[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20080306111525.GA23865@in.ibm.com>
Date: Thu, 6 Mar 2008 16:45:25 +0530
From: Gautham R Shenoy <ego@...ibm.com>
To: Oleg Nesterov <oleg@...sign.ru>
Cc: Yi Yang <yi.y.yang@...el.com>, Ingo Molnar <mingo@...e.hu>,
akpm@...ux-foundation.org, linux-kernel@...r.kernel.org,
"Rafael J. Wysocki" <rjw@...k.pl>,
Thomas Gleixner <tglx@...utronix.de>
Subject: Re: [BUG 2.6.25-rc3] scheduler/hotplug: some processes are
dealocked when cpu is set to offline
On Wed, Mar 05, 2008 at 04:53:14PM +0300, Oleg Nesterov wrote:
> On 03/05, Gautham R Shenoy wrote:
> >
> > On Tue, Mar 04, 2008 at 06:01:07PM +0300, Oleg Nesterov wrote:
> > > On 03/04, Gautham R Shenoy wrote:
> > > >
> > > > So at times, the callback thread is blocked on kthread_stop(k) in
> > > > softlockup.c, while other time, it was blocked in
> > > > cleanup_workqueue_threads() in workqueue.c.
> > >
> > > >From another message:
> > > >
> > > > However, it remains in R< state
> > >
> > > What about cwq->thread? Was it TASK_RUNNING too?
> >
> > No, it was in TASK_UNINTERRUPTIBLE state. The last thing it ever
> > executed was the wait_for_completion in flush_cpu_workqueue()
>
> You misunderstood ;) The task doing _cpu_down() hang in flush_cpu_workqueue(),
> because cwq->thread doesn't want to die. But what was cwq->thread->state?
>
Yup, sorry for that stupid reply.
> Could you try the patch? Yi constantly sees the tasks in R< state.
> I am starting to suspect some scheduling issue.
Me too. With your patch applied there were quite a few tasks in the
running state which didn't get the cpu for more than 120 seconds.
I could be a scheduler problem which CPU-Hotplug is able to catch more
easily. Will try a git bisect.
>
> Oleg.
--
Thanks and Regards
gautham
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists