linux/kernel/sched
Peter Zijlstra e3baac47f0 sched/idle: Optimize try-to-wake-up IPI
[ This series reduces the number of IPIs on Andy's workload by something like
  99%. It's down from many hundreds per second to very few.

  The basic idea behind this series is to make TIF_POLLING_NRFLAG be a
  reliable indication that the idle task is polling.  Once that's done,
  the rest is reasonably straightforward. ]

When enqueueing tasks on remote LLC domains, we send an IPI to do the
work 'locally' and avoid bouncing all the cachelines over.

However, when the remote CPU is idle (and polling, say x86 mwait), we
don't need to send an IPI, we can simply kick the TIF word to wake it
up and have the 'idle' loop do the work.

So when _TIF_POLLING_NRFLAG is set, but _TIF_NEED_RESCHED is not (yet)
set, set _TIF_NEED_RESCHED and avoid sending the IPI.

Much-requested-by: Andy Lutomirski <luto@amacapital.net>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
[Edited by Andy Lutomirski, but this is mostly Peter Zijlstra's code.]
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: nicolas.pitre@linaro.org
Cc: daniel.lezcano@linaro.org
Cc: Mike Galbraith <umgwanakikbuti@gmail.com>
Cc: umgwanakikbuti@gmail.com
Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: linux-kernel@vger.kernel.org
Link: http://lkml.kernel.org/r/ce06f8b02e7e337be63e97597fc4b248d3aa6f9b.1401902905.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-06-05 12:09:53 +02:00
..
auto_group.c sched: Replace hardcoding of -20 and 19 with MIN_NICE and MAX_NICE 2014-02-22 18:15:54 +01:00
auto_group.h Revert "sched/autogroup: Fix crash on reboot when autogroup is disabled" 2012-12-11 10:23:45 +01:00
clock.c kernel: use macros from compiler.h instead of __attribute__((...)) 2014-04-07 16:36:11 -07:00
completion.c sched: Move completion code from core.c to completion.c 2013-11-06 07:49:19 +01:00
core.c sched/idle: Optimize try-to-wake-up IPI 2014-06-05 12:09:53 +02:00
cpuacct.c cgroup: clean up cgroup_subsys names and initialization 2014-02-08 10:36:58 -05:00
cpuacct.h sched/cpuacct: Initialize root cpuacct earlier 2013-04-10 13:54:20 +02:00
cpudeadline.c sched/deadline: Replace NR_CPUS arrays 2014-05-22 10:21:28 +02:00
cpudeadline.h sched/deadline: Replace NR_CPUS arrays 2014-05-22 10:21:28 +02:00
cpupri.c sched/cpupri: Replace NR_CPUS arrays 2014-05-22 10:21:29 +02:00
cpupri.h sched/cpupri: Replace NR_CPUS arrays 2014-05-22 10:21:29 +02:00
cputime.c sched: Sanitize irq accounting madness 2014-05-07 11:51:30 +02:00
deadline.c sched/rt: Fix 'struct sched_dl_entity' and dl_task_time() comments, to match the current upstream code 2014-05-22 11:16:37 +02:00
debug.c Merge branch 'for-3.15' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup 2014-04-03 13:05:42 -07:00
fair.c sched: Rename capacity related flags 2014-06-05 11:52:32 +02:00
features.h sched: Rename capacity related flags 2014-06-05 11:52:32 +02:00
idle.c sched/idle: Optimize try-to-wake-up IPI 2014-06-05 12:09:53 +02:00
idle_task.c sched/fair: Push down check for high priority class task into idle_balance() 2014-03-11 12:05:37 +01:00
Makefile sched/idle: Move cpu/idle.c to sched/idle.c 2014-02-11 09:58:30 +01:00
proc.c sched: Change get_rq_runnable_load() to static and inline 2013-06-27 10:07:44 +02:00
rt.c sched: Remove redundant assignment to "rt_rq" in update_curr_rt(...) 2014-06-05 11:52:34 +02:00
sched.h sched/idle: Optimize try-to-wake-up IPI 2014-06-05 12:09:53 +02:00
stats.c kernel: audit/fix non-modular users of module_init in core code 2014-04-03 16:21:07 -07:00
stats.h sched: Micro-optimize by dropping unnecessary task_rq() calls 2013-09-25 13:51:06 +02:00
stop_task.c sched, nohz: Change rq->nr_running to always use wrappers 2014-05-22 11:16:33 +02:00
wait.c sched: Move wait code from core.c to wait.c 2013-11-06 07:49:18 +01:00