Re: [RFC][PATCH 1/4] sched: Fix a race between __kthread_bind() and sched_setaffinity()

From: Tejun Heo <tj@kernel.org>
To: Peter Zijlstra <peterz@infradead.org>
Cc: mingo@kernel.org, riel@redhat.com, dedekind1@gmail.com,
	linux-kernel@vger.kernel.org, mgorman@suse.de,
	rostedt@goodmis.org, juri.lelli@arm.com,
	Oleg Nesterov <oleg@redhat.com>
Subject: Re: [RFC][PATCH 1/4] sched: Fix a race between __kthread_bind() and sched_setaffinity()
Date: Fri, 15 May 2015 11:56:53 -0400	[thread overview]
Message-ID: <20150515155653.GA23692@htj.duckdns.org> (raw)
In-Reply-To: <20150515154833.545640346@infradead.org>

On Fri, May 15, 2015 at 05:43:34PM +0200, Peter Zijlstra wrote:
> Because sched_setscheduler() checks p->flags & PF_NO_SETAFFINITY
> without locks, a caller might observe an old value and race with the
> set_cpus_allowed_ptr() call from __kthread_bind() and effectively undo
> it.
> 
> 	__kthread_bind()
> 	  do_set_cpus_allowed()
> 						<SYSCALL>
> 						  sched_setaffinity()
> 						    if (p->flags & PF_NO_SETAFFINITIY)
> 						    set_cpus_allowed_ptr()
> 	  p->flags |= PF_NO_SETAFFINITY
> 
> Fix the issue by putting everything under the regular scheduler locks.
> 
> This also closes a hole in the serialization of
> task_struct::{nr_,}cpus_allowed.
> 
> Cc: Tejun Heo <tj@kernel.org>
> Cc: Oleg Nesterov <oleg@redhat.com>
> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>

For workqueue part,

 Acked-by: Tejun Heo <tj@kernel.org>

Please route anyway you see fit.

Thanks.

-- 
tejun