Re: [v4.8-rc1 Regression] sched/fair: Apply more PELT fixes

From: Vincent Guittot <vincent.guittot@linaro.org>
To: Dietmar Eggemann <dietmar.eggemann@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>,
	Joseph Salisbury <joseph.salisbury@canonical.com>,
	Ingo Molnar <mingo@kernel.org>,
	Linus Torvalds <torvalds@linux-foundation.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	LKML <linux-kernel@vger.kernel.org>,
	Mike Galbraith <efault@gmx.de>,
	omer.akram@canonical.com
Subject: Re: [v4.8-rc1 Regression] sched/fair: Apply more PELT fixes
Date: Wed, 19 Oct 2016 19:50:57 +0200	[thread overview]
Message-ID: <CAKfTPtC-jdC5PFX6AWkT=tm6BptxC-Ty06Q-rHCxR4-Lk6OYBg@mail.gmail.com> (raw)
In-Reply-To: <ee53f6b5-0fc9-537a-14ee-bfff87338369@arm.com>

On 19 October 2016 at 17:33, Dietmar Eggemann <dietmar.eggemann@arm.com> wrote:
> On 19/10/16 12:25, Vincent Guittot wrote:
>> On 19 October 2016 at 11:46, Dietmar Eggemann <dietmar.eggemann@arm.com> wrote:
>>> On 18/10/16 12:56, Vincent Guittot wrote:
>>>> Le Tuesday 18 Oct 2016 à 12:34:12 (+0200), Peter Zijlstra a écrit :
>>>>> On Tue, Oct 18, 2016 at 11:45:48AM +0200, Vincent Guittot wrote:
>>>>>> On 18 October 2016 at 11:07, Peter Zijlstra <peterz@infradead.org> wrote:
>
> [...]
>
>>> But this test only makes sure that we don't see any ghost contribution
>>> (from non-existing cpus) any more.
>>>
>>> We should study the tg->se[i]->avg.load_avg for the hierarchy of tg's
>>> (with the highest tg having a task enqueued) a little bit more, with and
>>> without your v5 'sched: reflect sched_entity move into task_group's load'.
>>
>> Can you elaborate ?
>
> I try :-)
>
> I thought I will see some different behaviour because of the fact that
> the tg se's are initialized differently [1024 versus 0].

This difference should be noticeable (if noticeable) only during few
hundreds of ms after the creation of the task group until the load_avg
has reached its real value.

>
> But I can't spot any difference. The test case is running a sysbench
> thread affine to cpu1 in tg_root/tg_1/tg_11/tg_111 on tip/sched/core on
> an ARM64 Juno (6 logical cpus).
> The moment the sysbench task is put into tg_111
> tg_111->se[1]->avg.load_avg gets updated to 0 any way because of the
> huge time difference between creating this tg and attaching a task to
> it. So the tg->se[2]->avg.load_avg signals for tg_111, tg_11 and tg_1
> look exactly the same w/o and w/ your patch.
>
> But your patch helps in this (very synthetic) test case as well. W/o
> your patch I see remaining tg->load_avg for tg_1 and tg_11 after the
> test case has finished because the tg's were exclusively used on cpu1.
>
> # cat /proc/sched_debug
>
>  cfs_rq[1]:/tg_1
>    .tg_load_avg_contrib           : 0
>    .tg_load_avg                   : 5120 (5 (unused cpus) * 1024 * 1)
>  cfs_rq[1]:/tg_1/tg_11/tg_111
>    .tg_load_avg_contrib           : 0
>    .tg_load_avg                   : 0
>  cfs_rq[1]:/tg_1/tg_11
>    .tg_load_avg_contrib           : 0
>    .tg_load_avg                   : 5120
>
> With your patch applied all the .tg_load_avg are 0.