linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Valentin Schneider <valentin.schneider@arm.com>
To: Peter Zijlstra <peterz@infradead.org>
Cc: tglx@linutronix.de, mingo@kernel.org,
	linux-kernel@vger.kernel.org, bigeasy@linutronix.de,
	qais.yousef@arm.com, swood@redhat.com, juri.lelli@redhat.com,
	vincent.guittot@linaro.org, dietmar.eggemann@arm.com,
	rostedt@goodmis.org, bsegall@google.com, mgorman@suse.de,
	bristot@redhat.com, vincent.donnefort@arm.com, tj@kernel.org
Subject: Re: [PATCH -v2 09/17] sched: Add migrate_disable()
Date: Tue, 06 Oct 2020 12:20:27 +0100	[thread overview]
Message-ID: <jhjwo03mwhw.mognet@arm.com> (raw)
In-Reply-To: <20201005150921.859981475@infradead.org>


On 05/10/20 15:57, Peter Zijlstra wrote:
> Add the base migrate_disable() support (under protest).
>
> While migrate_disable() is (currently) required for PREEMPT_RT, it is
> also one of the biggest flaws in the system.
>
> Notably this is just the base implementation, it is broken vs
> sched_setaffinity() and hotplug, both solved in additional patches for
> ease of review.
>
> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
> ---
[...]
> @@ -2363,7 +2451,7 @@ int select_task_rq(struct task_struct *p
>  {
>       lockdep_assert_held(&p->pi_lock);
>
> -	if (p->nr_cpus_allowed > 1)
> +	if (p->nr_cpus_allowed > 1 && !is_migration_disabled(p))

So dissociating migrate_disable() from p->nr_cpus_allowed unlocks the whole
DL+RT push/pull thingie, but it also means we need to be somewhat careful
regarding checks like the above.

Looking at current p->nr_cpus_allowed readers, it's mostly DL/RT; I was
somewhat afraid CFS load balance would use it but turns out it doesn't. The
only borderline case I see is call_on_cpu(), and even then it would be just a
performance deal rather than a correctness one.

All that to say: it seems fine to me.

>               cpu = p->sched_class->select_task_rq(p, cpu, sd_flags, wake_flags);
>       else
>               cpu = cpumask_any(p->cpus_ptr);

  reply	other threads:[~2020-10-06 11:20 UTC|newest]

Thread overview: 58+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-10-05 14:57 [PATCH -v2 00/17] sched: Migrate disable support Peter Zijlstra
2020-10-05 14:57 ` [PATCH -v2 01/17] stop_machine: Add function and caller debug info Peter Zijlstra
2020-10-05 14:57 ` [PATCH -v2 02/17] sched: Fix balance_callback() Peter Zijlstra
2020-10-05 14:57 ` [PATCH -v2 03/17] sched/hotplug: Ensure only per-cpu kthreads run during hotplug Peter Zijlstra
2020-10-06  7:25   ` Peter Zijlstra
2020-10-05 14:57 ` [PATCH -v2 04/17] sched/core: Wait for tasks being pushed away on hotplug Peter Zijlstra
2020-10-05 14:57 ` [PATCH -v2 05/17] workqueue: Manually break affinity " Peter Zijlstra
2020-10-05 16:31   ` Tejun Heo
2020-10-05 14:57 ` [PATCH -v2 06/17] sched/hotplug: Consolidate task migration on CPU unplug Peter Zijlstra
2020-10-05 14:57 ` [PATCH -v2 07/17] sched: Fix hotplug vs CPU bandwidth control Peter Zijlstra
2020-10-06 12:46   ` Vincent Guittot
2020-10-06 13:33     ` Peter Zijlstra
2020-10-09 20:41   ` Dietmar Eggemann
2020-10-12 12:52     ` Peter Zijlstra
2020-10-12 13:18       ` Peter Zijlstra
2020-10-12 14:14         ` Dietmar Eggemann
2020-10-05 14:57 ` [PATCH -v2 08/17] sched: Massage set_cpus_allowed() Peter Zijlstra
2020-10-06 11:20   ` Valentin Schneider
2020-10-05 14:57 ` [PATCH -v2 09/17] sched: Add migrate_disable() Peter Zijlstra
2020-10-06 11:20   ` Valentin Schneider [this message]
2020-10-05 14:57 ` [PATCH -v2 10/17] sched: Fix migrate_disable() vs set_cpus_allowed_ptr() Peter Zijlstra
     [not found]   ` <CH2PR14MB41833F828B4D3BA5A7B6CE7B9A0B0@CH2PR14MB4183.namprd14.prod.outlook.com>
2020-10-12 11:36     ` Peter Zijlstra
     [not found]       ` <CH2PR14MB4183BF26F02A4AD4F45CADA59A070@CH2PR14MB4183.namprd14.prod.outlook.com>
2020-10-13 13:20         ` Valentin Schneider
2020-10-05 14:57 ` [PATCH -v2 11/17] sched/core: Make migrate disable and CPU hotplug cooperative Peter Zijlstra
2020-10-05 14:57 ` [PATCH -v2 12/17] sched,rt: Use cpumask_any*_distribute() Peter Zijlstra
2020-10-06 14:09   ` Juri Lelli
2020-10-06 14:20     ` Peter Zijlstra
2020-10-06 15:55   ` Qais Yousef
2020-10-07  7:22     ` Peter Zijlstra
2020-10-05 14:57 ` [PATCH -v2 13/17] sched,rt: Use the full cpumask for balancing Peter Zijlstra
2020-10-05 14:57 ` [PATCH -v2 14/17] sched, lockdep: Annotate ->pi_lock recursion Peter Zijlstra
2020-10-05 14:57 ` [PATCH -v2 15/17] sched: Fix migrate_disable() vs rt/dl balancing Peter Zijlstra
2020-10-06  7:59   ` Peter Zijlstra
2020-10-06 13:44     ` Steven Rostedt
2020-10-06 13:50       ` Peter Zijlstra
2020-10-06 11:20   ` Valentin Schneider
2020-10-06 13:48     ` Peter Zijlstra
2020-10-06 14:37       ` Juri Lelli
2020-10-06 14:48         ` Peter Zijlstra
2020-10-07  5:29           ` Juri Lelli
2020-10-06 16:19       ` Valentin Schneider
2020-10-07 19:22   ` Valentin Schneider
2020-10-07 21:08     ` Peter Zijlstra
2020-10-07 22:32       ` Valentin Schneider
2020-10-08 10:48   ` Sebastian Andrzej Siewior
2020-10-12  9:56   ` Dietmar Eggemann
2020-10-12 11:28     ` Peter Zijlstra
2020-10-12 12:22       ` Dietmar Eggemann
2020-10-05 14:57 ` [PATCH -v2 16/17] sched/proc: Print accurate cpumask vs migrate_disable() Peter Zijlstra
2020-10-05 14:57 ` [PATCH -v2 17/17] sched: Add migrate_disable() tracepoints Peter Zijlstra
2020-10-13 14:01 ` [PATCH 1/2] sched: Deny self-issued __set_cpus_allowed_ptr() when migrate_disable() Valentin Schneider
2020-10-13 14:15   ` Sebastian Andrzej Siewior
2020-10-13 14:21     ` Peter Zijlstra
2020-10-15  9:14       ` Valentin Schneider
2020-10-13 14:22     ` Valentin Schneider
2020-11-11  8:23   ` [tip: sched/core] " tip-bot2 for Valentin Schneider
2020-10-13 14:01 ` [PATCH 2/2] sched: Comment affine_move_task() Valentin Schneider
2020-11-11  8:23   ` [tip: sched/core] " tip-bot2 for Valentin Schneider

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=jhjwo03mwhw.mognet@arm.com \
    --to=valentin.schneider@arm.com \
    --cc=bigeasy@linutronix.de \
    --cc=bristot@redhat.com \
    --cc=bsegall@google.com \
    --cc=dietmar.eggemann@arm.com \
    --cc=juri.lelli@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mgorman@suse.de \
    --cc=mingo@kernel.org \
    --cc=peterz@infradead.org \
    --cc=qais.yousef@arm.com \
    --cc=rostedt@goodmis.org \
    --cc=swood@redhat.com \
    --cc=tglx@linutronix.de \
    --cc=tj@kernel.org \
    --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;
as well as URLs for NNTP newsgroup(s).