From: Dietmar Eggemann <dietmar.eggemann@arm.com>
To: Valentin Schneider <valentin.schneider@arm.com>,
Vincent Donnefort <vincent.donnefort@arm.com>,
peterz@infradead.org, mingo@redhat.com,
vincent.guittot@linaro.org
Cc: linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2] sched/fair: Fix per-CPU kthread and wakee stacking for asym CPU capacity
Date: Mon, 29 Nov 2021 17:56:11 +0100 [thread overview]
Message-ID: <13c4ca53-61d6-b90d-7344-94f724c0cf15@arm.com> (raw)
In-Reply-To: <87mtlnb5ae.mognet@arm.com>
On 29.11.21 11:54, Valentin Schneider wrote:
> On 25/11/21 10:12, Vincent Donnefort wrote:
>> select_idle_sibling() has a special case for tasks woken up by a per-CPU
>> kthread. For this case, the chosen CPU is the previous one. This is an
>> issue for asymmetric CPU capacity systems where the wakee might not fit
>> that CPU anymore. Evaluate asym_fits_capacity() for prev_cpu before using
>> the exit path described above.
>>
>> Fixes: b4c9c9f15649 ("sched/fair: Prefer prev cpu in asymmetric wakeup path")
>> Signed-off-by: Vincent Donnefort <vincent.donnefort@arm.com>
>
> Per our discussion on v1, the asym check was intentionally omitted, the
> assumption being: we'd be putting p back on its prev CPU, its utilization
> cannot be bigger now than it was then so it should still pass the capacity
> fitness criterion (unless we dequeued it right before crossing the next
> PELT window boundary would have made it cross the tipping point...)
... and assume ~0 sleep time for p so sync_entity_load_avg() can't decay
p's util_avg.
>
> Uclamp goes against this, p's uclamp.min can completely change between its
> dequeue and wakeup, which warrants adding the check.
>
> I'd like to see (at least some of) the above in the changelog, but
+1 (since otherwise the WHY the wakee could potentially not fit on prev
anymore is hard to grasp).
> pedantism aside:
>> Reviewed-by: Valentin Schneider <Valentin.Schneider@arm.com>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
>
>>
>> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>> index 6291876a9d32..b90dc6fd86ca 100644
>> --- a/kernel/sched/fair.c
>> +++ b/kernel/sched/fair.c
>> @@ -6410,7 +6410,8 @@ static int select_idle_sibling(struct task_struct *p, int prev, int target)
>> */
>> if (is_per_cpu_kthread(current) &&
>> prev == smp_processor_id() &&
>> - this_rq()->nr_running <= 1) {
>> + this_rq()->nr_running <= 1 &&
>> + asym_fits_capacity(task_util, prev)) {
>> return prev;
>> }
>>
>> --
>> 2.25.1
prev parent reply other threads:[~2021-11-29 16:58 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-11-25 10:12 Vincent Donnefort
2021-11-29 10:54 ` Valentin Schneider
2021-11-29 16:56 ` Dietmar Eggemann [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=13c4ca53-61d6-b90d-7344-94f724c0cf15@arm.com \
--to=dietmar.eggemann@arm.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
--cc=valentin.schneider@arm.com \
--cc=vincent.donnefort@arm.com \
--cc=vincent.guittot@linaro.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®