mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Yong Zhang <yong.zhang0@gmail.com>
To: Peter Zijlstra <peterz@infradead.org>
Cc: samu.p.onkalo@nokia.com, mingo@elte.hu,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	tglx <tglx@linutronix.de>
Subject: Re: Bug in scheduler when using rt_mutex
Date: Wed, 19 Jan 2011 10:38:57 +0800	[thread overview]
Message-ID: <AANLkTimLSict5oq5PefDaMLB9q2fK91E8XgbUDEohP-_@mail.gmail.com> (raw)
In-Reply-To: <1295357746.30950.681.camel@laptop>

On Tue, Jan 18, 2011 at 9:35 PM, Peter Zijlstra <peterz@infradead.org> wrote:
> Subject: sched: Fix switch_to_fair()
> From: Peter Zijlstra <a.p.zijlstra@chello.nl>
> Date: Mon Jan 17 17:03:27 CET 2011
>
> When a task is placed back into fair_sched_class, we must update its
> placement, since we don't know how long its been gone, hence its
> vruntime is stale and cannot be trusted.
>
> There is also a case where it was moved from fair_sched_class when it
> was in a blocked state and moved back while it is running, this causes
> an imbalance between DEQUEUE_SLEEP/DEQUEUE_WAKEUP for the fair class
> and leaves vruntime way out there (due to the min_vruntime
> adjustment).
>
> Also update sysrq-n to call the ->switch_{to,from} methods.
>
> Reported-by: Onkalo Samu <samu.p.onkalo@nokia.com>
> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
> ---
>  kernel/sched.c      |    4 ++++
>  kernel/sched_fair.c |   16 ++++++++++++++++
>  2 files changed, 20 insertions(+)
>
> Index: linux-2.6/kernel/sched_fair.c
> ===================================================================
> --- linux-2.6.orig/kernel/sched_fair.c
> +++ linux-2.6/kernel/sched_fair.c
> @@ -4075,6 +4075,22 @@ static void prio_changed_fair(struct rq
>  static void switched_to_fair(struct rq *rq, struct task_struct *p,
>                             int running)
>  {
> +       struct sched_entity *se = &p->se;
> +       struct cfs_rq *cfs_rq = cfs_rq_of(se);
> +
> +       if (se->on_rq && cfs_rq->curr != se)

(cfs_rq->curr != se) equals to (!running), no?

> +               __dequeue_entity(cfs_rq, se);
> +
> +       /*
> +        * se->vruntime can be completely out there, there is no telling
> +        * how long this task was !fair and on what CPU if any it became
> +        * !fair. Therefore, reset it to a known, reasonable value.
> +        */
> +       se->vruntime = cfs_rq->min_vruntime;

But this is not fair for !SLEEP task.
You know se->vruntime -= cfs_rq->min_vruntime for !SLEEP task,
then after it go through sched_fair-->sched_rt-->sched_fair by some
means, current cfs_rq->min_vruntime is added back.

But here se is putted before where it should be. Is this what we want?

Thanks,
Yong

-- 
Only stand for myself

  parent reply	other threads:[~2011-01-19  2:39 UTC|newest]

Thread overview: 42+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-01-17 14:42 Onkalo Samu
2011-01-17 15:00 ` Peter Zijlstra
2011-01-17 15:15   ` samu.p.onkalo
2011-01-17 15:28     ` Peter Zijlstra
2011-01-17 16:00 ` Peter Zijlstra
2011-01-18  8:23   ` Onkalo Samu
2011-01-18  8:59     ` Yong Zhang
2011-01-18 13:35       ` Peter Zijlstra
2011-01-18 14:25         ` Onkalo Samu
2011-01-19  2:38         ` Yong Zhang [this message]
2011-01-19  3:43           ` Mike Galbraith
2011-01-19  4:35             ` Yong Zhang
2011-01-19  5:40               ` Mike Galbraith
2011-01-19  6:09                 ` Yong Zhang
2011-01-19  6:37                   ` Mike Galbraith
2011-01-19  7:19                     ` Ingo Molnar
2011-01-19  7:41                       ` Mike Galbraith
2011-01-19  9:44           ` Peter Zijlstra
2011-01-19 10:38             ` Peter Zijlstra
2011-01-19 11:30               ` Peter Zijlstra
2011-01-19 12:58                 ` Onkalo Samu
2011-01-19 13:13                   ` Onkalo Samu
2011-01-19 13:30                     ` Peter Zijlstra
2011-01-20  4:18                       ` Yong Zhang
2011-01-20  4:27                         ` Yong Zhang
2011-01-20  5:32                           ` Yong Zhang
2011-01-20  4:59                         ` Mike Galbraith
2011-01-20  5:30                           ` Yong Zhang
2011-01-20  6:12                             ` Mike Galbraith
2011-01-20  7:06                               ` Yong Zhang
2011-01-20  8:37                                 ` Mike Galbraith
2011-01-20  9:07                                   ` Yong Zhang
2011-01-20 10:07                                     ` Mike Galbraith
2011-01-21 11:08                                       ` Peter Zijlstra
2011-01-21 12:24                                         ` Yong Zhang
2011-01-21 13:40                                           ` Peter Zijlstra
2011-01-21 15:03                                             ` Yong Zhang
2011-01-21 15:10                                               ` Peter Zijlstra
2011-01-21 13:15                                       ` Yong Zhang
2011-01-20  7:07                       ` Onkalo Samu
2011-01-21  6:25                         ` Onkalo Samu
2011-01-20  3:10             ` Yong Zhang

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=AANLkTimLSict5oq5PefDaMLB9q2fK91E8XgbUDEohP-_@mail.gmail.com \
    --to=yong.zhang0@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=peterz@infradead.org \
    --cc=samu.p.onkalo@nokia.com \
    --cc=tglx@linutronix.de \
    /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®