Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1583373 > unrolled thread

[PATCH v2 0/2] sched/loadavg: Fix loadavg spikes and sprinkle {READ,WRITE}_ONCE()

Back to article view | Back to linux.kernel

Contents

  [PATCH v2 0/2] sched/loadavg: Fix loadavg spikes and sprinkle {READ,WRITE}_ONCE() Matt Fleming <matt@codeblueprint.co.uk> - 2017-02-17 13:10 +0100
    [PATCH v2 1/2] sched/loadavg: Avoid loadavg spikes caused by delayed NO_HZ accounting Matt Fleming <matt@codeblueprint.co.uk> - 2017-02-17 13:10 +0100
      Re: [PATCH v2 1/2] sched/loadavg: Avoid loadavg spikes caused by  delayed NO_HZ accounting Frederic Weisbecker <fweisbec@gmail.com> - 2017-02-22 16:20 +0100

#1583373 — [PATCH v2 0/2] sched/loadavg: Fix loadavg spikes and sprinkle {READ,WRITE}_ONCE()

Peter, Frederic, 

Here's a v2 of the patch series to fix the loadavg spikes I'm seeing
on a v3.12 based kernel, caused by delaying load sampling if a sample
period was crossed while in NO_HZ idle.

I tried to make the changelog for PATCH 1 clearer this time around by
incorporating suggestions from both of you. Please let me know if it's
still unclear.

PATCH 2 addresses Peter's comment:

  "Irrespective the above though; should we not make this:
   
   +       this_rq->calc_load_update = READ_ONCE(calc_load_update);
   
   because if for some reason we do a double load of calc_load_update and
   see two different values, weird stuff could happen.
   
   And because, on general principle, a READ_ONCE() should be paired with a
   WRITE_ONCE(), that should be done too I suppose."

The v1 of the patch can be found here:

  https://lkml.kernel.org/r/20170208132924.3038-1-matt@codeblueprint.co.uk

Matt Fleming (2):
  sched/loadavg: Avoid loadavg spikes caused by delayed NO_HZ accounting
  sched/loadavg: Use {READ,WRITE}_ONCE() for sample window

 kernel/sched/loadavg.c | 20 ++++++++++++--------
 1 file changed, 12 insertions(+), 8 deletions(-)

-- 
2.10.0

[toc] | [next] | [standalone]

#1583374 — [PATCH v2 1/2] sched/loadavg: Avoid loadavg spikes caused by delayed NO_HZ accounting

If we crossed a sample window while in NO_HZ we will add LOAD_FREQ to
the pending sample window time on exit, setting the next update not
one window into the future, but two.

This situation on exiting NO_HZ is described by:

  this_rq->calc_load_update < jiffies < calc_load_update

In this scenario, what we should be doing is:

  this_rq->calc_load_update = calc_load_update		     [ next window ]

But what we actually do is:

  this_rq->calc_load_update = calc_load_update + LOAD_FREQ   [ next+1 window ]

This has the effect of delaying load average updates for potentially
up to ~9seconds.

This can result in huge spikes in the load average values due to
per-cpu uninterruptible task counts being out of sync when accumulated
across all CPUs.

It's safe to update the per-cpu active count if we wake between sample
windows because any load that we left in 'calc_load_idle' will have
been zero'd when the idle load was folded in calc_global_load().

This issue is easy to reproduce before,

  commit 9d89c257dfb9 ("sched/fair: Rewrite runnable load and utilization average tracking")

just by forking short-lived process pipelines built from ps(1) and
grep(1) in a loop. I'm unable to reproduce the spikes after that
commit, but the bug still seems to be present from code review.

Fixes: commit 5167e8d ("sched/nohz: Rewrite and fix load-avg computation -- again")
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Mike Galbraith <umgwanakikbuti@gmail.com>
Cc: Morten Rasmussen <morten.rasmussen@arm.com>
Cc: Vincent Guittot <vincent.guittot@linaro.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: <stable@vger.kernel.org> # v3.5+
Signed-off-by: Matt Fleming <matt@codeblueprint.co.uk>
---
 kernel/sched/loadavg.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

Changes in v2:

 - Folded in Peter's suggestion for how to fix this.

 - Tried to clairfy the changelog based on feedback from Peter and
   Frederic

diff --git a/kernel/sched/loadavg.c b/kernel/sched/loadavg.c
index a2d6eb71f06b..ec91fcc09bfe 100644
--- a/kernel/sched/loadavg.c
+++ b/kernel/sched/loadavg.c
@@ -201,8 +201,9 @@ void calc_load_exit_idle(void)
 	struct rq *this_rq = this_rq();
 
 	/*
-	 * If we're still before the sample window, we're done.
+	 * If we're still before the pending sample window, we're done.
 	 */
+	this_rq->calc_load_update = calc_load_update;
 	if (time_before(jiffies, this_rq->calc_load_update))
 		return;
 
@@ -211,7 +212,6 @@ void calc_load_exit_idle(void)
 	 * accounted through the nohz accounting, so skip the entire deal and
 	 * sync up for the next window.
 	 */
-	this_rq->calc_load_update = calc_load_update;
 	if (time_before(jiffies, this_rq->calc_load_update + 10))
 		this_rq->calc_load_update += LOAD_FREQ;
 }
-- 
2.10.0

[toc] | [prev] | [next] | [standalone]

#1586252 — Re: [PATCH v2 1/2] sched/loadavg: Avoid loadavg spikes caused by delayed NO_HZ accounting

On Fri, Feb 17, 2017 at 12:07:30PM +0000, Matt Fleming wrote:
> If we crossed a sample window while in NO_HZ we will add LOAD_FREQ to
> the pending sample window time on exit, setting the next update not
> one window into the future, but two.
> 
> This situation on exiting NO_HZ is described by:
> 
>   this_rq->calc_load_update < jiffies < calc_load_update
> 
> In this scenario, what we should be doing is:
> 
>   this_rq->calc_load_update = calc_load_update		     [ next window ]
> 
> But what we actually do is:
> 
>   this_rq->calc_load_update = calc_load_update + LOAD_FREQ   [ next+1 window ]
> 
> This has the effect of delaying load average updates for potentially
> up to ~9seconds.
> 
> This can result in huge spikes in the load average values due to
> per-cpu uninterruptible task counts being out of sync when accumulated
> across all CPUs.
> 
> It's safe to update the per-cpu active count if we wake between sample
> windows because any load that we left in 'calc_load_idle' will have
> been zero'd when the idle load was folded in calc_global_load().
> 
> This issue is easy to reproduce before,
> 
>   commit 9d89c257dfb9 ("sched/fair: Rewrite runnable load and utilization average tracking")
> 
> just by forking short-lived process pipelines built from ps(1) and
> grep(1) in a loop. I'm unable to reproduce the spikes after that
> commit, but the bug still seems to be present from code review.
> 
> Fixes: commit 5167e8d ("sched/nohz: Rewrite and fix load-avg computation -- again")
> Cc: Peter Zijlstra <peterz@infradead.org>
> Cc: Mike Galbraith <umgwanakikbuti@gmail.com>
> Cc: Morten Rasmussen <morten.rasmussen@arm.com>
> Cc: Vincent Guittot <vincent.guittot@linaro.org>

Acked-by: Frederic Weisbecker <fweisbec@gmail.com>

Thanks it's much clearer now!

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web