With the removal of the ENODATA case from padata_get_next, the cpu_index
field is no longer useful, so it can go away.
Signed-off-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Acked-by: Steffen Klassert <steffen.klassert@secunet.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Tejun Heo <tj@kernel.org>
Cc: linux-crypto@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Padata binds the parallel part of a job to a single CPU and round-robins
over all CPUs in the system for each successive job. Though the serial
parts rely on per-CPU queues for correct ordering, they're not necessary
for parallel work, and it improves performance to run the job locally on
NUMA machines and let the scheduler pick the CPU within a node on a busy
system.
So, make the parallel workqueue unbound.
Update the parallel workqueue's cpumask when the instance's parallel
cpumask changes.
Now that parallel jobs no longer run on max_active=1 workqueues, two or
more parallel works that hash to the same CPU may run simultaneously,
finish out of order, and so be serialized out of order. Prevent this by
keeping the works sorted on the reorder list by sequence number and
checking that in the reordering logic.
padata_get_next becomes padata_find_next so it can be reused for the end
of padata_reorder, where it's used to avoid uselessly queueing work when
the next job by sequence number isn't finished yet but a later job that
hashed to the same CPU has.
The ENODATA case in padata_find_next no longer makes sense because
parallel jobs aren't bound to specific CPUs. The EINPROGRESS case takes
care of the scenario where a parallel job is potentially running on the
same CPU as padata_find_next, and with only one error code left, just
use NULL instead.
Signed-off-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steffen Klassert <steffen.klassert@secunet.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: linux-crypto@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
padata currently uses one per-CPU workqueue per instance for all work.
Prepare for running parallel jobs on an unbound workqueue by introducing
dedicated workqueues for parallel and serial work.
Signed-off-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Acked-by: Steffen Klassert <steffen.klassert@secunet.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Tejun Heo <tj@kernel.org>
Cc: linux-crypto@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
padata_do_parallel currently returns -EINVAL if the callback CPU isn't
in the callback cpumask.
pcrypt tries to prevent this situation by keeping its own callback
cpumask in sync with padata's and checks that the callback CPU it passes
to padata is valid. Make padata handle this instead.
padata_do_parallel now takes a pointer to the callback CPU and updates
it for the caller if an alternate CPU is used. Overall behavior in
terms of which callback CPUs are chosen stays the same.
Prepares for removal of the padata cpumask notifier in pcrypt, which
will fix a lockdep complaint about nested acquisition of the CPU hotplug
lock later in the series.
Signed-off-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Acked-by: Steffen Klassert <steffen.klassert@secunet.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Tejun Heo <tj@kernel.org>
Cc: linux-crypto@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Move workqueue allocation inside of padata to prepare for further
changes to how padata uses workqueues.
Guarantees the workqueue is created with max_active=1, which padata
relies on to work correctly. No functional change.
Signed-off-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Acked-by: Steffen Klassert <steffen.klassert@secunet.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Tejun Heo <tj@kernel.org>
Cc: linux-crypto@vger.kernel.org
Cc: linux-doc@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The function padata_reorder will use a timer when it cannot progress
while completed jobs are outstanding (pd->reorder_objects > 0). This
is suboptimal as if we do end up using the timer then it would have
introduced a gratuitous delay of one second.
In fact we can easily distinguish between whether completed jobs
are outstanding and whether we can make progress. All we have to
do is look at the next pqueue list.
This patch does that by replacing pd->processed with pd->cpu so
that the next pqueue is more accessible.
A work queue is used instead of the original try_again to avoid
hogging the CPU.
Note that we don't bother removing the work queue in
padata_flush_queues because the whole premise is broken. You
cannot flush async crypto requests so it makes no sense to even
try. A subsequent patch will fix it by replacing it with a ref
counting scheme.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Based on 1 normalized pattern(s):
this program is free software you can redistribute it and or modify
it under the terms and conditions of the gnu general public license
version 2 as published by the free software foundation this program
is distributed in the hope it will be useful but without any
warranty without even the implied warranty of merchantability or
fitness for a particular purpose see the gnu general public license
for more details you should have received a copy of the gnu general
public license along with this program if not write to the free
software foundation inc 51 franklin st fifth floor boston ma 02110
1301 usa
extracted by the scancode license scanner the SPDX license identifier
GPL-2.0-only
has been chosen to replace the boilerplate/reference in 111 file(s).
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Alexios Zavras <alexios.zavras@intel.com>
Reviewed-by: Allison Randal <allison@lohutok.net>
Cc: linux-spdx@vger.kernel.org
Link: https://lkml.kernel.org/r/20190530000436.567572064@linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If the algorithm we're parallelizing is asynchronous we might change
CPUs between padata_do_parallel() and padata_do_serial(). However, we
don't expect this to happen as we need to enqueue the padata object into
the per-cpu reorder queue we took it from, i.e. the same-cpu's parallel
queue.
Ensure we're not switching CPUs for a given padata object by tracking
the CPU within the padata object. If the serial callback gets called on
the wrong CPU, defer invoking padata_reorder() via a kernel worker on
the CPU we're expected to run on.
Signed-off-by: Mathias Krause <minipli@googlemail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The reorder timer function runs on the CPU where the timer interrupt was
handled which is not necessarily one of the CPUs of the 'pcpu' CPU mask
set.
Ensure the padata_reorder() callback runs on the correct CPU, which is
one in the 'pcpu' CPU mask set and, preferrably, the next expected one.
Do so by comparing the current CPU with the expected target CPU. If they
match, call padata_reorder() right away. If they differ, schedule a work
item on the target CPU that does the padata_reorder() call for us.
Signed-off-by: Mathias Krause <minipli@googlemail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
No users outside of padata.c
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Acked-by: Ingo Molnar <mingo@kernel.org>
Cc: Steffen Klassert <steffen.klassert@secunet.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sebastian Siewior <bigeasy@linutronix.de>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: linux-crypto@vger.kernel.org
Link: http://lkml.kernel.org/r/20170524081547.491457256@linutronix.de
Install the callbacks via the state machine. CPU-hotplug multinstance support
is used with the nocalls() version. Maybe parts of padata_alloc() could be
moved into the online callback so that we could invoke ->startup callback for
instance and drop get_online_cpus().
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Steffen Klassert <steffen.klassert@secunet.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: linux-crypto@vger.kernel.org
Cc: rt@linutronix.de
Link: http://lkml.kernel.org/r/20160906170457.32393-14-bigeasy@linutronix.de
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
By accident I stumbled across code that has never been used. This
driver has EXPORT_SYMBOL functions, and the only user of the code is
pcrypt.c, but this only uses a subset of the exported symbols.
According to 'git log -G', the functions, padata_set_cpumasks,
padata_add_cpu, and padata_remove_cpu have never been used since they
were first introduced. This patch removes the unused code.
On one 64 bit build, with CRYPTO_PCRYPT built in, the text is more than
4k smaller.
kbuild_hp> size $KBUILD_OUTPUT/vmlinux
text data bss dec hex filename
10566658 4678360 1122304 16367322 f9beda vmlinux
10561984 4678360 1122304 16362648 f9ac98 vmlinux
On another config, 32 bit, the saving is about 0.5k bytes.
kbuild_hp-x86> size $KBUILD_OUTPUT/vmlinux
6012005 2409513 2785280 11206798 ab008e vmlinux
6011491 2409513 2785280 11206284 aafe8c vmlinux
Signed-off-by: Richard Cochran <rcochran@linutronix.de>
Cc: Steffen Klassert <steffen.klassert@secunet.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Using a spinlock to atomically increase a counter sounds wrong -- we've
atomic_t for this!
Also move 'seq_nr' to a different cache line than 'lock' to reduce cache
line trashing. This has the nice side effect of decreasing the size of
struct parallel_data from 192 to 128 bytes for a x86-64 build, e.g.
occupying only two instead of three cache lines.
Those changes results in a 5% performance increase on an IPsec test run
using pcrypt.
Btw. the seq_lock spinlock was never explicitly initialized -- one more
reason to get rid of it.
Signed-off-by: Mathias Krause <mathias.krause@secunet.com>
Acked-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
When padata_do_parallel() is called from multiple cpus for the same
padata instance, we can get object reordering on sequence number wrap
because testing for sequence number wrap and reseting the sequence
number must happen atomically but is implemented with two atomic
operations. This patch fixes this by converting the sequence number
from atomic_t to an unsigned int and protect the access with a
spin_lock. As a side effect, we get rid of the sequence number wrap
handling because the seqence number wraps back to null now without
the need to do anything.
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
parallel_data->queue is a percpu pointer but was missing __percpu markup.
Add it.
Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Acked-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
A function that copies the padata cpumasks to a user buffer
is a bit error prone. The cpumask can change any time so we
can't be sure to have the right cpumask when using this function.
A user who is interested in the padata cpumasks should register
to the padata cpumask notifier chain instead. Users of
padata_get_cpumask are already updated, so we can remove it.
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
We pass a pointer to the new padata cpumasks to the cpumask_change_notifier
chain. So users can access the cpumasks without the need of an extra
padata_get_cpumask function.
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
padata_set_cpumask needs to be protected by a lock. We make
__padata_set_cpumasks unlocked and static. So this function
can be used by the exported and locked padata_set_cpumask and
padata_set_cpumasks functions.
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
We rename padata_alloc to padata_alloc_possible because this
function allocates a padata_instance and uses the cpu_possible
mask for parallel and serial workers. Also we rename __padata_alloc
to padata_alloc to avoid to export underlined functions. Underlined
functions are considered to be private to padata. Users are updated
accordingly.
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Added sysfs primitives to padata subsystem. Now API user may
embedded kobject each padata instance contains into any sysfs
hierarchy. For now padata sysfs interface provides only
two objects:
serial_cpumask [RW] - cpumask for serial workers
parallel_cpumask [RW] - cpumask for parallel workers
Signed-off-by: Dan Kruchinin <dkruchinin@acm.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The aim of this patch is to make two separate cpumasks
for padata parallel and serial workers respectively.
It allows user to make more thin and sophisticated configurations
of padata framework. For example user may bind parallel and serial workers to non-intersecting
CPU groups to gain better performance. Also each padata instance has notifiers chain for its
cpumasks now. If either parallel or serial or both masks were changed all
interested subsystems will get notification about that. It's especially useful
if padata user uses algorithm for callback CPU selection according to serial cpumask.
Signed-off-by: Dan Kruchinin <dkruchinin@acm.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
We count the number of processed objects on a percpu basis,
so we need to go through all the percpu reorder queues to calculate
the sequence number of the next object that needs serialization.
This patch changes this to count the number of processed objects
global. So we can calculate the sequence number and the percpu
reorder queue of the next object that needs serialization without
searching through the percpu reorder queues. This avoids some
accesses to memory of foreign cpus.
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This patch introduces the PADATA_INVALID flag which is
checked on padata start. This will be used to mark a padata
instance as invalid, if the padata cpumask does not intersect
with the active cpumask. we change padata_start to return an
error if the PADATA_INVALID is set. Also we adapt the only
padata user, pcrypt to this change.
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
padata_get_next needs to check whether the next object that
need serialization must be parallel processed by the local cpu.
This check was wrong implemented and returned always true,
so the try_again loop in padata_reorder was never taken. This
can lead to object leaks in some rare cases due to a race that
appears with the trylock in padata_reorder. The try_again loop
was not a good idea after all, because a cpu could take that
loop frequently, so we handle this with a timer instead.
This patch adds a timer to handle the race that appears with
the trylock. If cpu1 queues an object to the reorder queue while
cpu2 holds the pd->lock but left the while loop in padata_reorder
already, cpu2 can't care for this object and cpu1 exits because
it can't get the lock. Usually the next cpu that takes the lock
cares for this object too. We need the timer just if this object
was the last one that arrives to the reorder queues. The timer
function sends it out in this case.
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This patch introduces an interface to process data objects
in parallel. The parallelized objects return after serialization
in the same order as they were before the parallelization.
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Daniel Jordan