linux/kernel/sched
Jason Low 1018016c70 sched, timer: Replace spinlocks with atomics in thread_group_cputimer(), to improve scalability
While running a database workload, we found a scalability issue with itimers.

Much of the problem was caused by the thread_group_cputimer spinlock.
Each time we account for group system/user time, we need to obtain a
thread_group_cputimer's spinlock to update the timers. On larger systems
(such as a 16 socket machine), this caused more than 30% of total time
spent trying to obtain this kernel lock to update these group timer stats.

This patch converts the timers to 64-bit atomic variables and use
atomic add to update them without a lock. With this patch, the percent
of total time spent updating thread group cputimer timers was reduced
from 30% down to less than 1%.

Note: On 32-bit systems using the generic 64-bit atomics, this causes
sample_group_cputimer() to take locks 3 times instead of just 1 time.
However, we tested this patch on a 32-bit system ARM system using the
generic atomics and did not find the overhead to be much of an issue.
An explanation for why this isn't an issue is that 32-bit systems usually
have small numbers of CPUs, and cacheline contention from extra spinlocks
called periodically is not really apparent on smaller systems.

Signed-off-by: Jason Low <jason.low2@hp.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Rik van Riel <riel@redhat.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Aswin Chandramouleeswaran <aswin@hp.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Mike Galbraith <umgwanakikbuti@gmail.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Preeti U Murthy <preeti@linux.vnet.ibm.com>
Cc: Scott J Norton <scott.norton@hp.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Waiman Long <Waiman.Long@hp.com>
Link: http://lkml.kernel.org/r/1430251224-5764-4-git-send-email-jason.low2@hp.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-05-08 12:15:31 +02:00
..
auto_group.c sched, timer: Convert usages of ACCESS_ONCE() in the scheduler to READ_ONCE()/WRITE_ONCE() 2015-05-08 12:11:32 +02:00
auto_group.h sched, timer: Convert usages of ACCESS_ONCE() in the scheduler to READ_ONCE()/WRITE_ONCE() 2015-05-08 12:11:32 +02:00
clock.c kernel/sched/clock.c: add another clock for use with the soft lockup watchdog 2015-02-12 18:54:13 -08:00
completion.c sched/completion: Serialize completion_done() with complete() 2015-02-18 14:27:40 +01:00
core.c sched, timer: Convert usages of ACCESS_ONCE() in the scheduler to READ_ONCE()/WRITE_ONCE() 2015-05-08 12:11:32 +02:00
cpuacct.c cgroup: rename cgroup_subsys->base_cftypes to ->legacy_cftypes 2014-07-15 11:05:09 -04:00
cpuacct.h sched/cpuacct: Initialize root cpuacct earlier 2013-04-10 13:54:20 +02:00
cpudeadline.c sched/deadline: Remove cpu_active_mask from cpudl_find() 2015-02-04 07:52:29 +01:00
cpudeadline.h sched/deadline: Modify cpudl::free_cpus to reflect rd->online 2015-01-30 19:39:16 +01:00
cpupri.c Merge commit '3cf2f34' into sched/core, to fix build error 2014-06-12 13:46:37 +02:00
cpupri.h sched/cpupri: Remove unnecessary definitions in cpupri.h 2014-11-16 10:58:59 +01:00
cputime.c sched, timer: Convert usages of ACCESS_ONCE() in the scheduler to READ_ONCE()/WRITE_ONCE() 2015-05-08 12:11:32 +02:00
deadline.c sched, timer: Convert usages of ACCESS_ONCE() in the scheduler to READ_ONCE()/WRITE_ONCE() 2015-05-08 12:11:32 +02:00
debug.c sched: Track group sched_entity usage contributions 2015-03-27 09:35:58 +01:00
fair.c sched/numa: Document usages of mm->numa_scan_seq 2015-05-08 12:13:13 +02:00
features.h sched/rt: Use IPI to trigger RT task push migration instead of pulling 2015-03-23 10:55:22 +01:00
idle_task.c sched: Provide update_curr callbacks for stop/idle scheduling classes 2014-11-23 14:14:40 -08:00
idle.c cpuidle: Run tick_broadcast_exit() with disabled interrupts 2015-04-29 15:19:21 +02:00
loadavg.c sched: Move the loadavg code to a more obvious location 2015-05-08 12:04:12 +02:00
Makefile sched: Move the loadavg code to a more obvious location 2015-05-08 12:04:12 +02:00
rt.c sched, timer: Convert usages of ACCESS_ONCE() in the scheduler to READ_ONCE()/WRITE_ONCE() 2015-05-08 12:11:32 +02:00
sched.h sched, timer: Convert usages of ACCESS_ONCE() in the scheduler to READ_ONCE()/WRITE_ONCE() 2015-05-08 12:11:32 +02:00
stats.c sched: use %*pb[l] to print bitmaps including cpumasks and nodemasks 2015-02-13 21:21:37 -08:00
stats.h sched, timer: Replace spinlocks with atomics in thread_group_cputimer(), to improve scalability 2015-05-08 12:15:31 +02:00
stop_task.c sched: Provide update_curr callbacks for stop/idle scheduling classes 2014-11-23 14:14:40 -08:00
wait.c sched, timer: Convert usages of ACCESS_ONCE() in the scheduler to READ_ONCE()/WRITE_ONCE() 2015-05-08 12:11:32 +02:00