From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id S1751246Ab1ADQOL (ORCPT <rfc822;w@1wt.eu>);
	Tue, 4 Jan 2011 11:14:11 -0500
Received: from mx1.redhat.com ([209.132.183.28]:37731 "EHLO mx1.redhat.com"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1751037Ab1ADQOJ (ORCPT <rfc822;linux-kernel@vger.kernel.org>);
	Tue, 4 Jan 2011 11:14:09 -0500
Date: Tue, 4 Jan 2011 17:06:12 +0100
From: Oleg Nesterov <oleg@redhat.com>
To: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Chris Mason <chris.mason@oracle.com>,
        Frank Rowand <frank.rowand@am.sony.com>, Ingo Molnar <mingo@elte.hu>,
        Thomas Gleixner <tglx@linutronix.de>, Mike Galbraith <efault@gmx.de>,
        Paul Turner <pjt@google.com>, Jens Axboe <axboe@kernel.dk>,
        Yong Zhang <yong.zhang0@gmail.com>, linux-kernel@vger.kernel.org
Subject: Re: [RFC][PATCH 16/17] sched: Move the second half of ttwu() to
	the remote cpu
Message-ID: <20110104160612.GA9177@redhat.com>
References: <20101224122338.172750730@chello.nl> <20101224123743.303699501@chello.nl> <20110104142805.GA4347@redhat.com> <1294152441.2016.148.camel@laptop> <20110104151826.GA6800@redhat.com> <1294155781.2016.160.camel@laptop>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1294155781.2016.160.camel@laptop>
User-Agent: Mutt/1.5.18 (2008-05-17)
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On 01/04, Peter Zijlstra wrote:
>
> On Tue, 2011-01-04 at 16:18 +0100, Oleg Nesterov wrote:
>
> > > I don't think so, nobody should be migrating a TASK_WAKING task.
> >
> > I am not sure...
> >
> > Suppose that p was TASK_INTERRUPTIBLE and p->on_rq == 1 before, when
> > set_cpus_allowed_ptr() was called. To simplify, suppose that
> > the caller is preempted right after it drops p->pi_lock and before
> > it does stop_one_cpu(migration_cpu_stop).
> >
> > After that p can complete chedule() and deactivate itself.
> >
> > Now, try_to_wake_up() can set TASK_WAKING, choose another CPU,
> > and do ttwu_queue_remote().
> >
> > Finally, the caller of set_cpus_allowed_ptr() resumes and
> > schedules migration_cpu_stop.
>
> But __migrate_task() will then find !p->on_rq and not actually do
> anything.

But if it races with ttwu_do_activate() (which can hold another
rq->lock != rq_src/rq_dest), it can check p->on_rq after
activate_task() was already called.



But I think this no longer matters,

> > It is very possible I missed something, but what is the new
> > locking rules for set_task_cpu() anyway? I mean, which rq->lock
> > it needs?
>
> In the patches I just posted: set_task_cpu(p, cpu) callers need to
> either hold task_rq(p)->lock (the old rq) or p->pi_lock.

Yes, thanks, I already noticed v4, and at first glance 11/18 plus
set_task_cpu() in try_to_wake_up() should close the problems...

Oleg.