All callers should still have RCU enabled.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lore.kernel.org/r/20230112195540.190860672@infradead.org
The whole disable-RCU, enable-IRQS dance is very intricate since
changing IRQ state is traced, which depends on RCU.
Add two helpers for the cpuidle case that mirror the entry code:
ct_cpuidle_enter()
ct_cpuidle_exit()
And fix all the cases where the enter/exit dance was buggy.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lore.kernel.org/r/20230112195540.130014793@infradead.org
Doing RCU-idle outside the driver, only to then temporarily enable it
again before going idle is suboptimal.
Notably: this converts all dt_init_idle_driver() and
__CPU_PM_CPU_IDLE_ENTER() users for they are inextrably intertwined.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lore.kernel.org/r/20230112195540.068981667@infradead.org
Doing RCU-idle outside the driver, only to then temporarily enable it
again, some *four* times, before going idle is suboptimal.
Notably three times explicitly using RCU_NONIDLE() and once implicitly
through cpu_pm_*().
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Reviewed-by: Tony Lindgren <tony@atomide.com>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/20230112195540.007918454@infradead.org
Doing RCU-idle outside the driver, only to then temporarily enable it
again before going idle is suboptimal.
Notably the cpu_pm_*() calls implicitly re-enable RCU for a bit.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/20230112195539.946630819@infradead.org
Doing RCU-idle outside the driver, only to then teporarily enable it
again before going idle is suboptimal.
Notably the cpu_pm_*() calls implicitly re-enable RCU for a bit.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Reviewed-by: Tony Lindgren <tony@atomide.com>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/20230112195539.883561913@infradead.org
Doing RCU-idle outside the driver, only to then temporarily enable it
again, at least twice, before going idle is suboptimal.
Notably both cpu_pm_enter() and cpu_cluster_pm_enter() implicity
re-enable RCU.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/20230112195539.821714572@infradead.org
Doing RCU-idle outside the driver, only to then temporarily enable it
again, at least twice, before going idle is suboptimal.
Notably once implicitly through the cpu_pm_*() calls and once
explicitly doing ct_irq_*_irqon().
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Kajetan Puchalski <kajetan.puchalski@arm.com>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Reviewed-by: Guo Ren <guoren@kernel.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/20230112195539.760296658@infradead.org
Doing RCU-idle outside the driver, only to then temporarily enable it
again, at least twice, before going idle is suboptimal.
Notably once implicitly through the cpu_pm_*() calls and once
explicitly doing RCU_NONIDLE().
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/20230112195539.699546331@infradead.org
Doing RCU-idle outside the driver, only to then temporarily enable it
again, at least twice, before going idle is suboptimal.
That is, once implicitly through the cpu_pm_*() calls and once
explicitly doing ct_irq_*_irqon().
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Anup Patel <anup@brainfault.org>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/20230112195539.637185846@infradead.org
Make cpuidle_enter_state() consistent with the s2idle variant and
verify ->enter() always returns with interrupts disabled.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lore.kernel.org/r/20230112195539.576412812@infradead.org
Make cpuidle_state::enter() methods IRQ state invariant on exit.
Additionally make sure to use raw_local_irq_*() methods since this
cpuidle callback will be called with RCU already disabled.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lore.kernel.org/r/20230112195539.515253662@infradead.org
Typical boot time setup; no need to suffer an indirect call for that.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Reviewed-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/20230112195539.453613251@infradead.org
The perf_lopwr_cb() function is called from the idle routines; there
is no RCU there, we must not enter tracing.
Use __always_inline, noidle annotations and existing no-trace methods.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lore.kernel.org/r/20230112195539.392862891@infradead.org
Two new auxiliary vector entries are introduced for rseq without
matching increment of the AT_VECTOR_SIZE_BASE, which causes failures
with CONFIG_HARDENED_USERCOPY=y.
Fixes: 317c8194e6 ("rseq: Introduce feature size and alignment ELF auxiliary vector entries")
Reported-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Nathan Chancellor <nathan@kernel.org>
Link: https://lore.kernel.org/r/20230104192054.34046-1-mathieu.desnoyers@efficios.com
The mm_numa_cid related rseq patches from the series were not picked up
into the tip tree, so enabling the mm_numa_cid test needs to be
reverted.
This reverts commit b344b8f2d8.
Reported-by: kernel test robot <oliver.sang@intel.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/oe-lkp/202301040903.2dd1e25b-oliver.sang@intel.com
The following commit:
c89970202a ("cputime: remove cputime_to_nsecs fallback")
Removed an <asm/cputime.h> inclusion from <linux/sched/cputime.h>, but this
broke the IA64 build:
arch/ia64/kernel/time.c:110:6: warning: no previous prototype for 'arch_vtime_task_switch' [-Wmissing-prototypes]
Add in the missing <asm/cputime.h> header to fix it.
Fixes: c89970202a ("cputime: remove cputime_to_nsecs fallback")
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: linux-kernel@vger.kernel.org
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Keep track of previously issued registrations and compare the result
with MEMBARRIER_CMD_GET_REGISTRATIONS return value.
Signed-off-by: Michal Clapinski <mclapinski@google.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Acked-by: Paul E. McKenney <paulmck@kernel.org>
Link: https://lore.kernel.org/r/20221207164338.1535591-3-mclapinski@google.com
Provide a method to query previously issued registrations.
Signed-off-by: Michal Clapinski <mclapinski@google.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Acked-by: Paul E. McKenney <paulmck@kernel.org>
Link: https://lore.kernel.org/r/20221207164338.1535591-2-mclapinski@google.com
The max CPU capacity is the same for all CPUs sharing frequency domain.
There is a way to avoid heavy operations in a loop for each CPU by
leveraging this knowledge. Thus, simplify the looping code in the
sugov_next_freq_shared() and drop heavy multiplications. Instead, use
simple max() to get the highest utilization from these CPUs.
This is useful for platforms with many (4 or 6) little CPUs. We avoid
heavy 2*PD_CPU_NUM multiplications in that loop, which is called billions
of times, since it's not limited by the schedutil time delta filter in
sugov_should_update_freq(). When there was no need to change frequency
the code bailed out, not updating the sg_policy::last_freq_update_time.
Then every visit after delta_ns time longer than the
sg_policy::freq_update_delay_ns goes through and triggers the next
frequency calculation code. Although, if the next frequency, as outcome
of that, would be the same as current frequency, we won't update the
sg_policy::last_freq_update_time and the story will be repeated (in
a very short period, sometimes a few microseconds).
The max CPU capacity must be fetched every time we are called, due to
difficulties during the policy setup, where we are not able to get the
normalized CPU capacity at the right time.
The fetched CPU capacity value is than used in sugov_iowait_apply() to
calculate the right boost. This required a few changes in the local
functions and arguments. The capacity value should hopefully be fetched
once when needed and then passed over CPU registers to those functions.
Signed-off-by: Lukasz Luba <lukasz.luba@arm.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20221208160256.859-2-lukasz.luba@arm.com
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Patrick Bellasi <patrick.bellasi@arm.com>
Cc: Vincent Guittot <vincent.guittot@linaro.org>
Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
ttwu_do_activate() is used for a complete wakeup, in which we will
activate_task() and use ttwu_do_wakeup() to mark the task runnable
and perform wakeup-preemption, also call class->task_woken() callback
and update the rq->idle_stamp.
Since ttwu_runnable() is not a complete wakeup, don't need all those
done in ttwu_do_wakeup(), so we can move those to ttwu_do_activate()
to simplify ttwu_do_wakeup(), making it only mark the task runnable
to be reused in ttwu_runnable() and try_to_wake_up().
This patch should not have any functional changes.
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Chengming Zhou <zhouchengming@bytedance.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20221223103257.4962-2-zhouchengming@bytedance.com
ttwu_runnable() is used as a fast wakeup path when the wakee task
is running on CPU or runnable on RQ, in both cases we can just
set its state to TASK_RUNNING to prevent a sleep.
If the wakee task is on_cpu running, we don't need to update_rq_clock()
or check_preempt_curr().
But if the wakee task is on_rq && !on_cpu (e.g. an IRQ hit before
the task got to schedule() and the task been preempted), we should
check_preempt_curr() to see if it can preempt the current running.
This also removes the class->task_woken() callback from ttwu_runnable(),
which wasn't required per the RT/DL implementations: any required push
operation would have been queued during class->set_next_task() when p
got preempted.
ttwu_runnable() also loses the update to rq->idle_stamp, as by definition
the rq cannot be idle in this scenario.
Suggested-by: Valentin Schneider <vschneid@redhat.com>
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Chengming Zhou <zhouchengming@bytedance.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Reviewed-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Link: https://lore.kernel.org/r/20221223103257.4962-1-zhouchengming@bytedance.com
Add a document explaining the util clamp feature: what it is and
how to use it. The new document hopefully covers everything one needs to
know about uclamp.
Signed-off-by: Qais Yousef <qais.yousef@arm.com>
Signed-off-by: Qais Yousef (Google) <qyousef@layalina.io>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Lukasz Luba <lukasz.luba@arm.com>
Link: https://lore.kernel.org/r/20221216235716.201923-1-qyousef@layalina.io
Cc: Jonathan Corbet <corbet@lwn.net>
sched_mm_cid_after_execve() does not expect NULL t->mm, but it may happen
if a usermodehelper kthread fails when attempting to execute a binary.
sched_mm_cid_fork() can be issued from a usermodehelper kthread, which
has t->flags PF_KTHREAD set.
Fixes: af7f588d8f ("sched: Introduce per-memory-map concurrency ID")
Reported-by: kernel test robot <yujie.liu@intel.com>
Reported-by: Borislav Petkov <bp@alien8.de>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/oe-lkp/202212301353.5c959d72-yujie.liu@intel.com
The archs that use cputime_to_nsecs() internally provide their own
definition and don't need the fallback. cputime_to_usecs() unused except
in this fallback, and is not defined anywhere.
This removes the final remnant of the cputime_t code from the kernel.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Alexander Gordeev <agordeev@linux.ibm.com>
Link: https://lore.kernel.org/r/20221220070705.2958959-1-npiggin@gmail.com
When select_idle_capacity() starts scanning for an idle CPU, it starts
with target CPU that has already been checked in select_idle_sibling().
So we start checking from the next CPU and try the target CPU at the end.
Similarly for task_numa_assign(), we have just checked numa_migrate_on
of dst_cpu, so start from the next CPU. This also works for
steal_cookie_task(), the first scan must fail and start directly
from the next one.
Signed-off-by: Hao Jia <jiahao.os@bytedance.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Link: https://lore.kernel.org/r/20221216062406.7812-3-jiahao.os@bytedance.com
In update_numa_stats() we try to find an idle cpu on the NUMA node,
preferably an idle core. we can stop looking for the next idle core
or idle cpu after finding an idle core. But we can't stop the
whole loop of scanning the CPU, because we need to calculate
approximate NUMA stats at a point in time. For example,
the src and dst nr_running is needed by task_numa_find_cpu().
Signed-off-by: Hao Jia <jiahao.os@bytedance.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Link: https://lore.kernel.org/r/20221216062406.7812-2-jiahao.os@bytedance.com
With a modified container_of() that preserves constness, the compiler
finds some pointers which should have been marked as const. task_of()
also needs to become const-preserving for the !FAIR_GROUP_SCHED case so
that cfs_rq_of() can take a const argument. No change to generated code.
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20221212144946.2657785-1-willy@infradead.org
Add the mm_cid field to the rseq_update event, allowing tracers to
follow which mm_cid is observed by user-space, and whether negative
mm_cid values are visible in case of internal scheduler implementation
issues.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20221122203932.231377-22-mathieu.desnoyers@efficios.com
Adapt to the rseq.h API changes introduced by commits
"selftests/rseq: <arch>: Template memory ordering and percpu access mode".
Build a new param_test_mm_cid, param_test_mm_cid_benchmark, and
param_test_mm_cid_compare_twice executables to test the new "mm_cid"
rseq field.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20221122203932.231377-20-mathieu.desnoyers@efficios.com
Adapt to the rseq.h API changes introduced by commits
"selftests/rseq: <arch>: Template memory ordering and percpu access mode".
Build a new basic_percpu_ops_mm_cid_test to test the new "mm_cid" rseq
field.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20221122203932.231377-19-mathieu.desnoyers@efficios.com
Introduce a rseq-riscv-bits.h template header which is internally included
to generate the static inline functions covering:
- relaxed and release memory ordering,
- per-cpu-id and per-mm-cid per-cpu data access.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20221122203932.231377-18-mathieu.desnoyers@efficios.com
Introduce a rseq-s390-bits.h template header which is internally included
to generate the static inline functions covering:
- relaxed and release memory ordering,
- per-cpu-id and per-mm-cid per-cpu data access.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20221122203932.231377-17-mathieu.desnoyers@efficios.com
Introduce a rseq-ppc-bits.h template header which is internally included
to generate the static inline functions covering:
- relaxed and release memory ordering,
- per-cpu-id and per-mm-cid per-cpu data access.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20221122203932.231377-16-mathieu.desnoyers@efficios.com
Introduce a rseq-mips-bits.h template header which is internally
included to generate the static inline functions covering:
- relaxed and release memory ordering,
- per-cpu-id and per-mm-cid per-cpu data access.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20221122203932.231377-15-mathieu.desnoyers@efficios.com
Introduce a rseq-arm64-bits.h template header which is internally
included to generate the static inline functions covering:
- relaxed and release memory ordering,
- per-cpu-id and per-mm-cid per-cpu data access.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20221122203932.231377-14-mathieu.desnoyers@efficios.com
Introduce a rseq-arm-bits.h template header which is internally included
to generate the static inline functions covering:
- relaxed and release memory ordering,
- per-cpu-id and per-mm-cid per-cpu data access.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20221122203932.231377-13-mathieu.desnoyers@efficios.com
Introduce a rseq-x86-bits.h template header which is internally included
to generate the static inline functions covering:
- relaxed and release memory ordering,
- per-cpu-id and per-mm-cid per-cpu data access.
This introduces changes to the rseq.h selftests API which require to
update the rseq selftest programs. Similar API/templating changes need
to be done for other architectures.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20221122203932.231377-12-mathieu.desnoyers@efficios.com
This code is not currently build by the test Makefile, adds complexity,
and is not overall useful considering that the abort handling loops to
retry the fast-path.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20221122203932.231377-10-mathieu.desnoyers@efficios.com
If a memory map has fewer threads than there are cores on the system, or
is limited to run on few cores concurrently through sched affinity or
cgroup cpusets, the concurrency IDs will be values close to 0, thus
allowing efficient use of user-space memory for per-cpu data structures.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20221122203932.231377-9-mathieu.desnoyers@efficios.com
This feature allows the scheduler to expose a per-memory map concurrency
ID to user-space. This concurrency ID is within the possible cpus range,
and is temporarily (and uniquely) assigned while threads are actively
running within a memory map. If a memory map has fewer threads than
cores, or is limited to run on few cores concurrently through sched
affinity or cgroup cpusets, the concurrency IDs will be values close
to 0, thus allowing efficient use of user-space memory for per-cpu
data structures.
This feature is meant to be exposed by a new rseq thread area field.
The primary purpose of this feature is to do the heavy-lifting needed
by memory allocators to allow them to use per-cpu data structures
efficiently in the following situations:
- Single-threaded applications,
- Multi-threaded applications on large systems (many cores) with limited
cpu affinity mask,
- Multi-threaded applications on large systems (many cores) with
restricted cgroup cpuset per container.
One of the key concern from scheduler maintainers is the overhead
associated with additional spin locks or atomic operations in the
scheduler fast-path. This is why the following optimization is
implemented.
On context switch between threads belonging to the same memory map,
transfer the mm_cid from prev to next without any atomic ops. This
takes care of use-cases involving frequent context switch between
threads belonging to the same memory map.
Additional optimizations can be done if the spin locks added when
context switching between threads belonging to different memory maps end
up being a performance bottleneck. Those are left out of this patch
though. A performance impact would have to be clearly demonstrated to
justify the added complexity.
The credit goes to Paul Turner (Google) for the original virtual cpu id
idea. This feature is implemented based on the discussions with Paul
Turner and Peter Oskolkov (Google), but I took the liberty to implement
scheduler fast-path optimizations and my own NUMA-awareness scheme. The
rumor has it that Google have been running a rseq vcpu_id extension
internally in production for a year. The tcmalloc source code indeed has
comments hinting at a vcpu_id prototype extension to the rseq system
call [1].
The following benchmarks do not show any significant overhead added to
the scheduler context switch by this feature:
* perf bench sched messaging (process)
Baseline: 86.5±0.3 ms
With mm_cid: 86.7±2.6 ms
* perf bench sched messaging (threaded)
Baseline: 84.3±3.0 ms
With mm_cid: 84.7±2.6 ms
* hackbench (process)
Baseline: 82.9±2.7 ms
With mm_cid: 82.9±2.9 ms
* hackbench (threaded)
Baseline: 85.2±2.6 ms
With mm_cid: 84.4±2.9 ms
[1] https://github.com/google/tcmalloc/blob/master/tcmalloc/internal/linux_syscall_support.h#L26
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20221122203932.231377-8-mathieu.desnoyers@efficios.com
Test the NUMA node id extension rseq field. Compare it against the value
returned by the getcpu(2) system call while pinned on a specific core.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20221122203932.231377-7-mathieu.desnoyers@efficios.com
Adding the NUMA node id to struct rseq is a straightforward thing to do,
and a good way to figure out if anything in the user-space ecosystem
prevents extending struct rseq.
This NUMA node id field allows memory allocators such as tcmalloc to
take advantage of fast access to the current NUMA node id to perform
NUMA-aware memory allocation.
It can also be useful for implementing fast-paths for NUMA-aware
user-space mutexes.
It also allows implementing getcpu(2) purely in user-space.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20221122203932.231377-5-mathieu.desnoyers@efficios.com
Introduce the extensible rseq ABI, where the feature size supported by
the kernel and the required alignment are communicated to user-space
through ELF auxiliary vectors.
This allows user-space to call rseq registration with a rseq_len of
either 32 bytes for the original struct rseq size (which includes
padding), or larger.
If rseq_len is larger than 32 bytes, then it must be large enough to
contain the feature size communicated to user-space through ELF
auxiliary vectors.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20221122203932.231377-4-mathieu.desnoyers@efficios.com
Export the rseq feature size supported by the kernel as well as the
required allocation alignment for the rseq per-thread area to user-space
through ELF auxiliary vector entries.
This is part of the extensible rseq ABI.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20221122203932.231377-3-mathieu.desnoyers@efficios.com
Peter Zijlstra