.051309] [<ffffff80086e24b8>] netlink_sendmsg+0x2f0/0x358
[ 3370.056956] [<ffffff80086a0070>] sock_sendmsg+0x18/0x30
[ 3370.062168] [<ffffff80086a21cc>] ___sys_sendmsg+0x26c/0x280
[ 3370.067728] [<ffffff80086a30ac>] __sys_sendmsg+0x44/0x88
[ 3370.073027] [<ffffff80086a3100>] SyS_sendmsg+0x10/0x20
[ 3370.078153] [<ffffff8008085e70>] el0_svc_naked+0x24/0x28
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Reported-by: Dennis Chen <dennis.chen@arm.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Steve Capper <steve.capper@arm.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Will Deacon <will.deacon@arm.com>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2016-06-03 22:20:04 +08:00
|
|
|
|
rdp->grpmask = leaf_node_cpu_bit(rdp->mynode, cpu);
|
2020-08-08 22:56:31 +08:00
|
|
|
|
INIT_WORK(&rdp->strict_work, strict_work_handler);
|
2022-06-08 22:40:30 +08:00
|
|
|
|
WARN_ON_ONCE(ct->dynticks_nesting != 1);
|
2022-06-08 22:40:29 +08:00
|
|
|
|
WARN_ON_ONCE(rcu_dynticks_in_eqs(rcu_dynticks_snap(cpu)));
|
2021-12-15 05:15:18 +08:00
|
|
|
|
rdp->barrier_seq_snap = rcu_state.barrier_sequence;
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rdp->rcu_ofl_gp_seq = rcu_state.gp_seq;
|
2018-05-09 05:18:57 +08:00
|
|
|
|
rdp->rcu_ofl_gp_flags = RCU_GP_CLEANED;
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rdp->rcu_onl_gp_seq = rcu_state.gp_seq;
|
2018-05-09 05:18:57 +08:00
|
|
|
|
rdp->rcu_onl_gp_flags = RCU_GP_CLEANED;
|
2022-02-24 09:29:37 +08:00
|
|
|
|
rdp->last_sched_clock = jiffies;
|
2009-08-16 00:53:46 +08:00
|
|
|
|
rdp->cpu = cpu;
|
2012-08-20 12:35:53 +08:00
|
|
|
|
rcu_boot_init_nocb_percpu_data(rdp);
|
2009-08-16 00:53:46 +08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
2018-07-04 08:22:34 +08:00
|
|
|
|
* Invoked early in the CPU-online process, when pretty much all services
|
|
|
|
|
* are available. The incoming CPU is not present.
|
|
|
|
|
*
|
|
|
|
|
* Initializes a CPU's per-CPU RCU data. Note that only one online or
|
2018-05-02 05:34:08 +08:00
|
|
|
|
* offline event can be happening at a given time. Note also that we can
|
|
|
|
|
* accept some slop in the rsp->gp_seq access due to the fact that this
|
2019-05-15 00:50:49 +08:00
|
|
|
|
* CPU cannot possibly have any non-offloaded RCU callbacks in flight yet.
|
|
|
|
|
* And any offloaded callbacks are being numbered elsewhere.
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
*/
|
2018-07-04 08:22:34 +08:00
|
|
|
|
int rcutree_prepare_cpu(unsigned int cpu)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
{
|
|
|
|
|
unsigned long flags;
|
2022-06-08 22:40:30 +08:00
|
|
|
|
struct context_tracking *ct = per_cpu_ptr(&context_tracking, cpu);
|
2018-07-04 06:37:16 +08:00
|
|
|
|
struct rcu_data *rdp = per_cpu_ptr(&rcu_data, cpu);
|
2018-07-04 08:22:34 +08:00
|
|
|
|
struct rcu_node *rnp = rcu_get_root();
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
|
|
|
|
/* Set up local state, ensuring consistent view of global state. */
|
2015-10-09 06:36:54 +08:00
|
|
|
|
raw_spin_lock_irqsave_rcu_node(rnp, flags);
|
2009-10-15 01:15:55 +08:00
|
|
|
|
rdp->qlen_last_fqs_check = 0;
|
2021-07-20 21:16:27 +08:00
|
|
|
|
rdp->n_force_qs_snap = READ_ONCE(rcu_state.n_force_qs);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
rdp->blimit = blimit;
|
2022-06-08 22:40:30 +08:00
|
|
|
|
ct->dynticks_nesting = 1; /* CPU not up, no tearing. */
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_unlock_rcu_node(rnp); /* irqs remain disabled. */
|
rcu/nocb: Only (re-)initialize segcblist when needed on CPU up
At the start of a CPU-hotplug operation, the incoming CPU's callback
list can be in a number of states:
1. Disabled and empty. This is the case when the boot CPU has
not invoked call_rcu(), when a non-boot CPU first comes online,
and when a non-offloaded CPU comes back online. In this case,
it is both necessary and permissible to initialize ->cblist.
Because either the CPU is currently running with interrupts
disabled (boot CPU) or is not yet running at all (other CPUs),
it is not necessary to acquire ->nocb_lock.
In this case, initialization is required.
2. Disabled and non-empty. This cannot occur, because early boot
call_rcu() invocations enable the callback list before enqueuing
their callback.
3. Enabled, whether empty or not. In this case, the callback
list has already been initialized. This case occurs when the
boot CPU has executed an early boot call_rcu() and also when
an offloaded CPU comes back online. In both cases, there is
no need to initialize the callback list: In the boot-CPU case,
the CPU has not (yet) gone offline, and in the offloaded case,
the rcuo kthreads are taking care of business.
Because it is not necessary to initialize the callback list,
it is also not necessary to acquire ->nocb_lock.
Therefore, checking if the segcblist is enabled suffices. This commit
therefore initializes the callback list at rcutree_prepare_cpu() time
only if that list is disabled.
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: Joel Fernandes <joel@joelfernandes.org>
Cc: Neeraj Upadhyay <neeraju@codeaurora.org>
Cc: Boqun Feng <boqun.feng@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-01-29 01:12:10 +08:00
|
|
|
|
|
2020-11-13 20:13:18 +08:00
|
|
|
|
/*
|
rcu/nocb: Only (re-)initialize segcblist when needed on CPU up
At the start of a CPU-hotplug operation, the incoming CPU's callback
list can be in a number of states:
1. Disabled and empty. This is the case when the boot CPU has
not invoked call_rcu(), when a non-boot CPU first comes online,
and when a non-offloaded CPU comes back online. In this case,
it is both necessary and permissible to initialize ->cblist.
Because either the CPU is currently running with interrupts
disabled (boot CPU) or is not yet running at all (other CPUs),
it is not necessary to acquire ->nocb_lock.
In this case, initialization is required.
2. Disabled and non-empty. This cannot occur, because early boot
call_rcu() invocations enable the callback list before enqueuing
their callback.
3. Enabled, whether empty or not. In this case, the callback
list has already been initialized. This case occurs when the
boot CPU has executed an early boot call_rcu() and also when
an offloaded CPU comes back online. In both cases, there is
no need to initialize the callback list: In the boot-CPU case,
the CPU has not (yet) gone offline, and in the offloaded case,
the rcuo kthreads are taking care of business.
Because it is not necessary to initialize the callback list,
it is also not necessary to acquire ->nocb_lock.
Therefore, checking if the segcblist is enabled suffices. This commit
therefore initializes the callback list at rcutree_prepare_cpu() time
only if that list is disabled.
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: Joel Fernandes <joel@joelfernandes.org>
Cc: Neeraj Upadhyay <neeraju@codeaurora.org>
Cc: Boqun Feng <boqun.feng@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-01-29 01:12:10 +08:00
|
|
|
|
* Only non-NOCB CPUs that didn't have early-boot callbacks need to be
|
|
|
|
|
* (re-)initialized.
|
2020-11-13 20:13:18 +08:00
|
|
|
|
*/
|
rcu/nocb: Only (re-)initialize segcblist when needed on CPU up
At the start of a CPU-hotplug operation, the incoming CPU's callback
list can be in a number of states:
1. Disabled and empty. This is the case when the boot CPU has
not invoked call_rcu(), when a non-boot CPU first comes online,
and when a non-offloaded CPU comes back online. In this case,
it is both necessary and permissible to initialize ->cblist.
Because either the CPU is currently running with interrupts
disabled (boot CPU) or is not yet running at all (other CPUs),
it is not necessary to acquire ->nocb_lock.
In this case, initialization is required.
2. Disabled and non-empty. This cannot occur, because early boot
call_rcu() invocations enable the callback list before enqueuing
their callback.
3. Enabled, whether empty or not. In this case, the callback
list has already been initialized. This case occurs when the
boot CPU has executed an early boot call_rcu() and also when
an offloaded CPU comes back online. In both cases, there is
no need to initialize the callback list: In the boot-CPU case,
the CPU has not (yet) gone offline, and in the offloaded case,
the rcuo kthreads are taking care of business.
Because it is not necessary to initialize the callback list,
it is also not necessary to acquire ->nocb_lock.
Therefore, checking if the segcblist is enabled suffices. This commit
therefore initializes the callback list at rcutree_prepare_cpu() time
only if that list is disabled.
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: Joel Fernandes <joel@joelfernandes.org>
Cc: Neeraj Upadhyay <neeraju@codeaurora.org>
Cc: Boqun Feng <boqun.feng@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-01-29 01:12:10 +08:00
|
|
|
|
if (!rcu_segcblist_is_enabled(&rdp->cblist))
|
2020-11-13 20:13:18 +08:00
|
|
|
|
rcu_segcblist_init(&rdp->cblist); /* Re-enable callbacks. */
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
rcu: Process offlining and onlining only at grace-period start
Races between CPU hotplug and grace periods can be difficult to resolve,
so the ->onoff_mutex is used to exclude the two events. Unfortunately,
this means that it is impossible for an outgoing CPU to perform the
last bits of its offlining from its last pass through the idle loop,
because sleeplocks cannot be acquired in that context.
This commit avoids these problems by buffering online and offline events
in a new ->qsmaskinitnext field in the leaf rcu_node structures. When a
grace period starts, the events accumulated in this mask are applied to
the ->qsmaskinit field, and, if needed, up the rcu_node tree. The special
case of all CPUs corresponding to a given leaf rcu_node structure being
offline while there are still elements in that structure's ->blkd_tasks
list is handled using a new ->wait_blkd_tasks field. In this case,
propagating the offline bits up the tree is deferred until the beginning
of the grace period after all of the tasks have exited their RCU read-side
critical sections and removed themselves from the list, at which point
the ->wait_blkd_tasks flag is cleared. If one of that leaf rcu_node
structure's CPUs comes back online before the list empties, then the
->wait_blkd_tasks flag is simply cleared.
This of course means that RCU's notion of which CPUs are offline can be
out of date. This is OK because RCU need only wait on CPUs that were
online at the time that the grace period started. In addition, RCU's
force-quiescent-state actions will handle the case where a CPU goes
offline after the grace period starts.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2015-01-24 13:52:37 +08:00
|
|
|
|
/*
|
|
|
|
|
* Add CPU to leaf rcu_node pending-online bitmask. Any needed
|
|
|
|
|
* propagation up the rcu_node tree will happen at the beginning
|
|
|
|
|
* of the next grace period.
|
|
|
|
|
*/
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
rnp = rdp->mynode;
|
2015-10-08 18:24:23 +08:00
|
|
|
|
raw_spin_lock_rcu_node(rnp); /* irqs already disabled. */
|
2020-01-05 03:33:17 +08:00
|
|
|
|
rdp->gp_seq = READ_ONCE(rnp->gp_seq);
|
|
|
|
|
rdp->gp_seq_needed = rdp->gp_seq;
|
2015-08-07 06:16:57 +08:00
|
|
|
|
rdp->cpu_no_qs.b.norm = true;
|
2015-08-07 02:31:51 +08:00
|
|
|
|
rdp->core_needs_qs = false;
|
2017-08-18 08:05:59 +08:00
|
|
|
|
rdp->rcu_iw_pending = false;
|
2020-06-15 17:51:29 +08:00
|
|
|
|
rdp->rcu_iw = IRQ_WORK_INIT_HARD(rcu_iw_handler);
|
2020-01-05 03:33:17 +08:00
|
|
|
|
rdp->rcu_iw_gp_seq = rdp->gp_seq - 1;
|
2018-07-04 08:22:34 +08:00
|
|
|
|
trace_rcu_grace_period(rcu_state.name, rdp->gp_seq, TPS("cpuonl"));
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
|
rcu: Make RCU priority boosting work on single-CPU rcu_node structures
When any CPU comes online, it checks to see if an RCU-boost kthread has
already been created for that CPU's leaf rcu_node structure, and if
not, it creates one. Unfortunately, it also verifies that this leaf
rcu_node structure actually has at least one online CPU, and if not,
it declines to create the kthread. Although this behavior makes sense
during early boot, especially on systems that claim far more CPUs than
they actually have, it makes no sense for the first CPU to come online
for a given rcu_node structure. There is no point in checking because
we know there is a CPU on its way in.
The problem is that timing differences can cause this incoming CPU to not
yet be reflected in the various bit masks even at rcutree_online_cpu()
time, and there is no chance at rcutree_prepare_cpu() time. Plus it
would be better to create the RCU-boost kthread at rcutree_prepare_cpu()
to handle the case where the CPU is involved in an RCU priority inversion
very shortly after it comes online.
This commit therefore moves the checking to rcu_prepare_kthreads(), which
is called only at early boot, when the check is appropriate. In addition,
it makes rcutree_prepare_cpu() invoke rcu_spawn_one_boost_kthread(), which
no longer does any checking for online CPUs.
With this change, RCU priority boosting tests now pass for short rcutorture
runs, even with single-CPU leaf rcu_node structures.
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Scott Wood <swood@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-04-06 11:42:09 +08:00
|
|
|
|
rcu_spawn_one_boost_kthread(rnp);
|
2018-11-28 05:55:53 +08:00
|
|
|
|
rcu_spawn_cpu_nocb_kthread(cpu);
|
rcu: Fix single-CPU check in rcu_blocking_is_gp()
Currently, for CONFIG_PREEMPTION=n kernels, rcu_blocking_is_gp() uses
num_online_cpus() to determine whether there is only one CPU online. When
there is only a single CPU online, the simple fact that synchronize_rcu()
could be legally called implies that a full grace period has elapsed.
Therefore, in the single-CPU case, synchronize_rcu() simply returns
immediately. Unfortunately, num_online_cpus() is unreliable while a
CPU-hotplug operation is transitioning to or from single-CPU operation
because:
1. num_online_cpus() uses atomic_read(&__num_online_cpus) to
locklessly sample the number of online CPUs. The hotplug locks
are not held, which means that an incoming CPU can concurrently
update this count. This in turn means that an RCU read-side
critical section on the incoming CPU might observe updates
prior to the grace period, but also that this critical section
might extend beyond the end of the optimized synchronize_rcu().
This breaks RCU's fundamental guarantee.
2. In addition, num_online_cpus() does no ordering, thus providing
another way that RCU's fundamental guarantee can be broken by
the current code.
3. The most probable failure mode happens on outgoing CPUs.
The outgoing CPU updates the count of online CPUs in the
CPUHP_TEARDOWN_CPU stop-machine handler, which is fine in
and of itself due to preemption being disabled at the call
to num_online_cpus(). Unfortunately, after that stop-machine
handler returns, the CPU takes one last trip through the
scheduler (which has RCU readers) and, after the resulting
context switch, one final dive into the idle loop. During this
time, RCU needs to keep track of two CPUs, but num_online_cpus()
will say that there is only one, which in turn means that the
surviving CPU will incorrectly ignore the outgoing CPU's RCU
read-side critical sections.
This problem is illustrated by the following litmus test in which P0()
corresponds to synchronize_rcu() and P1() corresponds to the incoming CPU.
The herd7 tool confirms that the "exists" clause can be satisfied,
thus demonstrating that this breakage can happen according to the Linux
kernel memory model.
{
int x = 0;
atomic_t numonline = ATOMIC_INIT(1);
}
P0(int *x, atomic_t *numonline)
{
int r0;
WRITE_ONCE(*x, 1);
r0 = atomic_read(numonline);
if (r0 == 1) {
smp_mb();
} else {
synchronize_rcu();
}
WRITE_ONCE(*x, 2);
}
P1(int *x, atomic_t *numonline)
{
int r0; int r1;
atomic_inc(numonline);
smp_mb();
rcu_read_lock();
r0 = READ_ONCE(*x);
smp_rmb();
r1 = READ_ONCE(*x);
rcu_read_unlock();
}
locations [x;numonline;]
exists (1:r0=0 /\ 1:r1=2)
It is important to note that these problems arise only when the system
is transitioning to or from single-CPU operation.
One solution would be to hold the CPU-hotplug locks while sampling
num_online_cpus(), which was in fact the intent of the (redundant)
preempt_disable() and preempt_enable() surrounding this call to
num_online_cpus(). Actually blocking CPU hotplug would not only result
in excessive overhead, but would also unnecessarily impede CPU-hotplug
operations.
This commit therefore follows long-standing RCU tradition by maintaining
a separate RCU-specific set of CPU-hotplug books.
This separate set of books is implemented by a new ->n_online_cpus field
in the rcu_state structure that maintains RCU's count of the online CPUs.
This count is incremented early in the CPU-online process, so that
the critical transition away from single-CPU operation will occur when
there is only a single CPU. Similarly for the critical transition to
single-CPU operation, the counter is decremented late in the CPU-offline
process, again while there is only a single CPU. Because there is only
ever a single CPU when the ->n_online_cpus field undergoes the critical
1->2 and 2->1 transitions, full memory ordering and mutual exclusion is
provided implicitly and, better yet, for free.
In the case where the CPU is coming online, nothing will happen until
the current CPU helps it come online. Therefore, the new CPU will see
all accesses prior to the optimized grace period, which means that RCU
does not need to further delay this new CPU. In the case where the CPU
is going offline, the outgoing CPU is totally out of the picture before
the optimized grace period starts, which means that this outgoing CPU
cannot see any of the accesses following that grace period. Again,
RCU needs no further interaction with the outgoing CPU.
This does mean that synchronize_rcu() will unnecessarily do a few grace
periods the hard way just before the second CPU comes online and just
after the second-to-last CPU goes offline, but it is not worth optimizing
this uncommon case.
Reviewed-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Neeraj Upadhyay <neeraju@codeaurora.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-09-23 15:29:33 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.n_online_cpus, rcu_state.n_online_cpus + 1);
|
2016-07-14 01:17:03 +08:00
|
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2017-03-24 04:21:30 +08:00
|
|
|
|
/*
|
|
|
|
|
* Update RCU priority boot kthread affinity for CPU-hotplug changes.
|
|
|
|
|
*/
|
2016-07-14 01:17:03 +08:00
|
|
|
|
static void rcutree_affinity_setting(unsigned int cpu, int outgoing)
|
|
|
|
|
{
|
2018-07-04 06:37:16 +08:00
|
|
|
|
struct rcu_data *rdp = per_cpu_ptr(&rcu_data, cpu);
|
2016-07-14 01:17:03 +08:00
|
|
|
|
|
|
|
|
|
rcu_boost_kthread_setaffinity(rdp->mynode, outgoing);
|
|
|
|
|
}
|
|
|
|
|
|
rcu-tasks: Stop rcu_tasks_invoke_cbs() from using never-onlined CPUs
The rcu_tasks_invoke_cbs() function relies on queue_work_on() to silently
fall back to WORK_CPU_UNBOUND when the specified CPU is offline. However,
the queue_work_on() function's silent fallback mechanism relies on that
CPU having been online at some time in the past. When queue_work_on()
is passed a CPU that has never been online, workqueue lockups ensue,
which can be bad for your kernel's general health and well-being.
This commit therefore checks whether a given CPU has ever been online,
and, if not substitutes WORK_CPU_UNBOUND in the subsequent call to
queue_work_on(). Why not simply omit the queue_work_on() call entirely?
Because this function is flooding callback-invocation notifications
to all CPUs, and must deal with possibilities that include a sparse
cpu_possible_mask.
This commit also moves the setting of the rcu_data structure's
->beenonline field to rcu_cpu_starting(), which executes on the
incoming CPU before that CPU has ever enabled interrupts. This ensures
that the required workqueues are present. In addition, because the
incoming CPU has not yet enabled its interrupts, there cannot yet have
been any softirq handlers running on this CPU, which means that the
WARN_ON_ONCE(!rdp->beenonline) within the RCU_SOFTIRQ handler cannot
have triggered yet.
Fixes: d363f833c6d88 ("rcu-tasks: Use workqueues for multiple rcu_tasks_invoke_cbs() invocations")
Reported-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2023-04-27 02:11:29 +08:00
|
|
|
|
/*
|
|
|
|
|
* Has the specified (known valid) CPU ever been fully online?
|
|
|
|
|
*/
|
|
|
|
|
bool rcu_cpu_beenfullyonline(int cpu)
|
|
|
|
|
{
|
|
|
|
|
struct rcu_data *rdp = per_cpu_ptr(&rcu_data, cpu);
|
|
|
|
|
|
|
|
|
|
return smp_load_acquire(&rdp->beenonline);
|
|
|
|
|
}
|
|
|
|
|
|
2017-03-24 04:21:30 +08:00
|
|
|
|
/*
|
|
|
|
|
* Near the end of the CPU-online process. Pretty much all services
|
|
|
|
|
* enabled, and the CPU is now very much alive.
|
|
|
|
|
*/
|
2016-07-14 01:17:03 +08:00
|
|
|
|
int rcutree_online_cpu(unsigned int cpu)
|
|
|
|
|
{
|
2017-08-18 08:05:59 +08:00
|
|
|
|
unsigned long flags;
|
|
|
|
|
struct rcu_data *rdp;
|
|
|
|
|
struct rcu_node *rnp;
|
|
|
|
|
|
2018-07-05 06:35:00 +08:00
|
|
|
|
rdp = per_cpu_ptr(&rcu_data, cpu);
|
|
|
|
|
rnp = rdp->mynode;
|
|
|
|
|
raw_spin_lock_irqsave_rcu_node(rnp, flags);
|
|
|
|
|
rnp->ffmask |= rdp->grpmask;
|
|
|
|
|
raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
|
2017-08-18 08:05:59 +08:00
|
|
|
|
if (rcu_scheduler_active == RCU_SCHEDULER_INACTIVE)
|
|
|
|
|
return 0; /* Too early in boot for scheduler work. */
|
|
|
|
|
sync_sched_exp_online_cleanup(cpu);
|
|
|
|
|
rcutree_affinity_setting(cpu, -1);
|
2019-08-03 06:12:47 +08:00
|
|
|
|
|
|
|
|
|
// Stop-machine done, so allow nohz_full to disable tick.
|
|
|
|
|
tick_dep_clear(TICK_DEP_BIT_RCU);
|
2016-07-14 01:17:03 +08:00
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2017-03-24 04:21:30 +08:00
|
|
|
|
/*
|
|
|
|
|
* Near the beginning of the process. The CPU is still very much alive
|
|
|
|
|
* with pretty much all services enabled.
|
|
|
|
|
*/
|
2016-07-14 01:17:03 +08:00
|
|
|
|
int rcutree_offline_cpu(unsigned int cpu)
|
|
|
|
|
{
|
2017-08-18 08:05:59 +08:00
|
|
|
|
unsigned long flags;
|
|
|
|
|
struct rcu_data *rdp;
|
|
|
|
|
struct rcu_node *rnp;
|
|
|
|
|
|
2018-07-05 06:35:00 +08:00
|
|
|
|
rdp = per_cpu_ptr(&rcu_data, cpu);
|
|
|
|
|
rnp = rdp->mynode;
|
|
|
|
|
raw_spin_lock_irqsave_rcu_node(rnp, flags);
|
|
|
|
|
rnp->ffmask &= ~rdp->grpmask;
|
|
|
|
|
raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
|
2017-08-18 08:05:59 +08:00
|
|
|
|
|
2016-07-14 01:17:03 +08:00
|
|
|
|
rcutree_affinity_setting(cpu, cpu);
|
2019-08-03 06:12:47 +08:00
|
|
|
|
|
|
|
|
|
// nohz_full CPUs need the tick for stop-machine to work quickly
|
|
|
|
|
tick_dep_set(TICK_DEP_BIT_RCU);
|
2016-07-14 01:17:03 +08:00
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2016-07-01 04:58:26 +08:00
|
|
|
|
/*
|
|
|
|
|
* Mark the specified CPU as being online so that subsequent grace periods
|
|
|
|
|
* (both expedited and normal) will wait on it. Note that this means that
|
|
|
|
|
* incoming CPUs are not allowed to use RCU read-side critical sections
|
|
|
|
|
* until this function is called. Failing to observe this restriction
|
|
|
|
|
* will result in lockdep splats.
|
2017-03-24 04:21:30 +08:00
|
|
|
|
*
|
|
|
|
|
* Note that this function is special in that it is invoked directly
|
|
|
|
|
* from the incoming CPU rather than from the cpuhp_step mechanism.
|
|
|
|
|
* This is because this function must be invoked at a precise location.
|
2023-04-28 01:50:47 +08:00
|
|
|
|
* This incoming CPU must not have enabled interrupts yet.
|
2016-07-01 04:58:26 +08:00
|
|
|
|
*/
|
|
|
|
|
void rcu_cpu_starting(unsigned int cpu)
|
|
|
|
|
{
|
|
|
|
|
unsigned long mask;
|
|
|
|
|
struct rcu_data *rdp;
|
|
|
|
|
struct rcu_node *rnp;
|
2020-04-20 05:57:15 +08:00
|
|
|
|
bool newcpu;
|
2016-07-01 04:58:26 +08:00
|
|
|
|
|
2023-04-28 01:50:47 +08:00
|
|
|
|
lockdep_assert_irqs_disabled();
|
2020-07-25 11:22:05 +08:00
|
|
|
|
rdp = per_cpu_ptr(&rcu_data, cpu);
|
|
|
|
|
if (rdp->cpu_started)
|
2018-05-23 00:50:53 +08:00
|
|
|
|
return;
|
2020-07-25 11:22:05 +08:00
|
|
|
|
rdp->cpu_started = true;
|
2018-05-23 00:50:53 +08:00
|
|
|
|
|
2018-07-05 06:35:00 +08:00
|
|
|
|
rnp = rdp->mynode;
|
|
|
|
|
mask = rdp->grpmask;
|
2021-02-16 23:04:34 +08:00
|
|
|
|
arch_spin_lock(&rcu_state.ofl_lock);
|
rcu: Move rcu_dynticks_eqs_online() to rcu_cpu_starting()
The purpose of rcu_dynticks_eqs_online() is to adjust the ->dynticks
counter of an incoming CPU when required. It is currently invoked
from rcutree_prepare_cpu(), which runs before the incoming CPU is
running, and thus on some other CPU. This makes the per-CPU accesses in
rcu_dynticks_eqs_online() iffy at best, and it all "works" only because
the running CPU cannot possibly be in dyntick-idle mode, which means
that rcu_dynticks_eqs_online() never has any effect.
It is currently OK for rcu_dynticks_eqs_online() to have no effect, but
only because the CPU-offline process just happens to leave ->dynticks in
the correct state. After all, if ->dynticks were in the wrong state on a
just-onlined CPU, rcutorture would complain bitterly the next time that
CPU went idle, at least in kernels built with CONFIG_RCU_EQS_DEBUG=y,
for example, those built by rcutorture scenario TREE04. One could
argue that this means that rcu_dynticks_eqs_online() is unnecessary,
however, removing it would make the CPU-online process vulnerable to
slight changes in the CPU-offline process.
One could also ask why it is safe to move the rcu_dynticks_eqs_online()
call so late in the CPU-online process. Indeed, there was a time when it
would not have been safe, which does much to explain its current location.
However, the marking of a CPU as online from an RCU perspective has long
since moved from rcutree_prepare_cpu() to rcu_cpu_starting(), and all
that is required is that ->dynticks be set correctly by the time that
the CPU is marked as online from an RCU perspective. After all, the RCU
grace-period kthread does not check to see if offline CPUs are also idle.
(In case you were curious, this is one reason why there is quiescent-state
reporting as part of the offlining process.)
This commit therefore moves the call to rcu_dynticks_eqs_online() from
rcutree_prepare_cpu() to rcu_cpu_starting(), this latter being guaranteed
to be running on the incoming CPU. The call to this function must of
course be placed before this rcu_cpu_starting() announces this CPU's
presence to RCU.
Reported-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-07-29 03:38:42 +08:00
|
|
|
|
rcu_dynticks_eqs_online();
|
rcu: Make rcu_barrier() no longer block CPU-hotplug operations
This commit removes the cpus_read_lock() and cpus_read_unlock() calls
from rcu_barrier(), thus allowing CPUs to come and go during the course
of rcu_barrier() execution. Posting of the ->barrier_head callbacks does
synchronize with portions of RCU's CPU-hotplug notifiers, but these locks
are held for short time periods on both sides. Thus, full CPU-hotplug
operations could both start and finish during the execution of a given
rcu_barrier() invocation.
Additional synchronization is provided by a global ->barrier_lock.
Since the ->barrier_lock is only used during rcu_barrier() execution and
during onlining/offlining a CPU, the contention for this lock should
be low. It might be tempting to make use of a per-CPU lock just on
general principles, but straightforward attempts to do this have the
problems shown below.
Initial state: 3 CPUs present, CPU 0 and CPU1 do not have
any callback and CPU2 has callbacks.
1. CPU0 calls rcu_barrier().
2. CPU1 starts offlining for CPU2. CPU1 calls
rcutree_migrate_callbacks(). rcu_barrier_entrain() is called
from rcutree_migrate_callbacks(), with CPU2's rdp->barrier_lock.
It does not entrain ->barrier_head for CPU2, as rcu_barrier()
on CPU0 hasn't started the barrier sequence (by calling
rcu_seq_start(&rcu_state.barrier_sequence)) yet.
3. CPU0 starts new barrier sequence. It iterates over
CPU0 and CPU1, after acquiring their per-cpu ->barrier_lock
and finds 0 segcblist length. It updates ->barrier_seq_snap
for CPU0 and CPU1 and continues loop iteration to CPU2.
for_each_possible_cpu(cpu) {
raw_spin_lock_irqsave(&rdp->barrier_lock, flags);
if (!rcu_segcblist_n_cbs(&rdp->cblist)) {
WRITE_ONCE(rdp->barrier_seq_snap, gseq);
raw_spin_unlock_irqrestore(&rdp->barrier_lock, flags);
rcu_barrier_trace(TPS("NQ"), cpu, rcu_state.barrier_sequence);
continue;
}
4. rcutree_migrate_callbacks() completes execution on CPU1.
Segcblist len for CPU2 becomes 0.
5. The loop iteration on CPU0, checks rcu_segcblist_n_cbs(&rdp->cblist)
for CPU2 and completes the loop iteration after setting
->barrier_seq_snap.
6. As there isn't any ->barrier_head callback entrained; at
this point, rcu_barrier() in CPU0 returns.
7. The callbacks, which migrated from CPU2 to CPU1, execute.
Straightforward per-CPU locking is also subject to the following race
condition noted by Boqun Feng:
1. CPU0 calls rcu_barrier(), starting a new barrier sequence by invoking
rcu_seq_start() and init_completion(), but does not yet initialize
rcu_state.barrier_cpu_count.
2. CPU1 starts offlining for CPU2, calling rcutree_migrate_callbacks(),
which in turn calls rcu_barrier_entrain() holding CPU2's.
rdp->barrier_lock. It then entrains ->barrier_head for CPU2
and atomically increments rcu_state.barrier_cpu_count, which is
unfortunately not yet initialized to the value 2.
3. The just-entrained RCU callback is invoked. It atomically
decrements rcu_state.barrier_cpu_count and sees that it is
now zero. This callback therefore invokes complete().
4. CPU0 continues executing rcu_barrier(), but is not blocked
by its call to wait_for_completion(). This results in rcu_barrier()
returning before all pre-existing callbacks have been invoked,
which is a bug.
Therefore, synchronization is provided by rcu_state.barrier_lock,
which is also held across the initialization sequence, especially the
rcu_seq_start() and the atomic_set() that sets rcu_state.barrier_cpu_count
to the value 2. In addition, this lock is held when entraining the
rcu_barrier() callback, when deciding whether or not a CPU has callbacks
that rcu_barrier() must wait on, when setting the ->qsmaskinitnext for
incoming CPUs, and when migrating callbacks from a CPU that is going
offline.
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Co-developed-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-12-15 05:35:17 +08:00
|
|
|
|
raw_spin_lock(&rcu_state.barrier_lock);
|
2021-02-16 23:04:34 +08:00
|
|
|
|
raw_spin_lock_rcu_node(rnp);
|
2020-01-04 07:44:23 +08:00
|
|
|
|
WRITE_ONCE(rnp->qsmaskinitnext, rnp->qsmaskinitnext | mask);
|
rcu: Make rcu_barrier() no longer block CPU-hotplug operations
This commit removes the cpus_read_lock() and cpus_read_unlock() calls
from rcu_barrier(), thus allowing CPUs to come and go during the course
of rcu_barrier() execution. Posting of the ->barrier_head callbacks does
synchronize with portions of RCU's CPU-hotplug notifiers, but these locks
are held for short time periods on both sides. Thus, full CPU-hotplug
operations could both start and finish during the execution of a given
rcu_barrier() invocation.
Additional synchronization is provided by a global ->barrier_lock.
Since the ->barrier_lock is only used during rcu_barrier() execution and
during onlining/offlining a CPU, the contention for this lock should
be low. It might be tempting to make use of a per-CPU lock just on
general principles, but straightforward attempts to do this have the
problems shown below.
Initial state: 3 CPUs present, CPU 0 and CPU1 do not have
any callback and CPU2 has callbacks.
1. CPU0 calls rcu_barrier().
2. CPU1 starts offlining for CPU2. CPU1 calls
rcutree_migrate_callbacks(). rcu_barrier_entrain() is called
from rcutree_migrate_callbacks(), with CPU2's rdp->barrier_lock.
It does not entrain ->barrier_head for CPU2, as rcu_barrier()
on CPU0 hasn't started the barrier sequence (by calling
rcu_seq_start(&rcu_state.barrier_sequence)) yet.
3. CPU0 starts new barrier sequence. It iterates over
CPU0 and CPU1, after acquiring their per-cpu ->barrier_lock
and finds 0 segcblist length. It updates ->barrier_seq_snap
for CPU0 and CPU1 and continues loop iteration to CPU2.
for_each_possible_cpu(cpu) {
raw_spin_lock_irqsave(&rdp->barrier_lock, flags);
if (!rcu_segcblist_n_cbs(&rdp->cblist)) {
WRITE_ONCE(rdp->barrier_seq_snap, gseq);
raw_spin_unlock_irqrestore(&rdp->barrier_lock, flags);
rcu_barrier_trace(TPS("NQ"), cpu, rcu_state.barrier_sequence);
continue;
}
4. rcutree_migrate_callbacks() completes execution on CPU1.
Segcblist len for CPU2 becomes 0.
5. The loop iteration on CPU0, checks rcu_segcblist_n_cbs(&rdp->cblist)
for CPU2 and completes the loop iteration after setting
->barrier_seq_snap.
6. As there isn't any ->barrier_head callback entrained; at
this point, rcu_barrier() in CPU0 returns.
7. The callbacks, which migrated from CPU2 to CPU1, execute.
Straightforward per-CPU locking is also subject to the following race
condition noted by Boqun Feng:
1. CPU0 calls rcu_barrier(), starting a new barrier sequence by invoking
rcu_seq_start() and init_completion(), but does not yet initialize
rcu_state.barrier_cpu_count.
2. CPU1 starts offlining for CPU2, calling rcutree_migrate_callbacks(),
which in turn calls rcu_barrier_entrain() holding CPU2's.
rdp->barrier_lock. It then entrains ->barrier_head for CPU2
and atomically increments rcu_state.barrier_cpu_count, which is
unfortunately not yet initialized to the value 2.
3. The just-entrained RCU callback is invoked. It atomically
decrements rcu_state.barrier_cpu_count and sees that it is
now zero. This callback therefore invokes complete().
4. CPU0 continues executing rcu_barrier(), but is not blocked
by its call to wait_for_completion(). This results in rcu_barrier()
returning before all pre-existing callbacks have been invoked,
which is a bug.
Therefore, synchronization is provided by rcu_state.barrier_lock,
which is also held across the initialization sequence, especially the
rcu_seq_start() and the atomic_set() that sets rcu_state.barrier_cpu_count
to the value 2. In addition, this lock is held when entraining the
rcu_barrier() callback, when deciding whether or not a CPU has callbacks
that rcu_barrier() must wait on, when setting the ->qsmaskinitnext for
incoming CPUs, and when migrating callbacks from a CPU that is going
offline.
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Co-developed-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-12-15 05:35:17 +08:00
|
|
|
|
raw_spin_unlock(&rcu_state.barrier_lock);
|
2020-04-20 05:57:15 +08:00
|
|
|
|
newcpu = !(rnp->expmaskinitnext & mask);
|
2018-07-05 06:35:00 +08:00
|
|
|
|
rnp->expmaskinitnext |= mask;
|
|
|
|
|
/* Allow lockless access for expedited grace periods. */
|
2020-04-20 05:57:15 +08:00
|
|
|
|
smp_store_release(&rcu_state.ncpus, rcu_state.ncpus + newcpu); /* ^^^ */
|
2020-02-10 21:29:58 +08:00
|
|
|
|
ASSERT_EXCLUSIVE_WRITER(rcu_state.ncpus);
|
2018-07-05 06:35:00 +08:00
|
|
|
|
rcu_gpnum_ovf(rnp, rdp); /* Offline-induced counter wrap? */
|
2018-07-06 08:47:45 +08:00
|
|
|
|
rdp->rcu_onl_gp_seq = READ_ONCE(rcu_state.gp_seq);
|
|
|
|
|
rdp->rcu_onl_gp_flags = READ_ONCE(rcu_state.gp_flags);
|
2020-09-30 03:29:27 +08:00
|
|
|
|
|
|
|
|
|
/* An incoming CPU should never be blocking a grace period. */
|
|
|
|
|
if (WARN_ON_ONCE(rnp->qsmask & mask)) { /* RCU waiting on incoming CPU? */
|
2021-02-16 23:04:34 +08:00
|
|
|
|
/* rcu_report_qs_rnp() *really* wants some flags to restore */
|
2023-04-28 01:50:47 +08:00
|
|
|
|
unsigned long flags;
|
2021-02-16 23:04:34 +08:00
|
|
|
|
|
2023-04-28 01:50:47 +08:00
|
|
|
|
local_irq_save(flags);
|
2019-09-06 01:26:41 +08:00
|
|
|
|
rcu_disable_urgency_upon_qs(rdp);
|
2018-07-05 06:35:00 +08:00
|
|
|
|
/* Report QS -after- changing ->qsmaskinitnext! */
|
2023-04-28 01:50:47 +08:00
|
|
|
|
rcu_report_qs_rnp(mask, rnp, rnp->gp_seq, flags);
|
2018-07-05 06:35:00 +08:00
|
|
|
|
} else {
|
2021-02-16 23:04:34 +08:00
|
|
|
|
raw_spin_unlock_rcu_node(rnp);
|
2016-07-01 04:58:26 +08:00
|
|
|
|
}
|
2021-02-16 23:04:34 +08:00
|
|
|
|
arch_spin_unlock(&rcu_state.ofl_lock);
|
rcu-tasks: Stop rcu_tasks_invoke_cbs() from using never-onlined CPUs
The rcu_tasks_invoke_cbs() function relies on queue_work_on() to silently
fall back to WORK_CPU_UNBOUND when the specified CPU is offline. However,
the queue_work_on() function's silent fallback mechanism relies on that
CPU having been online at some time in the past. When queue_work_on()
is passed a CPU that has never been online, workqueue lockups ensue,
which can be bad for your kernel's general health and well-being.
This commit therefore checks whether a given CPU has ever been online,
and, if not substitutes WORK_CPU_UNBOUND in the subsequent call to
queue_work_on(). Why not simply omit the queue_work_on() call entirely?
Because this function is flooding callback-invocation notifications
to all CPUs, and must deal with possibilities that include a sparse
cpu_possible_mask.
This commit also moves the setting of the rcu_data structure's
->beenonline field to rcu_cpu_starting(), which executes on the
incoming CPU before that CPU has ever enabled interrupts. This ensures
that the required workqueues are present. In addition, because the
incoming CPU has not yet enabled its interrupts, there cannot yet have
been any softirq handlers running on this CPU, which means that the
WARN_ON_ONCE(!rdp->beenonline) within the RCU_SOFTIRQ handler cannot
have triggered yet.
Fixes: d363f833c6d88 ("rcu-tasks: Use workqueues for multiple rcu_tasks_invoke_cbs() invocations")
Reported-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2023-04-27 02:11:29 +08:00
|
|
|
|
smp_store_release(&rdp->beenonline, true);
|
rcu: Make expedited GPs correctly handle hardware CPU insertion
The update of the ->expmaskinitnext and of ->ncpus are unsynchronized,
with the value of ->ncpus being incremented long before the corresponding
->expmaskinitnext mask is updated. If an RCU expedited grace period
sees ->ncpus change, it will update the ->expmaskinit masks from the new
->expmaskinitnext masks. But it is possible that ->ncpus has already
been updated, but the ->expmaskinitnext masks still have their old values.
For the current expedited grace period, no harm done. The CPU could not
have been online before the grace period started, so there is no need to
wait for its non-existent pre-existing readers.
But the next RCU expedited grace period is in a world of hurt. The value
of ->ncpus has already been updated, so this grace period will assume
that the ->expmaskinitnext masks have not changed. But they have, and
they won't be taken into account until the next never-been-online CPU
comes online. This means that RCU will be ignoring some CPUs that it
should be paying attention to.
The solution is to update ->ncpus and ->expmaskinitnext while holding
the ->lock for the rcu_node structure containing the ->expmaskinitnext
mask. Because smp_store_release() is now used to update ->ncpus and
smp_load_acquire() is now used to locklessly read it, if the expedited
grace period sees ->ncpus change, then the updating CPU has to
already be holding the corresponding ->lock. Therefore, when the
expedited grace period later acquires that ->lock, it is guaranteed
to see the new value of ->expmaskinitnext.
On the other hand, if the expedited grace period loads ->ncpus just
before an update, earlier full memory barriers guarantee that
the incoming CPU isn't far enough along to be running any RCU readers.
This commit therefore makes the required change.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2017-06-09 07:55:40 +08:00
|
|
|
|
smp_mb(); /* Ensure RCU read-side usage follows above initialization. */
|
2016-07-01 04:58:26 +08:00
|
|
|
|
}
|
|
|
|
|
|
2016-02-27 02:43:44 +08:00
|
|
|
|
/*
|
2018-07-04 08:22:34 +08:00
|
|
|
|
* The outgoing function has no further need of RCU, so remove it from
|
|
|
|
|
* the rcu_node tree's ->qsmaskinitnext bit masks.
|
|
|
|
|
*
|
|
|
|
|
* Note that this function is special in that it is invoked directly
|
|
|
|
|
* from the outgoing CPU rather than from the cpuhp_step mechanism.
|
|
|
|
|
* This is because this function must be invoked at a precise location.
|
2016-02-27 02:43:44 +08:00
|
|
|
|
*/
|
2018-07-04 08:22:34 +08:00
|
|
|
|
void rcu_report_dead(unsigned int cpu)
|
2016-02-27 02:43:44 +08:00
|
|
|
|
{
|
2021-02-16 23:04:34 +08:00
|
|
|
|
unsigned long flags, seq_flags;
|
2016-02-27 02:43:44 +08:00
|
|
|
|
unsigned long mask;
|
2018-07-04 06:37:16 +08:00
|
|
|
|
struct rcu_data *rdp = per_cpu_ptr(&rcu_data, cpu);
|
2016-02-27 02:43:44 +08:00
|
|
|
|
struct rcu_node *rnp = rdp->mynode; /* Outgoing CPU's rdp & rnp. */
|
|
|
|
|
|
2020-12-23 08:49:11 +08:00
|
|
|
|
// Do any dangling deferred wakeups.
|
|
|
|
|
do_nocb_deferred_wakeup(rdp);
|
|
|
|
|
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_preempt_deferred_qs(current);
|
|
|
|
|
|
2016-02-27 02:43:44 +08:00
|
|
|
|
/* Remove outgoing CPU from mask in the leaf rcu_node structure. */
|
|
|
|
|
mask = rdp->grpmask;
|
2021-02-16 23:04:34 +08:00
|
|
|
|
local_irq_save(seq_flags);
|
|
|
|
|
arch_spin_lock(&rcu_state.ofl_lock);
|
2016-02-27 02:43:44 +08:00
|
|
|
|
raw_spin_lock_irqsave_rcu_node(rnp, flags); /* Enforce GP memory-order guarantee. */
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rdp->rcu_ofl_gp_seq = READ_ONCE(rcu_state.gp_seq);
|
|
|
|
|
rdp->rcu_ofl_gp_flags = READ_ONCE(rcu_state.gp_flags);
|
rcu: Suppress false-positive preempted-task splats
Consider the following sequence of events in a PREEMPT=y kernel:
1. All CPUs corresponding to a given rcu_node structure go offline.
A new grace period starts just after the CPU-hotplug code path
does its synchronize_rcu() for the last CPU, so at least this
CPU is present in that structure's ->qsmask.
2. Before the grace period ends, a CPU comes back online, and not
just any CPU, but the one corresponding to a non-zero bit in
the leaf rcu_node structure's ->qsmask.
3. A task running on the newly onlined CPU is preempted while in
an RCU read-side critical section. Because this CPU's ->qsmask
bit is net, not only does this task queue itself on the leaf
rcu_node structure's ->blkd_tasks list, it also sets that
structure's ->gp_tasks pointer to reference it.
4. The grace period started in #1 above comes to an end. This
results in rcu_gp_cleanup() being invoked, which, among other
things, checks to make sure that there are no tasks blocking the
just-ended grace period, that is, that all ->gp_tasks pointers
are NULL. The ->gp_tasks pointer corresponding to the task
preempted in #3 above is non-NULL, which results in a splat.
This splat is a false positive. The task's RCU read-side critical
section cannot have begun before the just-ended grace period because
this would mean either: (1) The CPU came online before the grace period
started, which cannot have happened because the grace period started
before that CPU was all the way offline, or (2) The task started its
RCU read-side critical section on some other CPU, but then it would
have had to have been preempted before migrating to this CPU, which
would mean that it would have instead queued itself on that other CPU's
rcu_node structure.
This commit eliminates this false positive by adding code to the end
of rcu_cleanup_dying_idle_cpu() that reports a quiescent state to RCU,
which has the side-effect of clearing that CPU's ->qsmask bit, preventing
the above scenario. This approach has the added benefit of more promptly
reporting quiescent states corresponding to offline CPUs.
Note well that the call to rcu_report_qs_rnp() reporting the quiescent
state must come -before- the clearing of this CPU's bit in the leaf
rcu_node structure's ->qsmaskinitnext field. Otherwise, lockdep-RCU
will complain bitterly about quiescent states coming from an offline CPU.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-05-03 11:04:12 +08:00
|
|
|
|
if (rnp->qsmask & mask) { /* RCU waiting on outgoing CPU? */
|
|
|
|
|
/* Report quiescent state -before- changing ->qsmaskinitnext! */
|
2022-05-26 09:55:12 +08:00
|
|
|
|
rcu_disable_urgency_upon_qs(rdp);
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_report_qs_rnp(mask, rnp, rnp->gp_seq, flags);
|
rcu: Suppress false-positive preempted-task splats
Consider the following sequence of events in a PREEMPT=y kernel:
1. All CPUs corresponding to a given rcu_node structure go offline.
A new grace period starts just after the CPU-hotplug code path
does its synchronize_rcu() for the last CPU, so at least this
CPU is present in that structure's ->qsmask.
2. Before the grace period ends, a CPU comes back online, and not
just any CPU, but the one corresponding to a non-zero bit in
the leaf rcu_node structure's ->qsmask.
3. A task running on the newly onlined CPU is preempted while in
an RCU read-side critical section. Because this CPU's ->qsmask
bit is net, not only does this task queue itself on the leaf
rcu_node structure's ->blkd_tasks list, it also sets that
structure's ->gp_tasks pointer to reference it.
4. The grace period started in #1 above comes to an end. This
results in rcu_gp_cleanup() being invoked, which, among other
things, checks to make sure that there are no tasks blocking the
just-ended grace period, that is, that all ->gp_tasks pointers
are NULL. The ->gp_tasks pointer corresponding to the task
preempted in #3 above is non-NULL, which results in a splat.
This splat is a false positive. The task's RCU read-side critical
section cannot have begun before the just-ended grace period because
this would mean either: (1) The CPU came online before the grace period
started, which cannot have happened because the grace period started
before that CPU was all the way offline, or (2) The task started its
RCU read-side critical section on some other CPU, but then it would
have had to have been preempted before migrating to this CPU, which
would mean that it would have instead queued itself on that other CPU's
rcu_node structure.
This commit eliminates this false positive by adding code to the end
of rcu_cleanup_dying_idle_cpu() that reports a quiescent state to RCU,
which has the side-effect of clearing that CPU's ->qsmask bit, preventing
the above scenario. This approach has the added benefit of more promptly
reporting quiescent states corresponding to offline CPUs.
Note well that the call to rcu_report_qs_rnp() reporting the quiescent
state must come -before- the clearing of this CPU's bit in the leaf
rcu_node structure's ->qsmaskinitnext field. Otherwise, lockdep-RCU
will complain bitterly about quiescent states coming from an offline CPU.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2018-05-03 11:04:12 +08:00
|
|
|
|
raw_spin_lock_irqsave_rcu_node(rnp, flags);
|
|
|
|
|
}
|
2020-01-04 07:44:23 +08:00
|
|
|
|
WRITE_ONCE(rnp->qsmaskinitnext, rnp->qsmaskinitnext & ~mask);
|
Merge branch 'smp-hotplug-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull cpu hotplug updates from Thomas Gleixner:
"This is the first part of the ongoing cpu hotplug rework:
- Initial implementation of the state machine
- Runs all online and prepare down callbacks on the plugged cpu and
not on some random processor
- Replaces busy loop waiting with completions
- Adds tracepoints so the states can be followed"
More detailed commentary on this work from an earlier email:
"What's wrong with the current cpu hotplug infrastructure?
- Asymmetry
The hotplug notifier mechanism is asymmetric versus the bringup and
teardown. This is mostly caused by the notifier mechanism.
- Largely undocumented dependencies
While some notifiers use explicitely defined notifier priorities,
we have quite some notifiers which use numerical priorities to
express dependencies without any documentation why.
- Control processor driven
Most of the bringup/teardown of a cpu is driven by a control
processor. While it is understandable, that preperatory steps,
like idle thread creation, memory allocation for and initialization
of essential facilities needs to be done before a cpu can boot,
there is no reason why everything else must run on a control
processor. Before this patch series, bringup looks like this:
Control CPU Booting CPU
do preparatory steps
kick cpu into life
do low level init
sync with booting cpu sync with control cpu
bring the rest up
- All or nothing approach
There is no way to do partial bringups. That's something which is
really desired because we waste e.g. at boot substantial amount of
time just busy waiting that the cpu comes to life. That's stupid
as we could very well do preparatory steps and the initial IPI for
other cpus and then go back and do the necessary low level
synchronization with the freshly booted cpu.
- Minimal debuggability
Due to the notifier based design, it's impossible to switch between
two stages of the bringup/teardown back and forth in order to test
the correctness. So in many hotplug notifiers the cancel
mechanisms are either not existant or completely untested.
- Notifier [un]registering is tedious
To [un]register notifiers we need to protect against hotplug at
every callsite. There is no mechanism that bringup/teardown
callbacks are issued on the online cpus, so every caller needs to
do it itself. That also includes error rollback.
What's the new design?
The base of the new design is a symmetric state machine, where both
the control processor and the booting/dying cpu execute a well
defined set of states. Each state is symmetric in the end, except
for some well defined exceptions, and the bringup/teardown can be
stopped and reversed at almost all states.
So the bringup of a cpu will look like this in the future:
Control CPU Booting CPU
do preparatory steps
kick cpu into life
do low level init
sync with booting cpu sync with control cpu
bring itself up
The synchronization step does not require the control cpu to wait.
That mechanism can be done asynchronously via a worker or some
other mechanism.
The teardown can be made very similar, so that the dying cpu cleans
up and brings itself down. Cleanups which need to be done after
the cpu is gone, can be scheduled asynchronously as well.
There is a long way to this, as we need to refactor the notion when a
cpu is available. Today we set the cpu online right after it comes
out of the low level bringup, which is not really correct.
The proper mechanism is to set it to available, i.e. cpu local
threads, like softirqd, hotplug thread etc. can be scheduled on that
cpu, and once it finished all booting steps, it's set to online, so
general workloads can be scheduled on it. The reverse happens on
teardown. First thing to do is to forbid scheduling of general
workloads, then teardown all the per cpu resources and finally shut it
off completely.
This patch series implements the basic infrastructure for this at the
core level. This includes the following:
- Basic state machine implementation with well defined states, so
ordering and prioritization can be expressed.
- Interfaces to [un]register state callbacks
This invokes the bringup/teardown callback on all online cpus with
the proper protection in place and [un]installs the callbacks in
the state machine array.
For callbacks which have no particular ordering requirement we have
a dynamic state space, so that drivers don't have to register an
explicit hotplug state.
If a callback fails, the code automatically does a rollback to the
previous state.
- Sysfs interface to drive the state machine to a particular step.
This is only partially functional today. Full functionality and
therefor testability will be achieved once we converted all
existing hotplug notifiers over to the new scheme.
- Run all CPU_ONLINE/DOWN_PREPARE notifiers on the booting/dying
processor:
Control CPU Booting CPU
do preparatory steps
kick cpu into life
do low level init
sync with booting cpu sync with control cpu
wait for boot
bring itself up
Signal completion to control cpu
In a previous step of this work we've done a full tree mechanical
conversion of all hotplug notifiers to the new scheme. The balance
is a net removal of about 4000 lines of code.
This is not included in this series, as we decided to take a
different approach. Instead of mechanically converting everything
over, we will do a proper overhaul of the usage sites one by one so
they nicely fit into the symmetric callback scheme.
I decided to do that after I looked at the ugliness of some of the
converted sites and figured out that their hotplug mechanism is
completely buggered anyway. So there is no point to do a
mechanical conversion first as we need to go through the usage
sites one by one again in order to achieve a full symmetric and
testable behaviour"
* 'smp-hotplug-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (23 commits)
cpu/hotplug: Document states better
cpu/hotplug: Fix smpboot thread ordering
cpu/hotplug: Remove redundant state check
cpu/hotplug: Plug death reporting race
rcu: Make CPU_DYING_IDLE an explicit call
cpu/hotplug: Make wait for dead cpu completion based
cpu/hotplug: Let upcoming cpu bring itself fully up
arch/hotplug: Call into idle with a proper state
cpu/hotplug: Move online calls to hotplugged cpu
cpu/hotplug: Create hotplug threads
cpu/hotplug: Split out the state walk into functions
cpu/hotplug: Unpark smpboot threads from the state machine
cpu/hotplug: Move scheduler cpu_online notifier to hotplug core
cpu/hotplug: Implement setup/removal interface
cpu/hotplug: Make target state writeable
cpu/hotplug: Add sysfs state interface
cpu/hotplug: Hand in target state to _cpu_up/down
cpu/hotplug: Convert the hotplugged cpu work to a state machine
cpu/hotplug: Convert to a state machine for the control processor
cpu/hotplug: Add tracepoints
...
2016-03-16 04:50:29 +08:00
|
|
|
|
raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
|
2021-02-16 23:04:34 +08:00
|
|
|
|
arch_spin_unlock(&rcu_state.ofl_lock);
|
|
|
|
|
local_irq_restore(seq_flags);
|
2018-05-23 00:50:53 +08:00
|
|
|
|
|
2020-07-25 11:22:05 +08:00
|
|
|
|
rdp->cpu_started = false;
|
2016-02-27 02:43:44 +08:00
|
|
|
|
}
|
rcu: Migrate callbacks earlier in the CPU-offline timeline
RCU callbacks must be migrated away from an outgoing CPU, and this is
done near the end of the CPU-hotplug operation, after the outgoing CPU is
long gone. Unfortunately, this means that other CPU-hotplug callbacks
can execute while the outgoing CPU's callbacks are still immobilized
on the long-gone CPU's callback lists. If any of these CPU-hotplug
callbacks must wait, either directly or indirectly, for the invocation
of any of the immobilized RCU callbacks, the system will hang.
This commit avoids such hangs by migrating the callbacks away from the
outgoing CPU immediately upon its departure, shortly after the return
from __cpu_die() in takedown_cpu(). Thus, RCU is able to advance these
callbacks and invoke them, which allows all the after-the-fact CPU-hotplug
callbacks to wait on these RCU callbacks without risk of a hang.
While in the neighborhood, this commit also moves rcu_send_cbs_to_orphanage()
and rcu_adopt_orphan_cbs() under a pre-existing #ifdef to avoid including
dead code on the one hand and to avoid define-without-use warnings on the
other hand.
Reported-by: Jeffrey Hugo <jhugo@codeaurora.org>
Link: http://lkml.kernel.org/r/db9c91f6-1b17-6136-84f0-03c3c2581ab4@codeaurora.org
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Anna-Maria Gleixner <anna-maria@linutronix.de>
Cc: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Cc: Richard Weinberger <richard@nod.at>
2017-06-21 03:11:34 +08:00
|
|
|
|
|
2020-11-06 18:25:49 +08:00
|
|
|
|
#ifdef CONFIG_HOTPLUG_CPU
|
2018-07-04 08:22:34 +08:00
|
|
|
|
/*
|
|
|
|
|
* The outgoing CPU has just passed through the dying-idle state, and we
|
|
|
|
|
* are being invoked from the CPU that was IPIed to continue the offline
|
|
|
|
|
* operation. Migrate the outgoing CPU's callbacks to the current CPU.
|
|
|
|
|
*/
|
|
|
|
|
void rcutree_migrate_callbacks(int cpu)
|
rcu: Migrate callbacks earlier in the CPU-offline timeline
RCU callbacks must be migrated away from an outgoing CPU, and this is
done near the end of the CPU-hotplug operation, after the outgoing CPU is
long gone. Unfortunately, this means that other CPU-hotplug callbacks
can execute while the outgoing CPU's callbacks are still immobilized
on the long-gone CPU's callback lists. If any of these CPU-hotplug
callbacks must wait, either directly or indirectly, for the invocation
of any of the immobilized RCU callbacks, the system will hang.
This commit avoids such hangs by migrating the callbacks away from the
outgoing CPU immediately upon its departure, shortly after the return
from __cpu_die() in takedown_cpu(). Thus, RCU is able to advance these
callbacks and invoke them, which allows all the after-the-fact CPU-hotplug
callbacks to wait on these RCU callbacks without risk of a hang.
While in the neighborhood, this commit also moves rcu_send_cbs_to_orphanage()
and rcu_adopt_orphan_cbs() under a pre-existing #ifdef to avoid including
dead code on the one hand and to avoid define-without-use warnings on the
other hand.
Reported-by: Jeffrey Hugo <jhugo@codeaurora.org>
Link: http://lkml.kernel.org/r/db9c91f6-1b17-6136-84f0-03c3c2581ab4@codeaurora.org
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Anna-Maria Gleixner <anna-maria@linutronix.de>
Cc: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Cc: Richard Weinberger <richard@nod.at>
2017-06-21 03:11:34 +08:00
|
|
|
|
{
|
|
|
|
|
unsigned long flags;
|
2017-06-27 03:23:46 +08:00
|
|
|
|
struct rcu_data *my_rdp;
|
2019-04-17 05:09:15 +08:00
|
|
|
|
struct rcu_node *my_rnp;
|
2018-07-04 06:37:16 +08:00
|
|
|
|
struct rcu_data *rdp = per_cpu_ptr(&rcu_data, cpu);
|
2018-04-22 23:49:24 +08:00
|
|
|
|
bool needwake;
|
rcu: Migrate callbacks earlier in the CPU-offline timeline
RCU callbacks must be migrated away from an outgoing CPU, and this is
done near the end of the CPU-hotplug operation, after the outgoing CPU is
long gone. Unfortunately, this means that other CPU-hotplug callbacks
can execute while the outgoing CPU's callbacks are still immobilized
on the long-gone CPU's callback lists. If any of these CPU-hotplug
callbacks must wait, either directly or indirectly, for the invocation
of any of the immobilized RCU callbacks, the system will hang.
This commit avoids such hangs by migrating the callbacks away from the
outgoing CPU immediately upon its departure, shortly after the return
from __cpu_die() in takedown_cpu(). Thus, RCU is able to advance these
callbacks and invoke them, which allows all the after-the-fact CPU-hotplug
callbacks to wait on these RCU callbacks without risk of a hang.
While in the neighborhood, this commit also moves rcu_send_cbs_to_orphanage()
and rcu_adopt_orphan_cbs() under a pre-existing #ifdef to avoid including
dead code on the one hand and to avoid define-without-use warnings on the
other hand.
Reported-by: Jeffrey Hugo <jhugo@codeaurora.org>
Link: http://lkml.kernel.org/r/db9c91f6-1b17-6136-84f0-03c3c2581ab4@codeaurora.org
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Anna-Maria Gleixner <anna-maria@linutronix.de>
Cc: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Cc: Richard Weinberger <richard@nod.at>
2017-06-21 03:11:34 +08:00
|
|
|
|
|
2020-11-12 08:51:21 +08:00
|
|
|
|
if (rcu_rdp_is_offloaded(rdp) ||
|
2019-04-13 06:58:34 +08:00
|
|
|
|
rcu_segcblist_empty(&rdp->cblist))
|
2017-06-27 01:49:50 +08:00
|
|
|
|
return; /* No callbacks to migrate. */
|
|
|
|
|
|
rcu: Make rcu_barrier() no longer block CPU-hotplug operations
This commit removes the cpus_read_lock() and cpus_read_unlock() calls
from rcu_barrier(), thus allowing CPUs to come and go during the course
of rcu_barrier() execution. Posting of the ->barrier_head callbacks does
synchronize with portions of RCU's CPU-hotplug notifiers, but these locks
are held for short time periods on both sides. Thus, full CPU-hotplug
operations could both start and finish during the execution of a given
rcu_barrier() invocation.
Additional synchronization is provided by a global ->barrier_lock.
Since the ->barrier_lock is only used during rcu_barrier() execution and
during onlining/offlining a CPU, the contention for this lock should
be low. It might be tempting to make use of a per-CPU lock just on
general principles, but straightforward attempts to do this have the
problems shown below.
Initial state: 3 CPUs present, CPU 0 and CPU1 do not have
any callback and CPU2 has callbacks.
1. CPU0 calls rcu_barrier().
2. CPU1 starts offlining for CPU2. CPU1 calls
rcutree_migrate_callbacks(). rcu_barrier_entrain() is called
from rcutree_migrate_callbacks(), with CPU2's rdp->barrier_lock.
It does not entrain ->barrier_head for CPU2, as rcu_barrier()
on CPU0 hasn't started the barrier sequence (by calling
rcu_seq_start(&rcu_state.barrier_sequence)) yet.
3. CPU0 starts new barrier sequence. It iterates over
CPU0 and CPU1, after acquiring their per-cpu ->barrier_lock
and finds 0 segcblist length. It updates ->barrier_seq_snap
for CPU0 and CPU1 and continues loop iteration to CPU2.
for_each_possible_cpu(cpu) {
raw_spin_lock_irqsave(&rdp->barrier_lock, flags);
if (!rcu_segcblist_n_cbs(&rdp->cblist)) {
WRITE_ONCE(rdp->barrier_seq_snap, gseq);
raw_spin_unlock_irqrestore(&rdp->barrier_lock, flags);
rcu_barrier_trace(TPS("NQ"), cpu, rcu_state.barrier_sequence);
continue;
}
4. rcutree_migrate_callbacks() completes execution on CPU1.
Segcblist len for CPU2 becomes 0.
5. The loop iteration on CPU0, checks rcu_segcblist_n_cbs(&rdp->cblist)
for CPU2 and completes the loop iteration after setting
->barrier_seq_snap.
6. As there isn't any ->barrier_head callback entrained; at
this point, rcu_barrier() in CPU0 returns.
7. The callbacks, which migrated from CPU2 to CPU1, execute.
Straightforward per-CPU locking is also subject to the following race
condition noted by Boqun Feng:
1. CPU0 calls rcu_barrier(), starting a new barrier sequence by invoking
rcu_seq_start() and init_completion(), but does not yet initialize
rcu_state.barrier_cpu_count.
2. CPU1 starts offlining for CPU2, calling rcutree_migrate_callbacks(),
which in turn calls rcu_barrier_entrain() holding CPU2's.
rdp->barrier_lock. It then entrains ->barrier_head for CPU2
and atomically increments rcu_state.barrier_cpu_count, which is
unfortunately not yet initialized to the value 2.
3. The just-entrained RCU callback is invoked. It atomically
decrements rcu_state.barrier_cpu_count and sees that it is
now zero. This callback therefore invokes complete().
4. CPU0 continues executing rcu_barrier(), but is not blocked
by its call to wait_for_completion(). This results in rcu_barrier()
returning before all pre-existing callbacks have been invoked,
which is a bug.
Therefore, synchronization is provided by rcu_state.barrier_lock,
which is also held across the initialization sequence, especially the
rcu_seq_start() and the atomic_set() that sets rcu_state.barrier_cpu_count
to the value 2. In addition, this lock is held when entraining the
rcu_barrier() callback, when deciding whether or not a CPU has callbacks
that rcu_barrier() must wait on, when setting the ->qsmaskinitnext for
incoming CPUs, and when migrating callbacks from a CPU that is going
offline.
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Co-developed-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-12-15 05:35:17 +08:00
|
|
|
|
raw_spin_lock_irqsave(&rcu_state.barrier_lock, flags);
|
2021-12-15 05:15:18 +08:00
|
|
|
|
WARN_ON_ONCE(rcu_rdp_cpu_online(rdp));
|
|
|
|
|
rcu_barrier_entrain(rdp);
|
2018-07-04 06:37:16 +08:00
|
|
|
|
my_rdp = this_cpu_ptr(&rcu_data);
|
2019-04-17 05:09:15 +08:00
|
|
|
|
my_rnp = my_rdp->mynode;
|
2019-05-16 00:56:40 +08:00
|
|
|
|
rcu_nocb_lock(my_rdp); /* irqs already disabled. */
|
2022-10-17 00:22:54 +08:00
|
|
|
|
WARN_ON_ONCE(!rcu_nocb_flush_bypass(my_rdp, NULL, jiffies, false));
|
2019-04-17 05:09:15 +08:00
|
|
|
|
raw_spin_lock_rcu_node(my_rnp); /* irqs already disabled. */
|
2018-04-22 23:49:24 +08:00
|
|
|
|
/* Leverage recent GPs and set GP for new callbacks. */
|
2019-04-17 05:09:15 +08:00
|
|
|
|
needwake = rcu_advance_cbs(my_rnp, rdp) ||
|
|
|
|
|
rcu_advance_cbs(my_rnp, my_rdp);
|
2017-06-27 22:44:06 +08:00
|
|
|
|
rcu_segcblist_merge(&my_rdp->cblist, &rdp->cblist);
|
rcu: Make rcu_barrier() no longer block CPU-hotplug operations
This commit removes the cpus_read_lock() and cpus_read_unlock() calls
from rcu_barrier(), thus allowing CPUs to come and go during the course
of rcu_barrier() execution. Posting of the ->barrier_head callbacks does
synchronize with portions of RCU's CPU-hotplug notifiers, but these locks
are held for short time periods on both sides. Thus, full CPU-hotplug
operations could both start and finish during the execution of a given
rcu_barrier() invocation.
Additional synchronization is provided by a global ->barrier_lock.
Since the ->barrier_lock is only used during rcu_barrier() execution and
during onlining/offlining a CPU, the contention for this lock should
be low. It might be tempting to make use of a per-CPU lock just on
general principles, but straightforward attempts to do this have the
problems shown below.
Initial state: 3 CPUs present, CPU 0 and CPU1 do not have
any callback and CPU2 has callbacks.
1. CPU0 calls rcu_barrier().
2. CPU1 starts offlining for CPU2. CPU1 calls
rcutree_migrate_callbacks(). rcu_barrier_entrain() is called
from rcutree_migrate_callbacks(), with CPU2's rdp->barrier_lock.
It does not entrain ->barrier_head for CPU2, as rcu_barrier()
on CPU0 hasn't started the barrier sequence (by calling
rcu_seq_start(&rcu_state.barrier_sequence)) yet.
3. CPU0 starts new barrier sequence. It iterates over
CPU0 and CPU1, after acquiring their per-cpu ->barrier_lock
and finds 0 segcblist length. It updates ->barrier_seq_snap
for CPU0 and CPU1 and continues loop iteration to CPU2.
for_each_possible_cpu(cpu) {
raw_spin_lock_irqsave(&rdp->barrier_lock, flags);
if (!rcu_segcblist_n_cbs(&rdp->cblist)) {
WRITE_ONCE(rdp->barrier_seq_snap, gseq);
raw_spin_unlock_irqrestore(&rdp->barrier_lock, flags);
rcu_barrier_trace(TPS("NQ"), cpu, rcu_state.barrier_sequence);
continue;
}
4. rcutree_migrate_callbacks() completes execution on CPU1.
Segcblist len for CPU2 becomes 0.
5. The loop iteration on CPU0, checks rcu_segcblist_n_cbs(&rdp->cblist)
for CPU2 and completes the loop iteration after setting
->barrier_seq_snap.
6. As there isn't any ->barrier_head callback entrained; at
this point, rcu_barrier() in CPU0 returns.
7. The callbacks, which migrated from CPU2 to CPU1, execute.
Straightforward per-CPU locking is also subject to the following race
condition noted by Boqun Feng:
1. CPU0 calls rcu_barrier(), starting a new barrier sequence by invoking
rcu_seq_start() and init_completion(), but does not yet initialize
rcu_state.barrier_cpu_count.
2. CPU1 starts offlining for CPU2, calling rcutree_migrate_callbacks(),
which in turn calls rcu_barrier_entrain() holding CPU2's.
rdp->barrier_lock. It then entrains ->barrier_head for CPU2
and atomically increments rcu_state.barrier_cpu_count, which is
unfortunately not yet initialized to the value 2.
3. The just-entrained RCU callback is invoked. It atomically
decrements rcu_state.barrier_cpu_count and sees that it is
now zero. This callback therefore invokes complete().
4. CPU0 continues executing rcu_barrier(), but is not blocked
by its call to wait_for_completion(). This results in rcu_barrier()
returning before all pre-existing callbacks have been invoked,
which is a bug.
Therefore, synchronization is provided by rcu_state.barrier_lock,
which is also held across the initialization sequence, especially the
rcu_seq_start() and the atomic_set() that sets rcu_state.barrier_cpu_count
to the value 2. In addition, this lock is held when entraining the
rcu_barrier() callback, when deciding whether or not a CPU has callbacks
that rcu_barrier() must wait on, when setting the ->qsmaskinitnext for
incoming CPUs, and when migrating callbacks from a CPU that is going
offline.
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Co-developed-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-12-15 05:35:17 +08:00
|
|
|
|
raw_spin_unlock(&rcu_state.barrier_lock); /* irqs remain disabled. */
|
2019-07-11 03:54:56 +08:00
|
|
|
|
needwake = needwake || rcu_advance_cbs(my_rnp, my_rdp);
|
2019-05-21 23:28:41 +08:00
|
|
|
|
rcu_segcblist_disable(&rdp->cblist);
|
2021-12-15 05:15:18 +08:00
|
|
|
|
WARN_ON_ONCE(rcu_segcblist_empty(&my_rdp->cblist) != !rcu_segcblist_n_cbs(&my_rdp->cblist));
|
rcu: Add rnp->cbovldmask check in rcutree_migrate_callbacks()
Currently, the rcu_node structure's ->cbovlmask field is set in call_rcu()
when a given CPU is suffering from callback overload. But if that CPU
goes offline, the outgoing CPU's callbacks is migrated to the running
CPU, which is likely to overload the running CPU. However, that CPU's
bit in its leaf rcu_node structure's ->cbovlmask field remains zero.
Initially, this is OK because the outgoing CPU's bit remains set.
However, that bit will be cleared at the next end of a grace period,
at which time it is quite possible that the running CPU will still
be overloaded. If the running CPU invokes call_rcu(), then overload
will be checked for and the bit will be set. Except that there is no
guarantee that the running CPU will invoke call_rcu(), in which case the
next grace period will fail to take the running CPU's overload condition
into account. Plus, because the bit is not set, the end of the grace
period won't check for overload on this CPU.
This commit therefore adds a call to check_cb_ovld_locked() in
rcutree_migrate_callbacks() to set the running CPU's ->cbovlmask bit
appropriately.
Signed-off-by: Zqiang <qiang1.zhang@intel.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Reviewed-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
2022-05-05 23:52:36 +08:00
|
|
|
|
check_cb_ovld_locked(my_rdp, my_rnp);
|
2020-11-12 08:51:21 +08:00
|
|
|
|
if (rcu_rdp_is_offloaded(my_rdp)) {
|
2019-05-16 00:56:40 +08:00
|
|
|
|
raw_spin_unlock_rcu_node(my_rnp); /* irqs remain disabled. */
|
|
|
|
|
__call_rcu_nocb_wake(my_rdp, true, flags);
|
|
|
|
|
} else {
|
|
|
|
|
rcu_nocb_unlock(my_rdp); /* irqs remain disabled. */
|
|
|
|
|
raw_spin_unlock_irqrestore_rcu_node(my_rnp, flags);
|
|
|
|
|
}
|
2018-04-22 23:49:24 +08:00
|
|
|
|
if (needwake)
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_gp_kthread_wake();
|
2019-05-16 00:56:40 +08:00
|
|
|
|
lockdep_assert_irqs_enabled();
|
rcu: Migrate callbacks earlier in the CPU-offline timeline
RCU callbacks must be migrated away from an outgoing CPU, and this is
done near the end of the CPU-hotplug operation, after the outgoing CPU is
long gone. Unfortunately, this means that other CPU-hotplug callbacks
can execute while the outgoing CPU's callbacks are still immobilized
on the long-gone CPU's callback lists. If any of these CPU-hotplug
callbacks must wait, either directly or indirectly, for the invocation
of any of the immobilized RCU callbacks, the system will hang.
This commit avoids such hangs by migrating the callbacks away from the
outgoing CPU immediately upon its departure, shortly after the return
from __cpu_die() in takedown_cpu(). Thus, RCU is able to advance these
callbacks and invoke them, which allows all the after-the-fact CPU-hotplug
callbacks to wait on these RCU callbacks without risk of a hang.
While in the neighborhood, this commit also moves rcu_send_cbs_to_orphanage()
and rcu_adopt_orphan_cbs() under a pre-existing #ifdef to avoid including
dead code on the one hand and to avoid define-without-use warnings on the
other hand.
Reported-by: Jeffrey Hugo <jhugo@codeaurora.org>
Link: http://lkml.kernel.org/r/db9c91f6-1b17-6136-84f0-03c3c2581ab4@codeaurora.org
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Anna-Maria Gleixner <anna-maria@linutronix.de>
Cc: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Cc: Richard Weinberger <richard@nod.at>
2017-06-21 03:11:34 +08:00
|
|
|
|
WARN_ONCE(rcu_segcblist_n_cbs(&rdp->cblist) != 0 ||
|
|
|
|
|
!rcu_segcblist_empty(&rdp->cblist),
|
|
|
|
|
"rcu_cleanup_dead_cpu: Callbacks on offline CPU %d: qlen=%lu, 1stCB=%p\n",
|
|
|
|
|
cpu, rcu_segcblist_n_cbs(&rdp->cblist),
|
|
|
|
|
rcu_segcblist_first_cb(&rdp->cblist));
|
|
|
|
|
}
|
2016-02-27 02:43:44 +08:00
|
|
|
|
#endif
|
|
|
|
|
|
2017-03-24 04:21:30 +08:00
|
|
|
|
/*
|
|
|
|
|
* On non-huge systems, use expedited RCU grace periods to make suspend
|
|
|
|
|
* and hibernation run faster.
|
|
|
|
|
*/
|
2013-04-22 06:12:42 +08:00
|
|
|
|
static int rcu_pm_notify(struct notifier_block *self,
|
|
|
|
|
unsigned long action, void *hcpu)
|
|
|
|
|
{
|
|
|
|
|
switch (action) {
|
|
|
|
|
case PM_HIBERNATION_PREPARE:
|
|
|
|
|
case PM_SUSPEND_PREPARE:
|
2023-01-12 08:52:22 +08:00
|
|
|
|
rcu_async_hurry();
|
2019-01-11 07:30:15 +08:00
|
|
|
|
rcu_expedite_gp();
|
2013-04-22 06:12:42 +08:00
|
|
|
|
break;
|
|
|
|
|
case PM_POST_HIBERNATION:
|
|
|
|
|
case PM_POST_SUSPEND:
|
2019-01-11 07:30:15 +08:00
|
|
|
|
rcu_unexpedite_gp();
|
2023-01-12 08:52:22 +08:00
|
|
|
|
rcu_async_relax();
|
2013-04-22 06:12:42 +08:00
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
return NOTIFY_OK;
|
|
|
|
|
}
|
|
|
|
|
|
rcu: Move expedited grace period (GP) work to RT kthread_worker
Enabling CONFIG_RCU_BOOST did not reduce RCU expedited grace-period
latency because its workqueues run at SCHED_OTHER, and thus can be
delayed by normal processes. This commit avoids these delays by moving
the expedited GP work items to a real-time-priority kthread_worker.
This option is controlled by CONFIG_RCU_EXP_KTHREAD and disabled by
default on PREEMPT_RT=y kernels which disable expedited grace periods
after boot by unconditionally setting rcupdate.rcu_normal_after_boot=1.
The results were evaluated on arm64 Android devices (6GB ram) running
5.10 kernel, and capturing trace data in critical user-level code.
The table below shows the resulting order-of-magnitude improvements
in synchronize_rcu_expedited() latency:
------------------------------------------------------------------------
| | workqueues | kthread_worker | Diff |
------------------------------------------------------------------------
| Count | 725 | 688 | |
------------------------------------------------------------------------
| Min Duration (ns) | 326 | 447 | 37.12% |
------------------------------------------------------------------------
| Q1 (ns) | 39,428 | 38,971 | -1.16% |
------------------------------------------------------------------------
| Q2 - Median (ns) | 98,225 | 69,743 | -29.00% |
------------------------------------------------------------------------
| Q3 (ns) | 342,122 | 126,638 | -62.98% |
------------------------------------------------------------------------
| Max Duration (ns) | 372,766,967 | 2,329,671 | -99.38% |
------------------------------------------------------------------------
| Avg Duration (ns) | 2,746,353 | 151,242 | -94.49% |
------------------------------------------------------------------------
| Standard Deviation (ns) | 19,327,765 | 294,408 | |
------------------------------------------------------------------------
The below table show the range of maximums/minimums for
synchronize_rcu_expedited() latency from all experiments:
------------------------------------------------------------------------
| | workqueues | kthread_worker | Diff |
------------------------------------------------------------------------
| Total No. of Experiments | 25 | 23 | |
------------------------------------------------------------------------
| Largest Maximum (ns) | 372,766,967 | 2,329,671 | -99.38% |
------------------------------------------------------------------------
| Smallest Maximum (ns) | 38,819 | 86,954 | 124.00% |
------------------------------------------------------------------------
| Range of Maximums (ns) | 372,728,148 | 2,242,717 | |
------------------------------------------------------------------------
| Largest Minimum (ns) | 88,623 | 27,588 | -68.87% |
------------------------------------------------------------------------
| Smallest Minimum (ns) | 326 | 447 | 37.12% |
------------------------------------------------------------------------
| Range of Minimums (ns) | 88,297 | 27,141 | |
------------------------------------------------------------------------
Cc: "Paul E. McKenney" <paulmck@kernel.org>
Cc: Tejun Heo <tj@kernel.org>
Reported-by: Tim Murray <timmurray@google.com>
Reported-by: Wei Wang <wvw@google.com>
Tested-by: Kyle Lin <kylelin@google.com>
Tested-by: Chunwei Lu <chunweilu@google.com>
Tested-by: Lulu Wang <luluw@google.com>
Signed-off-by: Kalesh Singh <kaleshsingh@google.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2022-04-09 08:35:27 +08:00
|
|
|
|
#ifdef CONFIG_RCU_EXP_KTHREAD
|
|
|
|
|
struct kthread_worker *rcu_exp_gp_kworker;
|
|
|
|
|
struct kthread_worker *rcu_exp_par_gp_kworker;
|
|
|
|
|
|
|
|
|
|
static void __init rcu_start_exp_gp_kworkers(void)
|
|
|
|
|
{
|
|
|
|
|
const char *par_gp_kworker_name = "rcu_exp_par_gp_kthread_worker";
|
|
|
|
|
const char *gp_kworker_name = "rcu_exp_gp_kthread_worker";
|
|
|
|
|
struct sched_param param = { .sched_priority = kthread_prio };
|
|
|
|
|
|
|
|
|
|
rcu_exp_gp_kworker = kthread_create_worker(0, gp_kworker_name);
|
|
|
|
|
if (IS_ERR_OR_NULL(rcu_exp_gp_kworker)) {
|
|
|
|
|
pr_err("Failed to create %s!\n", gp_kworker_name);
|
2024-01-12 23:46:16 +08:00
|
|
|
|
rcu_exp_gp_kworker = NULL;
|
rcu: Move expedited grace period (GP) work to RT kthread_worker
Enabling CONFIG_RCU_BOOST did not reduce RCU expedited grace-period
latency because its workqueues run at SCHED_OTHER, and thus can be
delayed by normal processes. This commit avoids these delays by moving
the expedited GP work items to a real-time-priority kthread_worker.
This option is controlled by CONFIG_RCU_EXP_KTHREAD and disabled by
default on PREEMPT_RT=y kernels which disable expedited grace periods
after boot by unconditionally setting rcupdate.rcu_normal_after_boot=1.
The results were evaluated on arm64 Android devices (6GB ram) running
5.10 kernel, and capturing trace data in critical user-level code.
The table below shows the resulting order-of-magnitude improvements
in synchronize_rcu_expedited() latency:
------------------------------------------------------------------------
| | workqueues | kthread_worker | Diff |
------------------------------------------------------------------------
| Count | 725 | 688 | |
------------------------------------------------------------------------
| Min Duration (ns) | 326 | 447 | 37.12% |
------------------------------------------------------------------------
| Q1 (ns) | 39,428 | 38,971 | -1.16% |
------------------------------------------------------------------------
| Q2 - Median (ns) | 98,225 | 69,743 | -29.00% |
------------------------------------------------------------------------
| Q3 (ns) | 342,122 | 126,638 | -62.98% |
------------------------------------------------------------------------
| Max Duration (ns) | 372,766,967 | 2,329,671 | -99.38% |
------------------------------------------------------------------------
| Avg Duration (ns) | 2,746,353 | 151,242 | -94.49% |
------------------------------------------------------------------------
| Standard Deviation (ns) | 19,327,765 | 294,408 | |
------------------------------------------------------------------------
The below table show the range of maximums/minimums for
synchronize_rcu_expedited() latency from all experiments:
------------------------------------------------------------------------
| | workqueues | kthread_worker | Diff |
------------------------------------------------------------------------
| Total No. of Experiments | 25 | 23 | |
------------------------------------------------------------------------
| Largest Maximum (ns) | 372,766,967 | 2,329,671 | -99.38% |
------------------------------------------------------------------------
| Smallest Maximum (ns) | 38,819 | 86,954 | 124.00% |
------------------------------------------------------------------------
| Range of Maximums (ns) | 372,728,148 | 2,242,717 | |
------------------------------------------------------------------------
| Largest Minimum (ns) | 88,623 | 27,588 | -68.87% |
------------------------------------------------------------------------
| Smallest Minimum (ns) | 326 | 447 | 37.12% |
------------------------------------------------------------------------
| Range of Minimums (ns) | 88,297 | 27,141 | |
------------------------------------------------------------------------
Cc: "Paul E. McKenney" <paulmck@kernel.org>
Cc: Tejun Heo <tj@kernel.org>
Reported-by: Tim Murray <timmurray@google.com>
Reported-by: Wei Wang <wvw@google.com>
Tested-by: Kyle Lin <kylelin@google.com>
Tested-by: Chunwei Lu <chunweilu@google.com>
Tested-by: Lulu Wang <luluw@google.com>
Signed-off-by: Kalesh Singh <kaleshsingh@google.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2022-04-09 08:35:27 +08:00
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
rcu_exp_par_gp_kworker = kthread_create_worker(0, par_gp_kworker_name);
|
|
|
|
|
if (IS_ERR_OR_NULL(rcu_exp_par_gp_kworker)) {
|
|
|
|
|
pr_err("Failed to create %s!\n", par_gp_kworker_name);
|
2024-01-12 23:46:15 +08:00
|
|
|
|
rcu_exp_par_gp_kworker = NULL;
|
rcu: Move expedited grace period (GP) work to RT kthread_worker
Enabling CONFIG_RCU_BOOST did not reduce RCU expedited grace-period
latency because its workqueues run at SCHED_OTHER, and thus can be
delayed by normal processes. This commit avoids these delays by moving
the expedited GP work items to a real-time-priority kthread_worker.
This option is controlled by CONFIG_RCU_EXP_KTHREAD and disabled by
default on PREEMPT_RT=y kernels which disable expedited grace periods
after boot by unconditionally setting rcupdate.rcu_normal_after_boot=1.
The results were evaluated on arm64 Android devices (6GB ram) running
5.10 kernel, and capturing trace data in critical user-level code.
The table below shows the resulting order-of-magnitude improvements
in synchronize_rcu_expedited() latency:
------------------------------------------------------------------------
| | workqueues | kthread_worker | Diff |
------------------------------------------------------------------------
| Count | 725 | 688 | |
------------------------------------------------------------------------
| Min Duration (ns) | 326 | 447 | 37.12% |
------------------------------------------------------------------------
| Q1 (ns) | 39,428 | 38,971 | -1.16% |
------------------------------------------------------------------------
| Q2 - Median (ns) | 98,225 | 69,743 | -29.00% |
------------------------------------------------------------------------
| Q3 (ns) | 342,122 | 126,638 | -62.98% |
------------------------------------------------------------------------
| Max Duration (ns) | 372,766,967 | 2,329,671 | -99.38% |
------------------------------------------------------------------------
| Avg Duration (ns) | 2,746,353 | 151,242 | -94.49% |
------------------------------------------------------------------------
| Standard Deviation (ns) | 19,327,765 | 294,408 | |
------------------------------------------------------------------------
The below table show the range of maximums/minimums for
synchronize_rcu_expedited() latency from all experiments:
------------------------------------------------------------------------
| | workqueues | kthread_worker | Diff |
------------------------------------------------------------------------
| Total No. of Experiments | 25 | 23 | |
------------------------------------------------------------------------
| Largest Maximum (ns) | 372,766,967 | 2,329,671 | -99.38% |
------------------------------------------------------------------------
| Smallest Maximum (ns) | 38,819 | 86,954 | 124.00% |
------------------------------------------------------------------------
| Range of Maximums (ns) | 372,728,148 | 2,242,717 | |
------------------------------------------------------------------------
| Largest Minimum (ns) | 88,623 | 27,588 | -68.87% |
------------------------------------------------------------------------
| Smallest Minimum (ns) | 326 | 447 | 37.12% |
------------------------------------------------------------------------
| Range of Minimums (ns) | 88,297 | 27,141 | |
------------------------------------------------------------------------
Cc: "Paul E. McKenney" <paulmck@kernel.org>
Cc: Tejun Heo <tj@kernel.org>
Reported-by: Tim Murray <timmurray@google.com>
Reported-by: Wei Wang <wvw@google.com>
Tested-by: Kyle Lin <kylelin@google.com>
Tested-by: Chunwei Lu <chunweilu@google.com>
Tested-by: Lulu Wang <luluw@google.com>
Signed-off-by: Kalesh Singh <kaleshsingh@google.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2022-04-09 08:35:27 +08:00
|
|
|
|
kthread_destroy_worker(rcu_exp_gp_kworker);
|
2024-01-12 23:46:16 +08:00
|
|
|
|
rcu_exp_gp_kworker = NULL;
|
rcu: Move expedited grace period (GP) work to RT kthread_worker
Enabling CONFIG_RCU_BOOST did not reduce RCU expedited grace-period
latency because its workqueues run at SCHED_OTHER, and thus can be
delayed by normal processes. This commit avoids these delays by moving
the expedited GP work items to a real-time-priority kthread_worker.
This option is controlled by CONFIG_RCU_EXP_KTHREAD and disabled by
default on PREEMPT_RT=y kernels which disable expedited grace periods
after boot by unconditionally setting rcupdate.rcu_normal_after_boot=1.
The results were evaluated on arm64 Android devices (6GB ram) running
5.10 kernel, and capturing trace data in critical user-level code.
The table below shows the resulting order-of-magnitude improvements
in synchronize_rcu_expedited() latency:
------------------------------------------------------------------------
| | workqueues | kthread_worker | Diff |
------------------------------------------------------------------------
| Count | 725 | 688 | |
------------------------------------------------------------------------
| Min Duration (ns) | 326 | 447 | 37.12% |
------------------------------------------------------------------------
| Q1 (ns) | 39,428 | 38,971 | -1.16% |
------------------------------------------------------------------------
| Q2 - Median (ns) | 98,225 | 69,743 | -29.00% |
------------------------------------------------------------------------
| Q3 (ns) | 342,122 | 126,638 | -62.98% |
------------------------------------------------------------------------
| Max Duration (ns) | 372,766,967 | 2,329,671 | -99.38% |
------------------------------------------------------------------------
| Avg Duration (ns) | 2,746,353 | 151,242 | -94.49% |
------------------------------------------------------------------------
| Standard Deviation (ns) | 19,327,765 | 294,408 | |
------------------------------------------------------------------------
The below table show the range of maximums/minimums for
synchronize_rcu_expedited() latency from all experiments:
------------------------------------------------------------------------
| | workqueues | kthread_worker | Diff |
------------------------------------------------------------------------
| Total No. of Experiments | 25 | 23 | |
------------------------------------------------------------------------
| Largest Maximum (ns) | 372,766,967 | 2,329,671 | -99.38% |
------------------------------------------------------------------------
| Smallest Maximum (ns) | 38,819 | 86,954 | 124.00% |
------------------------------------------------------------------------
| Range of Maximums (ns) | 372,728,148 | 2,242,717 | |
------------------------------------------------------------------------
| Largest Minimum (ns) | 88,623 | 27,588 | -68.87% |
------------------------------------------------------------------------
| Smallest Minimum (ns) | 326 | 447 | 37.12% |
------------------------------------------------------------------------
| Range of Minimums (ns) | 88,297 | 27,141 | |
------------------------------------------------------------------------
Cc: "Paul E. McKenney" <paulmck@kernel.org>
Cc: Tejun Heo <tj@kernel.org>
Reported-by: Tim Murray <timmurray@google.com>
Reported-by: Wei Wang <wvw@google.com>
Tested-by: Kyle Lin <kylelin@google.com>
Tested-by: Chunwei Lu <chunweilu@google.com>
Tested-by: Lulu Wang <luluw@google.com>
Signed-off-by: Kalesh Singh <kaleshsingh@google.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2022-04-09 08:35:27 +08:00
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
sched_setscheduler_nocheck(rcu_exp_gp_kworker->task, SCHED_FIFO, ¶m);
|
|
|
|
|
sched_setscheduler_nocheck(rcu_exp_par_gp_kworker->task, SCHED_FIFO,
|
|
|
|
|
¶m);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline void rcu_alloc_par_gp_wq(void)
|
|
|
|
|
{
|
|
|
|
|
}
|
|
|
|
|
#else /* !CONFIG_RCU_EXP_KTHREAD */
|
|
|
|
|
struct workqueue_struct *rcu_par_gp_wq;
|
|
|
|
|
|
|
|
|
|
static void __init rcu_start_exp_gp_kworkers(void)
|
|
|
|
|
{
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline void rcu_alloc_par_gp_wq(void)
|
|
|
|
|
{
|
|
|
|
|
rcu_par_gp_wq = alloc_workqueue("rcu_par_gp", WQ_MEM_RECLAIM, 0);
|
|
|
|
|
WARN_ON(!rcu_par_gp_wq);
|
|
|
|
|
}
|
|
|
|
|
#endif /* CONFIG_RCU_EXP_KTHREAD */
|
|
|
|
|
|
2012-06-19 09:36:08 +08:00
|
|
|
|
/*
|
2018-07-08 09:12:26 +08:00
|
|
|
|
* Spawn the kthreads that handle RCU's grace periods.
|
2012-06-19 09:36:08 +08:00
|
|
|
|
*/
|
|
|
|
|
static int __init rcu_spawn_gp_kthread(void)
|
|
|
|
|
{
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
struct rcu_node *rnp;
|
2014-12-12 23:37:48 +08:00
|
|
|
|
struct sched_param sp;
|
2012-06-19 09:36:08 +08:00
|
|
|
|
struct task_struct *t;
|
2022-02-16 23:42:07 +08:00
|
|
|
|
struct rcu_data *rdp = this_cpu_ptr(&rcu_data);
|
2012-06-19 09:36:08 +08:00
|
|
|
|
|
2014-07-14 03:00:53 +08:00
|
|
|
|
rcu_scheduler_fully_active = 1;
|
2018-07-05 06:35:00 +08:00
|
|
|
|
t = kthread_create(rcu_gp_kthread, NULL, "%s", rcu_state.name);
|
2018-10-22 23:04:03 +08:00
|
|
|
|
if (WARN_ONCE(IS_ERR(t), "%s: Could not start grace-period kthread, OOM is now expected behavior\n", __func__))
|
|
|
|
|
return 0;
|
2018-07-05 06:35:00 +08:00
|
|
|
|
if (kthread_prio) {
|
|
|
|
|
sp.sched_priority = kthread_prio;
|
|
|
|
|
sched_setscheduler_nocheck(t, SCHED_FIFO, &sp);
|
2012-06-19 09:36:08 +08:00
|
|
|
|
}
|
2018-07-05 06:35:00 +08:00
|
|
|
|
rnp = rcu_get_root();
|
|
|
|
|
raw_spin_lock_irqsave_rcu_node(rnp, flags);
|
2020-01-22 04:30:22 +08:00
|
|
|
|
WRITE_ONCE(rcu_state.gp_activity, jiffies);
|
|
|
|
|
WRITE_ONCE(rcu_state.gp_req_activity, jiffies);
|
|
|
|
|
// Reset .gp_activity and .gp_req_activity before setting .gp_kthread.
|
|
|
|
|
smp_store_release(&rcu_state.gp_kthread, t); /* ^^^ */
|
2018-07-05 06:35:00 +08:00
|
|
|
|
raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
|
|
|
|
|
wake_up_process(t);
|
2022-02-16 23:42:07 +08:00
|
|
|
|
/* This is a pre-SMP initcall, we expect a single CPU */
|
|
|
|
|
WARN_ON(num_online_cpus() > 1);
|
2022-02-16 23:42:08 +08:00
|
|
|
|
/*
|
|
|
|
|
* Those kthreads couldn't be created on rcu_init() -> rcutree_prepare_cpu()
|
|
|
|
|
* due to rcu_scheduler_fully_active.
|
|
|
|
|
*/
|
|
|
|
|
rcu_spawn_cpu_nocb_kthread(smp_processor_id());
|
2022-02-16 23:42:07 +08:00
|
|
|
|
rcu_spawn_one_boost_kthread(rdp->mynode);
|
2021-04-01 01:59:05 +08:00
|
|
|
|
rcu_spawn_core_kthreads();
|
rcu: Move expedited grace period (GP) work to RT kthread_worker
Enabling CONFIG_RCU_BOOST did not reduce RCU expedited grace-period
latency because its workqueues run at SCHED_OTHER, and thus can be
delayed by normal processes. This commit avoids these delays by moving
the expedited GP work items to a real-time-priority kthread_worker.
This option is controlled by CONFIG_RCU_EXP_KTHREAD and disabled by
default on PREEMPT_RT=y kernels which disable expedited grace periods
after boot by unconditionally setting rcupdate.rcu_normal_after_boot=1.
The results were evaluated on arm64 Android devices (6GB ram) running
5.10 kernel, and capturing trace data in critical user-level code.
The table below shows the resulting order-of-magnitude improvements
in synchronize_rcu_expedited() latency:
------------------------------------------------------------------------
| | workqueues | kthread_worker | Diff |
------------------------------------------------------------------------
| Count | 725 | 688 | |
------------------------------------------------------------------------
| Min Duration (ns) | 326 | 447 | 37.12% |
------------------------------------------------------------------------
| Q1 (ns) | 39,428 | 38,971 | -1.16% |
------------------------------------------------------------------------
| Q2 - Median (ns) | 98,225 | 69,743 | -29.00% |
------------------------------------------------------------------------
| Q3 (ns) | 342,122 | 126,638 | -62.98% |
------------------------------------------------------------------------
| Max Duration (ns) | 372,766,967 | 2,329,671 | -99.38% |
------------------------------------------------------------------------
| Avg Duration (ns) | 2,746,353 | 151,242 | -94.49% |
------------------------------------------------------------------------
| Standard Deviation (ns) | 19,327,765 | 294,408 | |
------------------------------------------------------------------------
The below table show the range of maximums/minimums for
synchronize_rcu_expedited() latency from all experiments:
------------------------------------------------------------------------
| | workqueues | kthread_worker | Diff |
------------------------------------------------------------------------
| Total No. of Experiments | 25 | 23 | |
------------------------------------------------------------------------
| Largest Maximum (ns) | 372,766,967 | 2,329,671 | -99.38% |
------------------------------------------------------------------------
| Smallest Maximum (ns) | 38,819 | 86,954 | 124.00% |
------------------------------------------------------------------------
| Range of Maximums (ns) | 372,728,148 | 2,242,717 | |
------------------------------------------------------------------------
| Largest Minimum (ns) | 88,623 | 27,588 | -68.87% |
------------------------------------------------------------------------
| Smallest Minimum (ns) | 326 | 447 | 37.12% |
------------------------------------------------------------------------
| Range of Minimums (ns) | 88,297 | 27,141 | |
------------------------------------------------------------------------
Cc: "Paul E. McKenney" <paulmck@kernel.org>
Cc: Tejun Heo <tj@kernel.org>
Reported-by: Tim Murray <timmurray@google.com>
Reported-by: Wei Wang <wvw@google.com>
Tested-by: Kyle Lin <kylelin@google.com>
Tested-by: Chunwei Lu <chunweilu@google.com>
Tested-by: Lulu Wang <luluw@google.com>
Signed-off-by: Kalesh Singh <kaleshsingh@google.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2022-04-09 08:35:27 +08:00
|
|
|
|
/* Create kthread worker for expedited GPs */
|
|
|
|
|
rcu_start_exp_gp_kworkers();
|
2012-06-19 09:36:08 +08:00
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
early_initcall(rcu_spawn_gp_kthread);
|
|
|
|
|
|
2010-04-03 07:17:17 +08:00
|
|
|
|
/*
|
rcu: Narrow early boot window of illegal synchronous grace periods
The current preemptible RCU implementation goes through three phases
during bootup. In the first phase, there is only one CPU that is running
with preemption disabled, so that a no-op is a synchronous grace period.
In the second mid-boot phase, the scheduler is running, but RCU has
not yet gotten its kthreads spawned (and, for expedited grace periods,
workqueues are not yet running. During this time, any attempt to do
a synchronous grace period will hang the system (or complain bitterly,
depending). In the third and final phase, RCU is fully operational and
everything works normally.
This has been OK for some time, but there has recently been some
synchronous grace periods showing up during the second mid-boot phase.
This code worked "by accident" for awhile, but started failing as soon
as expedited RCU grace periods switched over to workqueues in commit
8b355e3bc140 ("rcu: Drive expedited grace periods from workqueue").
Note that the code was buggy even before this commit, as it was subject
to failure on real-time systems that forced all expedited grace periods
to run as normal grace periods (for example, using the rcu_normal ksysfs
parameter). The callchain from the failure case is as follows:
early_amd_iommu_init()
|-> acpi_put_table(ivrs_base);
|-> acpi_tb_put_table(table_desc);
|-> acpi_tb_invalidate_table(table_desc);
|-> acpi_tb_release_table(...)
|-> acpi_os_unmap_memory
|-> acpi_os_unmap_iomem
|-> acpi_os_map_cleanup
|-> synchronize_rcu_expedited
The kernel showing this callchain was built with CONFIG_PREEMPT_RCU=y,
which caused the code to try using workqueues before they were
initialized, which did not go well.
This commit therefore reworks RCU to permit synchronous grace periods
to proceed during this mid-boot phase. This commit is therefore a
fix to a regression introduced in v4.9, and is therefore being put
forward post-merge-window in v4.10.
This commit sets a flag from the existing rcu_scheduler_starting()
function which causes all synchronous grace periods to take the expedited
path. The expedited path now checks this flag, using the requesting task
to drive the expedited grace period forward during the mid-boot phase.
Finally, this flag is updated by a core_initcall() function named
rcu_exp_runtime_mode(), which causes the runtime codepaths to be used.
Note that this arrangement assumes that tasks are not sent POSIX signals
(or anything similar) from the time that the first task is spawned
through core_initcall() time.
Fixes: 8b355e3bc140 ("rcu: Drive expedited grace periods from workqueue")
Reported-by: "Zheng, Lv" <lv.zheng@intel.com>
Reported-by: Borislav Petkov <bp@alien8.de>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Stan Kain <stan.kain@gmail.com>
Tested-by: Ivan <waffolz@hotmail.com>
Tested-by: Emanuel Castelo <emanuel.castelo@gmail.com>
Tested-by: Bruno Pesavento <bpesavento@infinito.it>
Tested-by: Borislav Petkov <bp@suse.de>
Tested-by: Frederic Bezies <fredbezies@gmail.com>
Cc: <stable@vger.kernel.org> # 4.9.0-
2017-01-10 18:28:26 +08:00
|
|
|
|
* This function is invoked towards the end of the scheduler's
|
|
|
|
|
* initialization process. Before this is called, the idle task might
|
|
|
|
|
* contain synchronous grace-period primitives (during which time, this idle
|
|
|
|
|
* task is booting the system, and such primitives are no-ops). After this
|
|
|
|
|
* function is called, any synchronous grace-period primitives are run as
|
|
|
|
|
* expedited, with the requesting task driving the grace period forward.
|
2017-02-11 06:32:54 +08:00
|
|
|
|
* A later core_initcall() rcu_set_runtime_mode() will switch to full
|
rcu: Narrow early boot window of illegal synchronous grace periods
The current preemptible RCU implementation goes through three phases
during bootup. In the first phase, there is only one CPU that is running
with preemption disabled, so that a no-op is a synchronous grace period.
In the second mid-boot phase, the scheduler is running, but RCU has
not yet gotten its kthreads spawned (and, for expedited grace periods,
workqueues are not yet running. During this time, any attempt to do
a synchronous grace period will hang the system (or complain bitterly,
depending). In the third and final phase, RCU is fully operational and
everything works normally.
This has been OK for some time, but there has recently been some
synchronous grace periods showing up during the second mid-boot phase.
This code worked "by accident" for awhile, but started failing as soon
as expedited RCU grace periods switched over to workqueues in commit
8b355e3bc140 ("rcu: Drive expedited grace periods from workqueue").
Note that the code was buggy even before this commit, as it was subject
to failure on real-time systems that forced all expedited grace periods
to run as normal grace periods (for example, using the rcu_normal ksysfs
parameter). The callchain from the failure case is as follows:
early_amd_iommu_init()
|-> acpi_put_table(ivrs_base);
|-> acpi_tb_put_table(table_desc);
|-> acpi_tb_invalidate_table(table_desc);
|-> acpi_tb_release_table(...)
|-> acpi_os_unmap_memory
|-> acpi_os_unmap_iomem
|-> acpi_os_map_cleanup
|-> synchronize_rcu_expedited
The kernel showing this callchain was built with CONFIG_PREEMPT_RCU=y,
which caused the code to try using workqueues before they were
initialized, which did not go well.
This commit therefore reworks RCU to permit synchronous grace periods
to proceed during this mid-boot phase. This commit is therefore a
fix to a regression introduced in v4.9, and is therefore being put
forward post-merge-window in v4.10.
This commit sets a flag from the existing rcu_scheduler_starting()
function which causes all synchronous grace periods to take the expedited
path. The expedited path now checks this flag, using the requesting task
to drive the expedited grace period forward during the mid-boot phase.
Finally, this flag is updated by a core_initcall() function named
rcu_exp_runtime_mode(), which causes the runtime codepaths to be used.
Note that this arrangement assumes that tasks are not sent POSIX signals
(or anything similar) from the time that the first task is spawned
through core_initcall() time.
Fixes: 8b355e3bc140 ("rcu: Drive expedited grace periods from workqueue")
Reported-by: "Zheng, Lv" <lv.zheng@intel.com>
Reported-by: Borislav Petkov <bp@alien8.de>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Stan Kain <stan.kain@gmail.com>
Tested-by: Ivan <waffolz@hotmail.com>
Tested-by: Emanuel Castelo <emanuel.castelo@gmail.com>
Tested-by: Bruno Pesavento <bpesavento@infinito.it>
Tested-by: Borislav Petkov <bp@suse.de>
Tested-by: Frederic Bezies <fredbezies@gmail.com>
Cc: <stable@vger.kernel.org> # 4.9.0-
2017-01-10 18:28:26 +08:00
|
|
|
|
* runtime RCU functionality.
|
2010-04-03 07:17:17 +08:00
|
|
|
|
*/
|
|
|
|
|
void rcu_scheduler_starting(void)
|
|
|
|
|
{
|
2022-08-06 06:42:25 +08:00
|
|
|
|
unsigned long flags;
|
|
|
|
|
struct rcu_node *rnp;
|
|
|
|
|
|
2010-04-03 07:17:17 +08:00
|
|
|
|
WARN_ON(num_online_cpus() != 1);
|
|
|
|
|
WARN_ON(nr_context_switches() > 0);
|
rcu: Narrow early boot window of illegal synchronous grace periods
The current preemptible RCU implementation goes through three phases
during bootup. In the first phase, there is only one CPU that is running
with preemption disabled, so that a no-op is a synchronous grace period.
In the second mid-boot phase, the scheduler is running, but RCU has
not yet gotten its kthreads spawned (and, for expedited grace periods,
workqueues are not yet running. During this time, any attempt to do
a synchronous grace period will hang the system (or complain bitterly,
depending). In the third and final phase, RCU is fully operational and
everything works normally.
This has been OK for some time, but there has recently been some
synchronous grace periods showing up during the second mid-boot phase.
This code worked "by accident" for awhile, but started failing as soon
as expedited RCU grace periods switched over to workqueues in commit
8b355e3bc140 ("rcu: Drive expedited grace periods from workqueue").
Note that the code was buggy even before this commit, as it was subject
to failure on real-time systems that forced all expedited grace periods
to run as normal grace periods (for example, using the rcu_normal ksysfs
parameter). The callchain from the failure case is as follows:
early_amd_iommu_init()
|-> acpi_put_table(ivrs_base);
|-> acpi_tb_put_table(table_desc);
|-> acpi_tb_invalidate_table(table_desc);
|-> acpi_tb_release_table(...)
|-> acpi_os_unmap_memory
|-> acpi_os_unmap_iomem
|-> acpi_os_map_cleanup
|-> synchronize_rcu_expedited
The kernel showing this callchain was built with CONFIG_PREEMPT_RCU=y,
which caused the code to try using workqueues before they were
initialized, which did not go well.
This commit therefore reworks RCU to permit synchronous grace periods
to proceed during this mid-boot phase. This commit is therefore a
fix to a regression introduced in v4.9, and is therefore being put
forward post-merge-window in v4.10.
This commit sets a flag from the existing rcu_scheduler_starting()
function which causes all synchronous grace periods to take the expedited
path. The expedited path now checks this flag, using the requesting task
to drive the expedited grace period forward during the mid-boot phase.
Finally, this flag is updated by a core_initcall() function named
rcu_exp_runtime_mode(), which causes the runtime codepaths to be used.
Note that this arrangement assumes that tasks are not sent POSIX signals
(or anything similar) from the time that the first task is spawned
through core_initcall() time.
Fixes: 8b355e3bc140 ("rcu: Drive expedited grace periods from workqueue")
Reported-by: "Zheng, Lv" <lv.zheng@intel.com>
Reported-by: Borislav Petkov <bp@alien8.de>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Stan Kain <stan.kain@gmail.com>
Tested-by: Ivan <waffolz@hotmail.com>
Tested-by: Emanuel Castelo <emanuel.castelo@gmail.com>
Tested-by: Bruno Pesavento <bpesavento@infinito.it>
Tested-by: Borislav Petkov <bp@suse.de>
Tested-by: Frederic Bezies <fredbezies@gmail.com>
Cc: <stable@vger.kernel.org> # 4.9.0-
2017-01-10 18:28:26 +08:00
|
|
|
|
rcu_test_sync_prims();
|
2022-08-06 06:42:25 +08:00
|
|
|
|
|
|
|
|
|
// Fix up the ->gp_seq counters.
|
|
|
|
|
local_irq_save(flags);
|
|
|
|
|
rcu_for_each_node_breadth_first(rnp)
|
|
|
|
|
rnp->gp_seq_needed = rnp->gp_seq = rcu_state.gp_seq;
|
|
|
|
|
local_irq_restore(flags);
|
|
|
|
|
|
|
|
|
|
// Switch out of early boot mode.
|
rcu: Narrow early boot window of illegal synchronous grace periods
The current preemptible RCU implementation goes through three phases
during bootup. In the first phase, there is only one CPU that is running
with preemption disabled, so that a no-op is a synchronous grace period.
In the second mid-boot phase, the scheduler is running, but RCU has
not yet gotten its kthreads spawned (and, for expedited grace periods,
workqueues are not yet running. During this time, any attempt to do
a synchronous grace period will hang the system (or complain bitterly,
depending). In the third and final phase, RCU is fully operational and
everything works normally.
This has been OK for some time, but there has recently been some
synchronous grace periods showing up during the second mid-boot phase.
This code worked "by accident" for awhile, but started failing as soon
as expedited RCU grace periods switched over to workqueues in commit
8b355e3bc140 ("rcu: Drive expedited grace periods from workqueue").
Note that the code was buggy even before this commit, as it was subject
to failure on real-time systems that forced all expedited grace periods
to run as normal grace periods (for example, using the rcu_normal ksysfs
parameter). The callchain from the failure case is as follows:
early_amd_iommu_init()
|-> acpi_put_table(ivrs_base);
|-> acpi_tb_put_table(table_desc);
|-> acpi_tb_invalidate_table(table_desc);
|-> acpi_tb_release_table(...)
|-> acpi_os_unmap_memory
|-> acpi_os_unmap_iomem
|-> acpi_os_map_cleanup
|-> synchronize_rcu_expedited
The kernel showing this callchain was built with CONFIG_PREEMPT_RCU=y,
which caused the code to try using workqueues before they were
initialized, which did not go well.
This commit therefore reworks RCU to permit synchronous grace periods
to proceed during this mid-boot phase. This commit is therefore a
fix to a regression introduced in v4.9, and is therefore being put
forward post-merge-window in v4.10.
This commit sets a flag from the existing rcu_scheduler_starting()
function which causes all synchronous grace periods to take the expedited
path. The expedited path now checks this flag, using the requesting task
to drive the expedited grace period forward during the mid-boot phase.
Finally, this flag is updated by a core_initcall() function named
rcu_exp_runtime_mode(), which causes the runtime codepaths to be used.
Note that this arrangement assumes that tasks are not sent POSIX signals
(or anything similar) from the time that the first task is spawned
through core_initcall() time.
Fixes: 8b355e3bc140 ("rcu: Drive expedited grace periods from workqueue")
Reported-by: "Zheng, Lv" <lv.zheng@intel.com>
Reported-by: Borislav Petkov <bp@alien8.de>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Stan Kain <stan.kain@gmail.com>
Tested-by: Ivan <waffolz@hotmail.com>
Tested-by: Emanuel Castelo <emanuel.castelo@gmail.com>
Tested-by: Bruno Pesavento <bpesavento@infinito.it>
Tested-by: Borislav Petkov <bp@suse.de>
Tested-by: Frederic Bezies <fredbezies@gmail.com>
Cc: <stable@vger.kernel.org> # 4.9.0-
2017-01-10 18:28:26 +08:00
|
|
|
|
rcu_scheduler_active = RCU_SCHEDULER_INIT;
|
|
|
|
|
rcu_test_sync_prims();
|
2010-04-03 07:17:17 +08:00
|
|
|
|
}
|
|
|
|
|
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
/*
|
2018-07-08 09:12:26 +08:00
|
|
|
|
* Helper function for rcu_init() that initializes the rcu_state structure.
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
*/
|
2018-07-04 08:22:34 +08:00
|
|
|
|
static void __init rcu_init_one(void)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
{
|
2015-06-03 14:18:30 +08:00
|
|
|
|
static const char * const buf[] = RCU_NODE_NAME_INIT;
|
|
|
|
|
static const char * const fqs[] = RCU_FQS_NAME_INIT;
|
2015-09-27 05:51:24 +08:00
|
|
|
|
static struct lock_class_key rcu_node_class[RCU_NUM_LVLS];
|
|
|
|
|
static struct lock_class_key rcu_fqs_class[RCU_NUM_LVLS];
|
2015-06-03 14:18:29 +08:00
|
|
|
|
|
|
|
|
|
int levelspread[RCU_NUM_LVLS]; /* kids/node in each level. */
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
int cpustride = 1;
|
|
|
|
|
int i;
|
|
|
|
|
int j;
|
|
|
|
|
struct rcu_node *rnp;
|
|
|
|
|
|
2015-06-03 14:18:28 +08:00
|
|
|
|
BUILD_BUG_ON(RCU_NUM_LVLS > ARRAY_SIZE(buf)); /* Fix buf[] init! */
|
2010-01-05 08:04:02 +08:00
|
|
|
|
|
2015-03-10 07:51:17 +08:00
|
|
|
|
/* Silence gcc 4.8 false positive about array index out of range. */
|
|
|
|
|
if (rcu_num_lvls <= 0 || rcu_num_lvls > RCU_NUM_LVLS)
|
|
|
|
|
panic("rcu_init_one: rcu_num_lvls out of range");
|
2012-11-30 05:49:00 +08:00
|
|
|
|
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
/* Initialize the level-tracking arrays. */
|
|
|
|
|
|
2012-04-24 06:52:53 +08:00
|
|
|
|
for (i = 1; i < rcu_num_lvls; i++)
|
2018-07-06 08:47:45 +08:00
|
|
|
|
rcu_state.level[i] =
|
|
|
|
|
rcu_state.level[i - 1] + num_rcu_lvl[i - 1];
|
2017-03-16 03:59:17 +08:00
|
|
|
|
rcu_init_levelspread(levelspread, num_rcu_lvl);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
|
|
|
|
|
/* Initialize the elements themselves, starting from the leaves. */
|
|
|
|
|
|
2012-04-24 06:52:53 +08:00
|
|
|
|
for (i = rcu_num_lvls - 1; i >= 0; i--) {
|
2015-06-03 14:18:29 +08:00
|
|
|
|
cpustride *= levelspread[i];
|
2018-07-06 08:47:45 +08:00
|
|
|
|
rnp = rcu_state.level[i];
|
2017-03-16 03:59:17 +08:00
|
|
|
|
for (j = 0; j < num_rcu_lvl[i]; j++, rnp++) {
|
2015-12-29 12:18:47 +08:00
|
|
|
|
raw_spin_lock_init(&ACCESS_PRIVATE(rnp, lock));
|
|
|
|
|
lockdep_set_class_and_name(&ACCESS_PRIVATE(rnp, lock),
|
2010-01-05 08:04:02 +08:00
|
|
|
|
&rcu_node_class[i], buf[i]);
|
2012-06-27 08:00:35 +08:00
|
|
|
|
raw_spin_lock_init(&rnp->fqslock);
|
|
|
|
|
lockdep_set_class_and_name(&rnp->fqslock,
|
|
|
|
|
&rcu_fqs_class[i], fqs[i]);
|
2018-07-06 08:47:45 +08:00
|
|
|
|
rnp->gp_seq = rcu_state.gp_seq;
|
|
|
|
|
rnp->gp_seq_needed = rcu_state.gp_seq;
|
|
|
|
|
rnp->completedqs = rcu_state.gp_seq;
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
rnp->qsmask = 0;
|
|
|
|
|
rnp->qsmaskinit = 0;
|
|
|
|
|
rnp->grplo = j * cpustride;
|
|
|
|
|
rnp->grphi = (j + 1) * cpustride - 1;
|
2014-03-19 01:22:26 +08:00
|
|
|
|
if (rnp->grphi >= nr_cpu_ids)
|
|
|
|
|
rnp->grphi = nr_cpu_ids - 1;
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
if (i == 0) {
|
|
|
|
|
rnp->grpnum = 0;
|
|
|
|
|
rnp->grpmask = 0;
|
|
|
|
|
rnp->parent = NULL;
|
|
|
|
|
} else {
|
2015-06-03 14:18:29 +08:00
|
|
|
|
rnp->grpnum = j % levelspread[i - 1];
|
2018-08-01 00:49:20 +08:00
|
|
|
|
rnp->grpmask = BIT(rnp->grpnum);
|
2018-07-06 08:47:45 +08:00
|
|
|
|
rnp->parent = rcu_state.level[i - 1] +
|
2015-06-03 14:18:29 +08:00
|
|
|
|
j / levelspread[i - 1];
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
|
|
|
|
rnp->level = i;
|
2010-11-30 13:56:39 +08:00
|
|
|
|
INIT_LIST_HEAD(&rnp->blkd_tasks);
|
2013-02-11 12:48:58 +08:00
|
|
|
|
rcu_init_one_nocb(rnp);
|
2016-01-31 09:57:35 +08:00
|
|
|
|
init_waitqueue_head(&rnp->exp_wq[0]);
|
|
|
|
|
init_waitqueue_head(&rnp->exp_wq[1]);
|
2016-03-17 07:47:55 +08:00
|
|
|
|
init_waitqueue_head(&rnp->exp_wq[2]);
|
|
|
|
|
init_waitqueue_head(&rnp->exp_wq[3]);
|
2016-01-31 09:57:35 +08:00
|
|
|
|
spin_lock_init(&rnp->exp_lock);
|
2021-12-09 07:41:53 +08:00
|
|
|
|
mutex_init(&rnp->boost_kthread_mutex);
|
2022-04-16 01:55:42 +08:00
|
|
|
|
raw_spin_lock_init(&rnp->exp_poll_lock);
|
|
|
|
|
rnp->exp_seq_poll_rq = RCU_GET_STATE_COMPLETED;
|
|
|
|
|
INIT_WORK(&rnp->exp_poll_wq, sync_rcu_do_polled_gp);
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
|
|
|
|
}
|
2010-03-28 11:12:30 +08:00
|
|
|
|
|
2018-07-06 08:47:45 +08:00
|
|
|
|
init_swait_queue_head(&rcu_state.gp_wq);
|
|
|
|
|
init_swait_queue_head(&rcu_state.expedited_wq);
|
2018-07-05 05:33:59 +08:00
|
|
|
|
rnp = rcu_first_leaf_node();
|
2010-03-28 11:12:30 +08:00
|
|
|
|
for_each_possible_cpu(i) {
|
2010-04-15 07:48:11 +08:00
|
|
|
|
while (i > rnp->grphi)
|
2010-03-28 11:12:30 +08:00
|
|
|
|
rnp++;
|
2018-07-04 06:37:16 +08:00
|
|
|
|
per_cpu_ptr(&rcu_data, i)->mynode = rnp;
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_boot_init_percpu_data(i);
|
2010-03-28 11:12:30 +08:00
|
|
|
|
}
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
|
|
|
|
|
2022-01-12 07:32:50 +08:00
|
|
|
|
/*
|
|
|
|
|
* Force priority from the kernel command-line into range.
|
|
|
|
|
*/
|
|
|
|
|
static void __init sanitize_kthread_prio(void)
|
|
|
|
|
{
|
|
|
|
|
int kthread_prio_in = kthread_prio;
|
|
|
|
|
|
|
|
|
|
if (IS_ENABLED(CONFIG_RCU_BOOST) && kthread_prio < 2
|
|
|
|
|
&& IS_BUILTIN(CONFIG_RCU_TORTURE_TEST))
|
|
|
|
|
kthread_prio = 2;
|
|
|
|
|
else if (IS_ENABLED(CONFIG_RCU_BOOST) && kthread_prio < 1)
|
|
|
|
|
kthread_prio = 1;
|
|
|
|
|
else if (kthread_prio < 0)
|
|
|
|
|
kthread_prio = 0;
|
|
|
|
|
else if (kthread_prio > 99)
|
|
|
|
|
kthread_prio = 99;
|
|
|
|
|
|
|
|
|
|
if (kthread_prio != kthread_prio_in)
|
|
|
|
|
pr_alert("%s: Limited prio to %d from %d\n",
|
|
|
|
|
__func__, kthread_prio, kthread_prio_in);
|
|
|
|
|
}
|
|
|
|
|
|
2012-04-24 06:52:53 +08:00
|
|
|
|
/*
|
|
|
|
|
* Compute the rcu_node tree geometry from kernel parameters. This cannot
|
2013-10-09 11:23:47 +08:00
|
|
|
|
* replace the definitions in tree.h because those are needed to size
|
2012-04-24 06:52:53 +08:00
|
|
|
|
* the ->node array in the rcu_state structure.
|
|
|
|
|
*/
|
srcu: Fix broken node geometry after early ssp init
An srcu_struct structure that is initialized before rcu_init_geometry()
will have its srcu_node hierarchy based on CONFIG_NR_CPUS. Once
rcu_init_geometry() is called, this hierarchy is compressed as needed
for the actual maximum number of CPUs for this system.
Later on, that srcu_struct structure is confused, sometimes referring
to its initial CONFIG_NR_CPUS-based hierarchy, and sometimes instead
to the new num_possible_cpus() hierarchy. For example, each of its
->mynode fields continues to reference the original leaf rcu_node
structures, some of which might no longer exist. On the other hand,
srcu_for_each_node_breadth_first() traverses to the new node hierarchy.
There are at least two bad possible outcomes to this:
1) a) A callback enqueued early on an srcu_data structure (call it
*sdp) is recorded pending on sdp->mynode->srcu_data_have_cbs in
srcu_funnel_gp_start() with sdp->mynode pointing to a deep leaf
(say 3 levels).
b) The grace period ends after rcu_init_geometry() shrinks the
nodes level to a single one. srcu_gp_end() walks through the new
srcu_node hierarchy without ever reaching the old leaves so the
callback is never executed.
This is easily reproduced on an 8 CPUs machine with CONFIG_NR_CPUS >= 32
and "rcupdate.rcu_self_test=1". The srcu_barrier() after early tests
verification never completes and the boot hangs:
[ 5413.141029] INFO: task swapper/0:1 blocked for more than 4915 seconds.
[ 5413.147564] Not tainted 5.12.0-rc4+ #28
[ 5413.151927] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[ 5413.159753] task:swapper/0 state:D stack: 0 pid: 1 ppid: 0 flags:0x00004000
[ 5413.168099] Call Trace:
[ 5413.170555] __schedule+0x36c/0x930
[ 5413.174057] ? wait_for_completion+0x88/0x110
[ 5413.178423] schedule+0x46/0xf0
[ 5413.181575] schedule_timeout+0x284/0x380
[ 5413.185591] ? wait_for_completion+0x88/0x110
[ 5413.189957] ? mark_held_locks+0x61/0x80
[ 5413.193882] ? mark_held_locks+0x61/0x80
[ 5413.197809] ? _raw_spin_unlock_irq+0x24/0x50
[ 5413.202173] ? wait_for_completion+0x88/0x110
[ 5413.206535] wait_for_completion+0xb4/0x110
[ 5413.210724] ? srcu_torture_stats_print+0x110/0x110
[ 5413.215610] srcu_barrier+0x187/0x200
[ 5413.219277] ? rcu_tasks_verify_self_tests+0x50/0x50
[ 5413.224244] ? rdinit_setup+0x2b/0x2b
[ 5413.227907] rcu_verify_early_boot_tests+0x2d/0x40
[ 5413.232700] do_one_initcall+0x63/0x310
[ 5413.236541] ? rdinit_setup+0x2b/0x2b
[ 5413.240207] ? rcu_read_lock_sched_held+0x52/0x80
[ 5413.244912] kernel_init_freeable+0x253/0x28f
[ 5413.249273] ? rest_init+0x250/0x250
[ 5413.252846] kernel_init+0xa/0x110
[ 5413.256257] ret_from_fork+0x22/0x30
2) An srcu_struct structure that is initialized before rcu_init_geometry()
and used afterward will always have stale rdp->mynode references,
resulting in callbacks to be missed in srcu_gp_end(), just like in
the previous scenario.
This commit therefore causes init_srcu_struct_nodes to initialize the
geometry, if needed. This ensures that the srcu_node hierarchy is
properly built and distributed from the get-go.
Suggested-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Cc: Boqun Feng <boqun.feng@gmail.com>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: Neeraj Upadhyay <neeraju@codeaurora.org>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Joel Fernandes <joel@joelfernandes.org>
Cc: Uladzislau Rezki <urezki@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-04-17 21:16:49 +08:00
|
|
|
|
void rcu_init_geometry(void)
|
2012-04-24 06:52:53 +08:00
|
|
|
|
{
|
2013-04-04 13:14:11 +08:00
|
|
|
|
ulong d;
|
2012-04-24 06:52:53 +08:00
|
|
|
|
int i;
|
srcu: Fix broken node geometry after early ssp init
An srcu_struct structure that is initialized before rcu_init_geometry()
will have its srcu_node hierarchy based on CONFIG_NR_CPUS. Once
rcu_init_geometry() is called, this hierarchy is compressed as needed
for the actual maximum number of CPUs for this system.
Later on, that srcu_struct structure is confused, sometimes referring
to its initial CONFIG_NR_CPUS-based hierarchy, and sometimes instead
to the new num_possible_cpus() hierarchy. For example, each of its
->mynode fields continues to reference the original leaf rcu_node
structures, some of which might no longer exist. On the other hand,
srcu_for_each_node_breadth_first() traverses to the new node hierarchy.
There are at least two bad possible outcomes to this:
1) a) A callback enqueued early on an srcu_data structure (call it
*sdp) is recorded pending on sdp->mynode->srcu_data_have_cbs in
srcu_funnel_gp_start() with sdp->mynode pointing to a deep leaf
(say 3 levels).
b) The grace period ends after rcu_init_geometry() shrinks the
nodes level to a single one. srcu_gp_end() walks through the new
srcu_node hierarchy without ever reaching the old leaves so the
callback is never executed.
This is easily reproduced on an 8 CPUs machine with CONFIG_NR_CPUS >= 32
and "rcupdate.rcu_self_test=1". The srcu_barrier() after early tests
verification never completes and the boot hangs:
[ 5413.141029] INFO: task swapper/0:1 blocked for more than 4915 seconds.
[ 5413.147564] Not tainted 5.12.0-rc4+ #28
[ 5413.151927] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[ 5413.159753] task:swapper/0 state:D stack: 0 pid: 1 ppid: 0 flags:0x00004000
[ 5413.168099] Call Trace:
[ 5413.170555] __schedule+0x36c/0x930
[ 5413.174057] ? wait_for_completion+0x88/0x110
[ 5413.178423] schedule+0x46/0xf0
[ 5413.181575] schedule_timeout+0x284/0x380
[ 5413.185591] ? wait_for_completion+0x88/0x110
[ 5413.189957] ? mark_held_locks+0x61/0x80
[ 5413.193882] ? mark_held_locks+0x61/0x80
[ 5413.197809] ? _raw_spin_unlock_irq+0x24/0x50
[ 5413.202173] ? wait_for_completion+0x88/0x110
[ 5413.206535] wait_for_completion+0xb4/0x110
[ 5413.210724] ? srcu_torture_stats_print+0x110/0x110
[ 5413.215610] srcu_barrier+0x187/0x200
[ 5413.219277] ? rcu_tasks_verify_self_tests+0x50/0x50
[ 5413.224244] ? rdinit_setup+0x2b/0x2b
[ 5413.227907] rcu_verify_early_boot_tests+0x2d/0x40
[ 5413.232700] do_one_initcall+0x63/0x310
[ 5413.236541] ? rdinit_setup+0x2b/0x2b
[ 5413.240207] ? rcu_read_lock_sched_held+0x52/0x80
[ 5413.244912] kernel_init_freeable+0x253/0x28f
[ 5413.249273] ? rest_init+0x250/0x250
[ 5413.252846] kernel_init+0xa/0x110
[ 5413.256257] ret_from_fork+0x22/0x30
2) An srcu_struct structure that is initialized before rcu_init_geometry()
and used afterward will always have stale rdp->mynode references,
resulting in callbacks to be missed in srcu_gp_end(), just like in
the previous scenario.
This commit therefore causes init_srcu_struct_nodes to initialize the
geometry, if needed. This ensures that the srcu_node hierarchy is
properly built and distributed from the get-go.
Suggested-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Cc: Boqun Feng <boqun.feng@gmail.com>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: Neeraj Upadhyay <neeraju@codeaurora.org>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Joel Fernandes <joel@joelfernandes.org>
Cc: Uladzislau Rezki <urezki@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-04-17 21:16:49 +08:00
|
|
|
|
static unsigned long old_nr_cpu_ids;
|
2015-06-03 14:18:28 +08:00
|
|
|
|
int rcu_capacity[RCU_NUM_LVLS];
|
srcu: Fix broken node geometry after early ssp init
An srcu_struct structure that is initialized before rcu_init_geometry()
will have its srcu_node hierarchy based on CONFIG_NR_CPUS. Once
rcu_init_geometry() is called, this hierarchy is compressed as needed
for the actual maximum number of CPUs for this system.
Later on, that srcu_struct structure is confused, sometimes referring
to its initial CONFIG_NR_CPUS-based hierarchy, and sometimes instead
to the new num_possible_cpus() hierarchy. For example, each of its
->mynode fields continues to reference the original leaf rcu_node
structures, some of which might no longer exist. On the other hand,
srcu_for_each_node_breadth_first() traverses to the new node hierarchy.
There are at least two bad possible outcomes to this:
1) a) A callback enqueued early on an srcu_data structure (call it
*sdp) is recorded pending on sdp->mynode->srcu_data_have_cbs in
srcu_funnel_gp_start() with sdp->mynode pointing to a deep leaf
(say 3 levels).
b) The grace period ends after rcu_init_geometry() shrinks the
nodes level to a single one. srcu_gp_end() walks through the new
srcu_node hierarchy without ever reaching the old leaves so the
callback is never executed.
This is easily reproduced on an 8 CPUs machine with CONFIG_NR_CPUS >= 32
and "rcupdate.rcu_self_test=1". The srcu_barrier() after early tests
verification never completes and the boot hangs:
[ 5413.141029] INFO: task swapper/0:1 blocked for more than 4915 seconds.
[ 5413.147564] Not tainted 5.12.0-rc4+ #28
[ 5413.151927] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[ 5413.159753] task:swapper/0 state:D stack: 0 pid: 1 ppid: 0 flags:0x00004000
[ 5413.168099] Call Trace:
[ 5413.170555] __schedule+0x36c/0x930
[ 5413.174057] ? wait_for_completion+0x88/0x110
[ 5413.178423] schedule+0x46/0xf0
[ 5413.181575] schedule_timeout+0x284/0x380
[ 5413.185591] ? wait_for_completion+0x88/0x110
[ 5413.189957] ? mark_held_locks+0x61/0x80
[ 5413.193882] ? mark_held_locks+0x61/0x80
[ 5413.197809] ? _raw_spin_unlock_irq+0x24/0x50
[ 5413.202173] ? wait_for_completion+0x88/0x110
[ 5413.206535] wait_for_completion+0xb4/0x110
[ 5413.210724] ? srcu_torture_stats_print+0x110/0x110
[ 5413.215610] srcu_barrier+0x187/0x200
[ 5413.219277] ? rcu_tasks_verify_self_tests+0x50/0x50
[ 5413.224244] ? rdinit_setup+0x2b/0x2b
[ 5413.227907] rcu_verify_early_boot_tests+0x2d/0x40
[ 5413.232700] do_one_initcall+0x63/0x310
[ 5413.236541] ? rdinit_setup+0x2b/0x2b
[ 5413.240207] ? rcu_read_lock_sched_held+0x52/0x80
[ 5413.244912] kernel_init_freeable+0x253/0x28f
[ 5413.249273] ? rest_init+0x250/0x250
[ 5413.252846] kernel_init+0xa/0x110
[ 5413.256257] ret_from_fork+0x22/0x30
2) An srcu_struct structure that is initialized before rcu_init_geometry()
and used afterward will always have stale rdp->mynode references,
resulting in callbacks to be missed in srcu_gp_end(), just like in
the previous scenario.
This commit therefore causes init_srcu_struct_nodes to initialize the
geometry, if needed. This ensures that the srcu_node hierarchy is
properly built and distributed from the get-go.
Suggested-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Cc: Boqun Feng <boqun.feng@gmail.com>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: Neeraj Upadhyay <neeraju@codeaurora.org>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Joel Fernandes <joel@joelfernandes.org>
Cc: Uladzislau Rezki <urezki@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2021-04-17 21:16:49 +08:00
|
|
|
|
static bool initialized;
|
|
|
|
|
|
|
|
|
|
if (initialized) {
|
|
|
|
|
/*
|
|
|
|
|
* Warn if setup_nr_cpu_ids() had not yet been invoked,
|
|
|
|
|
* unless nr_cpus_ids == NR_CPUS, in which case who cares?
|
|
|
|
|
*/
|
|
|
|
|
WARN_ON_ONCE(old_nr_cpu_ids != nr_cpu_ids);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
old_nr_cpu_ids = nr_cpu_ids;
|
|
|
|
|
initialized = true;
|
2012-04-24 06:52:53 +08:00
|
|
|
|
|
2013-04-04 13:14:11 +08:00
|
|
|
|
/*
|
|
|
|
|
* Initialize any unspecified boot parameters.
|
|
|
|
|
* The default values of jiffies_till_first_fqs and
|
|
|
|
|
* jiffies_till_next_fqs are set to the RCU_JIFFIES_TILL_FORCE_QS
|
|
|
|
|
* value, which is a function of HZ, then adding one for each
|
|
|
|
|
* RCU_JIFFIES_FQS_DIV CPUs that might be on the system.
|
|
|
|
|
*/
|
|
|
|
|
d = RCU_JIFFIES_TILL_FORCE_QS + nr_cpu_ids / RCU_JIFFIES_FQS_DIV;
|
|
|
|
|
if (jiffies_till_first_fqs == ULONG_MAX)
|
|
|
|
|
jiffies_till_first_fqs = d;
|
|
|
|
|
if (jiffies_till_next_fqs == ULONG_MAX)
|
|
|
|
|
jiffies_till_next_fqs = d;
|
2019-03-11 17:46:11 +08:00
|
|
|
|
adjust_jiffies_till_sched_qs();
|
2013-04-04 13:14:11 +08:00
|
|
|
|
|
2012-04-24 06:52:53 +08:00
|
|
|
|
/* If the compile-time values are accurate, just leave. */
|
2015-04-22 00:12:13 +08:00
|
|
|
|
if (rcu_fanout_leaf == RCU_FANOUT_LEAF &&
|
2012-09-07 06:38:02 +08:00
|
|
|
|
nr_cpu_ids == NR_CPUS)
|
2012-04-24 06:52:53 +08:00
|
|
|
|
return;
|
2018-05-15 04:27:33 +08:00
|
|
|
|
pr_info("Adjusting geometry for rcu_fanout_leaf=%d, nr_cpu_ids=%u\n",
|
2013-10-10 06:20:33 +08:00
|
|
|
|
rcu_fanout_leaf, nr_cpu_ids);
|
2012-04-24 06:52:53 +08:00
|
|
|
|
|
|
|
|
|
/*
|
2015-07-31 23:28:35 +08:00
|
|
|
|
* The boot-time rcu_fanout_leaf parameter must be at least two
|
|
|
|
|
* and cannot exceed the number of bits in the rcu_node masks.
|
|
|
|
|
* Complain and fall back to the compile-time values if this
|
|
|
|
|
* limit is exceeded.
|
2012-04-24 06:52:53 +08:00
|
|
|
|
*/
|
2015-07-31 23:28:35 +08:00
|
|
|
|
if (rcu_fanout_leaf < 2 ||
|
2015-06-03 14:18:23 +08:00
|
|
|
|
rcu_fanout_leaf > sizeof(unsigned long) * 8) {
|
2015-06-05 01:06:01 +08:00
|
|
|
|
rcu_fanout_leaf = RCU_FANOUT_LEAF;
|
2012-04-24 06:52:53 +08:00
|
|
|
|
WARN_ON(1);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Compute number of nodes that can be handled an rcu_node tree
|
2015-06-03 14:18:26 +08:00
|
|
|
|
* with the given number of levels.
|
2012-04-24 06:52:53 +08:00
|
|
|
|
*/
|
2015-06-03 14:18:26 +08:00
|
|
|
|
rcu_capacity[0] = rcu_fanout_leaf;
|
2015-06-03 14:18:28 +08:00
|
|
|
|
for (i = 1; i < RCU_NUM_LVLS; i++)
|
2015-04-21 05:27:43 +08:00
|
|
|
|
rcu_capacity[i] = rcu_capacity[i - 1] * RCU_FANOUT;
|
2012-04-24 06:52:53 +08:00
|
|
|
|
|
|
|
|
|
/*
|
2015-06-03 14:18:23 +08:00
|
|
|
|
* The tree must be able to accommodate the configured number of CPUs.
|
2015-07-31 23:28:35 +08:00
|
|
|
|
* If this limit is exceeded, fall back to the compile-time values.
|
2012-04-24 06:52:53 +08:00
|
|
|
|
*/
|
2015-07-31 23:28:35 +08:00
|
|
|
|
if (nr_cpu_ids > rcu_capacity[RCU_NUM_LVLS - 1]) {
|
|
|
|
|
rcu_fanout_leaf = RCU_FANOUT_LEAF;
|
|
|
|
|
WARN_ON(1);
|
|
|
|
|
return;
|
|
|
|
|
}
|
2012-04-24 06:52:53 +08:00
|
|
|
|
|
2015-06-03 14:18:25 +08:00
|
|
|
|
/* Calculate the number of levels in the tree. */
|
2015-06-03 14:18:26 +08:00
|
|
|
|
for (i = 0; nr_cpu_ids > rcu_capacity[i]; i++) {
|
2015-06-03 14:18:25 +08:00
|
|
|
|
}
|
2015-06-03 14:18:26 +08:00
|
|
|
|
rcu_num_lvls = i + 1;
|
2015-06-03 14:18:25 +08:00
|
|
|
|
|
2012-04-24 06:52:53 +08:00
|
|
|
|
/* Calculate the number of rcu_nodes at each level of the tree. */
|
2015-06-03 14:18:25 +08:00
|
|
|
|
for (i = 0; i < rcu_num_lvls; i++) {
|
2015-06-03 14:18:26 +08:00
|
|
|
|
int cap = rcu_capacity[(rcu_num_lvls - 1) - i];
|
2015-06-03 14:18:25 +08:00
|
|
|
|
num_rcu_lvl[i] = DIV_ROUND_UP(nr_cpu_ids, cap);
|
|
|
|
|
}
|
2012-04-24 06:52:53 +08:00
|
|
|
|
|
|
|
|
|
/* Calculate the total number of rcu_node structures. */
|
|
|
|
|
rcu_num_nodes = 0;
|
2015-06-03 14:18:25 +08:00
|
|
|
|
for (i = 0; i < rcu_num_lvls; i++)
|
2012-04-24 06:52:53 +08:00
|
|
|
|
rcu_num_nodes += num_rcu_lvl[i];
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-21 02:40:50 +08:00
|
|
|
|
/*
|
|
|
|
|
* Dump out the structure of the rcu_node combining tree associated
|
2018-07-08 09:12:26 +08:00
|
|
|
|
* with the rcu_state structure.
|
2015-04-21 02:40:50 +08:00
|
|
|
|
*/
|
2018-07-04 08:22:34 +08:00
|
|
|
|
static void __init rcu_dump_rcu_node_tree(void)
|
2015-04-21 02:40:50 +08:00
|
|
|
|
{
|
|
|
|
|
int level = 0;
|
|
|
|
|
struct rcu_node *rnp;
|
|
|
|
|
|
|
|
|
|
pr_info("rcu_node tree layout dump\n");
|
|
|
|
|
pr_info(" ");
|
2018-07-05 05:33:59 +08:00
|
|
|
|
rcu_for_each_node_breadth_first(rnp) {
|
2015-04-21 02:40:50 +08:00
|
|
|
|
if (rnp->level != level) {
|
|
|
|
|
pr_cont("\n");
|
|
|
|
|
pr_info(" ");
|
|
|
|
|
level = rnp->level;
|
|
|
|
|
}
|
|
|
|
|
pr_cont("%d:%d ^%d ", rnp->grplo, rnp->grphi, rnp->grpnum);
|
|
|
|
|
}
|
|
|
|
|
pr_cont("\n");
|
|
|
|
|
}
|
|
|
|
|
|
2018-01-09 06:35:52 +08:00
|
|
|
|
struct workqueue_struct *rcu_gp_wq;
|
|
|
|
|
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
static void __init kfree_rcu_batch_init(void)
|
|
|
|
|
{
|
|
|
|
|
int cpu;
|
2022-11-29 23:58:19 +08:00
|
|
|
|
int i, j;
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
|
2021-04-16 01:19:56 +08:00
|
|
|
|
/* Clamp it to [0:100] seconds interval. */
|
|
|
|
|
if (rcu_delay_page_cache_fill_msec < 0 ||
|
|
|
|
|
rcu_delay_page_cache_fill_msec > 100 * MSEC_PER_SEC) {
|
|
|
|
|
|
|
|
|
|
rcu_delay_page_cache_fill_msec =
|
|
|
|
|
clamp(rcu_delay_page_cache_fill_msec, 0,
|
|
|
|
|
(int) (100 * MSEC_PER_SEC));
|
|
|
|
|
|
|
|
|
|
pr_info("Adjusting rcutree.rcu_delay_page_cache_fill_msec to %d ms.\n",
|
|
|
|
|
rcu_delay_page_cache_fill_msec);
|
|
|
|
|
}
|
|
|
|
|
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
for_each_possible_cpu(cpu) {
|
|
|
|
|
struct kfree_rcu_cpu *krcp = per_cpu_ptr(&krc, cpu);
|
|
|
|
|
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
for (i = 0; i < KFREE_N_BATCHES; i++) {
|
|
|
|
|
INIT_RCU_WORK(&krcp->krw_arr[i].rcu_work, kfree_rcu_work);
|
2019-09-20 05:58:26 +08:00
|
|
|
|
krcp->krw_arr[i].krcp = krcp;
|
2022-11-29 23:58:19 +08:00
|
|
|
|
|
|
|
|
|
for (j = 0; j < FREE_N_CHANNELS; j++)
|
|
|
|
|
INIT_LIST_HEAD(&krcp->krw_arr[i].bulk_head_free[j]);
|
rcu: Support kfree_bulk() interface in kfree_rcu()
The kfree_rcu() logic can be improved further by using kfree_bulk()
interface along with "basic batching support" introduced earlier.
The are at least two advantages of using "bulk" interface:
- in case of large number of kfree_rcu() requests kfree_bulk()
reduces the per-object overhead caused by calling kfree()
per-object.
- reduces the number of cache-misses due to "pointer chasing"
between objects which can be far spread between each other.
This approach defines a new kfree_rcu_bulk_data structure that
stores pointers in an array with a specific size. Number of entries
in that array depends on PAGE_SIZE making kfree_rcu_bulk_data
structure to be exactly one page.
Since it deals with "block-chain" technique there is an extra
need in dynamic allocation when a new block is required. Memory
is allocated with GFP_NOWAIT | __GFP_NOWARN flags, i.e. that
allows to skip direct reclaim under low memory condition to
prevent stalling and fails silently under high memory pressure.
The "emergency path" gets maintained when a system is run out of
memory. In that case objects are linked into regular list.
The "rcuperf" was run to analyze this change in terms of memory
consumption and kfree_bulk() throughput.
1) Testing on the Intel(R) Xeon(R) W-2135 CPU @ 3.70GHz, 12xCPUs
with following parameters:
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1 kfree_vary_obj_size=1
dev.2020.01.10a branch
Default / CONFIG_SLAB
53607352517 ns, loops: 200000, batches: 1885, memory footprint: 1248MB
53529637912 ns, loops: 200000, batches: 1921, memory footprint: 1193MB
53570175705 ns, loops: 200000, batches: 1929, memory footprint: 1250MB
Patch / CONFIG_SLAB
23981587315 ns, loops: 200000, batches: 810, memory footprint: 1219MB
23879375281 ns, loops: 200000, batches: 822, memory footprint: 1190MB
24086841707 ns, loops: 200000, batches: 794, memory footprint: 1380MB
Default / CONFIG_SLUB
51291025022 ns, loops: 200000, batches: 1713, memory footprint: 741MB
51278911477 ns, loops: 200000, batches: 1671, memory footprint: 719MB
51256183045 ns, loops: 200000, batches: 1719, memory footprint: 647MB
Patch / CONFIG_SLUB
50709919132 ns, loops: 200000, batches: 1618, memory footprint: 456MB
50736297452 ns, loops: 200000, batches: 1633, memory footprint: 507MB
50660403893 ns, loops: 200000, batches: 1628, memory footprint: 429MB
in case of CONFIG_SLAB there is double increase in performance and
slightly higher memory usage. As for CONFIG_SLUB, the performance
figures are better together with lower memory usage.
2) Testing on the HiKey-960, arm64, 8xCPUs with below parameters:
CONFIG_SLAB=y
kfree_loops=200000 kfree_alloc_num=1000 kfree_rcu_test=1
102898760401 ns, loops: 200000, batches: 5822, memory footprint: 158MB
89947009882 ns, loops: 200000, batches: 6715, memory footprint: 115MB
rcuperf shows approximately ~12% better throughput in case of
using "bulk" interface. The "drain logic" or its RCU callback
does the work faster that leads to better throughput.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2020-01-20 22:42:25 +08:00
|
|
|
|
}
|
|
|
|
|
|
2022-11-29 23:58:19 +08:00
|
|
|
|
for (i = 0; i < FREE_N_CHANNELS; i++)
|
|
|
|
|
INIT_LIST_HEAD(&krcp->bulk_head[i]);
|
|
|
|
|
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
INIT_DELAYED_WORK(&krcp->monitor_work, kfree_rcu_monitor);
|
2021-04-16 01:19:56 +08:00
|
|
|
|
INIT_DELAYED_WORK(&krcp->page_cache_work, fill_page_cache_func);
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
krcp->initialized = true;
|
|
|
|
|
}
|
2022-06-01 11:22:24 +08:00
|
|
|
|
if (register_shrinker(&kfree_rcu_shrinker, "rcu-kfree"))
|
2020-03-17 00:32:27 +08:00
|
|
|
|
pr_err("Failed to register kfree_rcu() shrinker!\n");
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
}
|
|
|
|
|
|
2009-11-23 00:53:49 +08:00
|
|
|
|
void __init rcu_init(void)
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
{
|
2022-02-16 23:42:06 +08:00
|
|
|
|
int cpu = smp_processor_id();
|
2009-11-23 00:53:49 +08:00
|
|
|
|
|
2015-01-20 13:10:21 +08:00
|
|
|
|
rcu_early_boot_tests();
|
|
|
|
|
|
rcu: Add basic support for kfree_rcu() batching
Recently a discussion about stability and performance of a system
involving a high rate of kfree_rcu() calls surfaced on the list [1]
which led to another discussion how to prepare for this situation.
This patch adds basic batching support for kfree_rcu(). It is "basic"
because we do none of the slab management, dynamic allocation, code
moving or any of the other things, some of which previous attempts did
[2]. These fancier improvements can be follow-up patches and there are
different ideas being discussed in those regards. This is an effort to
start simple, and build up from there. In the future, an extension to
use kfree_bulk and possibly per-slab batching could be done to further
improve performance due to cache-locality and slab-specific bulk free
optimizations. By using an array of pointers, the worker thread
processing the work would need to read lesser data since it does not
need to deal with large rcu_head(s) any longer.
Torture tests follow in the next patch and show improvements of around
5x reduction in number of grace periods on a 16 CPU system. More
details and test data are in that patch.
There is an implication with rcu_barrier() with this patch. Since the
kfree_rcu() calls can be batched, and may not be handed yet to the RCU
machinery in fact, the monitor may not have even run yet to do the
queue_rcu_work(), there seems no easy way of implementing rcu_barrier()
to wait for those kfree_rcu()s that are already made. So this means a
kfree_rcu() followed by an rcu_barrier() does not imply that memory will
be freed once rcu_barrier() returns.
Another implication is higher active memory usage (although not
run-away..) until the kfree_rcu() flooding ends, in comparison to
without batching. More details about this are in the second patch which
adds an rcuperf test.
Finally, in the near future we will get rid of kfree_rcu() special casing
within RCU such as in rcu_do_batch and switch everything to just
batching. Currently we don't do that since timer subsystem is not yet up
and we cannot schedule the kfree_rcu() monitor as the timer subsystem's
lock are not initialized. That would also mean getting rid of
kfree_call_rcu_nobatch() entirely.
[1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org
[2] https://lkml.org/lkml/2017/12/19/824
Cc: kernel-team@android.com
Cc: kernel-team@lge.com
Co-developed-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ]
[ paulmck: Make it work during early boot. ]
[ paulmck: Add a crude early boot self-test. ]
[ paulmck: Style adjustments and experimental docbook structure header. ]
Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2019-08-06 06:22:27 +08:00
|
|
|
|
kfree_rcu_batch_init();
|
rcu: Merge preemptable-RCU functionality into hierarchical RCU
Create a kernel/rcutree_plugin.h file that contains definitions
for preemptable RCU (or, under the #else branch of the #ifdef,
empty definitions for the classic non-preemptable semantics).
These definitions fit into plugins defined in kernel/rcutree.c
for this purpose.
This variant of preemptable RCU uses a new algorithm whose
read-side expense is roughly that of classic hierarchical RCU
under CONFIG_PREEMPT. This new algorithm's update-side expense
is similar to that of classic hierarchical RCU, and, in absence
of read-side preemption or blocking, is exactly that of classic
hierarchical RCU. Perhaps more important, this new algorithm
has a much simpler implementation, saving well over 1,000 lines
of code compared to mainline's implementation of preemptable
RCU, which will hopefully be retired in favor of this new
algorithm.
The simplifications are obtained by maintaining per-task
nesting state for running tasks, and using a simple
lock-protected algorithm to handle accounting when tasks block
within RCU read-side critical sections, making use of lessons
learned while creating numerous user-level RCU implementations
over the past 18 months.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: laijs@cn.fujitsu.com
Cc: dipankar@in.ibm.com
Cc: akpm@linux-foundation.org
Cc: mathieu.desnoyers@polymtl.ca
Cc: josht@linux.vnet.ibm.com
Cc: dvhltc@us.ibm.com
Cc: niv@us.ibm.com
Cc: peterz@infradead.org
Cc: rostedt@goodmis.org
LKML-Reference: <12509746134003-git-send-email->
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-08-23 04:56:52 +08:00
|
|
|
|
rcu_bootup_announce();
|
2022-01-12 07:32:50 +08:00
|
|
|
|
sanitize_kthread_prio();
|
2012-04-24 06:52:53 +08:00
|
|
|
|
rcu_init_geometry();
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_init_one();
|
2015-04-21 02:40:50 +08:00
|
|
|
|
if (dump_tree)
|
2018-07-04 08:22:34 +08:00
|
|
|
|
rcu_dump_rcu_node_tree();
|
2019-03-21 05:13:33 +08:00
|
|
|
|
if (use_softirq)
|
|
|
|
|
open_softirq(RCU_SOFTIRQ, rcu_core_si);
|
2009-11-23 00:53:49 +08:00
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* We don't need protection against CPU-hotplug here because
|
|
|
|
|
* this is called early in boot, before either interrupts
|
|
|
|
|
* or the scheduler are operational.
|
|
|
|
|
*/
|
2013-04-22 06:12:42 +08:00
|
|
|
|
pm_notifier(rcu_pm_notify, 0);
|
2022-02-16 23:42:06 +08:00
|
|
|
|
WARN_ON(num_online_cpus() > 1); // Only one CPU this early in boot.
|
|
|
|
|
rcutree_prepare_cpu(cpu);
|
|
|
|
|
rcu_cpu_starting(cpu);
|
|
|
|
|
rcutree_online_cpu(cpu);
|
2018-01-09 06:35:52 +08:00
|
|
|
|
|
2021-03-31 04:47:42 +08:00
|
|
|
|
/* Create workqueue for Tree SRCU and for expedited GPs. */
|
2018-01-09 06:35:52 +08:00
|
|
|
|
rcu_gp_wq = alloc_workqueue("rcu_gp", WQ_MEM_RECLAIM, 0);
|
|
|
|
|
WARN_ON(!rcu_gp_wq);
|
rcu: Move expedited grace period (GP) work to RT kthread_worker
Enabling CONFIG_RCU_BOOST did not reduce RCU expedited grace-period
latency because its workqueues run at SCHED_OTHER, and thus can be
delayed by normal processes. This commit avoids these delays by moving
the expedited GP work items to a real-time-priority kthread_worker.
This option is controlled by CONFIG_RCU_EXP_KTHREAD and disabled by
default on PREEMPT_RT=y kernels which disable expedited grace periods
after boot by unconditionally setting rcupdate.rcu_normal_after_boot=1.
The results were evaluated on arm64 Android devices (6GB ram) running
5.10 kernel, and capturing trace data in critical user-level code.
The table below shows the resulting order-of-magnitude improvements
in synchronize_rcu_expedited() latency:
------------------------------------------------------------------------
| | workqueues | kthread_worker | Diff |
------------------------------------------------------------------------
| Count | 725 | 688 | |
------------------------------------------------------------------------
| Min Duration (ns) | 326 | 447 | 37.12% |
------------------------------------------------------------------------
| Q1 (ns) | 39,428 | 38,971 | -1.16% |
------------------------------------------------------------------------
| Q2 - Median (ns) | 98,225 | 69,743 | -29.00% |
------------------------------------------------------------------------
| Q3 (ns) | 342,122 | 126,638 | -62.98% |
------------------------------------------------------------------------
| Max Duration (ns) | 372,766,967 | 2,329,671 | -99.38% |
------------------------------------------------------------------------
| Avg Duration (ns) | 2,746,353 | 151,242 | -94.49% |
------------------------------------------------------------------------
| Standard Deviation (ns) | 19,327,765 | 294,408 | |
------------------------------------------------------------------------
The below table show the range of maximums/minimums for
synchronize_rcu_expedited() latency from all experiments:
------------------------------------------------------------------------
| | workqueues | kthread_worker | Diff |
------------------------------------------------------------------------
| Total No. of Experiments | 25 | 23 | |
------------------------------------------------------------------------
| Largest Maximum (ns) | 372,766,967 | 2,329,671 | -99.38% |
------------------------------------------------------------------------
| Smallest Maximum (ns) | 38,819 | 86,954 | 124.00% |
------------------------------------------------------------------------
| Range of Maximums (ns) | 372,728,148 | 2,242,717 | |
------------------------------------------------------------------------
| Largest Minimum (ns) | 88,623 | 27,588 | -68.87% |
------------------------------------------------------------------------
| Smallest Minimum (ns) | 326 | 447 | 37.12% |
------------------------------------------------------------------------
| Range of Minimums (ns) | 88,297 | 27,141 | |
------------------------------------------------------------------------
Cc: "Paul E. McKenney" <paulmck@kernel.org>
Cc: Tejun Heo <tj@kernel.org>
Reported-by: Tim Murray <timmurray@google.com>
Reported-by: Wei Wang <wvw@google.com>
Tested-by: Kyle Lin <kylelin@google.com>
Tested-by: Chunwei Lu <chunweilu@google.com>
Tested-by: Lulu Wang <luluw@google.com>
Signed-off-by: Kalesh Singh <kaleshsingh@google.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2022-04-09 08:35:27 +08:00
|
|
|
|
rcu_alloc_par_gp_wq();
|
2019-10-31 02:56:10 +08:00
|
|
|
|
|
|
|
|
|
/* Fill in default value for rcutree.qovld boot parameter. */
|
|
|
|
|
/* -After- the rcu_node ->lock fields are initialized! */
|
|
|
|
|
if (qovld < 0)
|
|
|
|
|
qovld_calc = DEFAULT_RCU_QOVLD_MULT * qhimark;
|
|
|
|
|
else
|
|
|
|
|
qovld_calc = qovld;
|
2022-04-16 01:55:42 +08:00
|
|
|
|
|
2023-01-13 02:48:29 +08:00
|
|
|
|
// Kick-start in case any polled grace periods started early.
|
|
|
|
|
(void)start_poll_synchronize_rcu_expedited();
|
2022-12-20 09:02:20 +08:00
|
|
|
|
|
|
|
|
|
rcu_test_sync_prims();
|
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-12-19 04:55:32 +08:00
|
|
|
|
}
|
|
|
|
|
|
2019-01-12 08:10:57 +08:00
|
|
|
|
#include "tree_stall.h"
|
2016-04-16 07:35:29 +08:00
|
|
|
|
#include "tree_exp.h"
|
2021-05-19 08:09:28 +08:00
|
|
|
|
#include "tree_nocb.h"
|
2013-10-09 11:23:47 +08:00
|
|
|
|
#include "tree_plugin.h"
|