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 S1753384AbbI3OCt (ORCPT <rfc822;w@1wt.eu>);
	Wed, 30 Sep 2015 10:02:49 -0400
Received: from mx1.redhat.com ([209.132.183.28]:40276 "EHLO mx1.redhat.com"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1750847AbbI3OCr (ORCPT <rfc822;linux-kernel@vger.kernel.org>);
	Wed, 30 Sep 2015 10:02:47 -0400
Date: Wed, 30 Sep 2015 15:59:37 +0200
From: Oleg Nesterov <oleg@redhat.com>
To: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: akpm@linux-foundation.org, rientjes@google.com, kwalker@redhat.com,
        mhocko@kernel.org, skozina@redhat.com, linux-kernel@vger.kernel.org
Subject: Re: [PATCH -mm 3/3] mm/oom_kill: fix the wrong task->mm == mm
	checks in
Message-ID: <20150930135937.GE32263@redhat.com>
References: <20150929141758.GA10934@redhat.com> <20150929141822.GA10962@redhat.com> <201509301116.GDA41748.HFOOVFLtMQSFOJ@I-love.SAKURA.ne.jp>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <201509301116.GDA41748.HFOOVFLtMQSFOJ@I-love.SAKURA.ne.jp>
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 09/30, Tetsuo Handa wrote:
>
> Oleg Nesterov wrote:
> > Both "child->mm == mm" and "p->mm != mm" checks in oom_kill_process()
> > are wrong. ->mm can be if task is the exited group leader. This means
>
> can be [missing word here?] if task

Yes thanks. Will fix in v2.

Hmm. And I just noticed that the subjects were corrupted... need to fix
my script.

> > +static bool process_has_mm(struct task_struct *p, struct mm_struct *mm)
> > +{
> > +	struct task_struct *t;
> > +
> > +	for_each_thread(p, t)
> > +		if (t->mm)
>
> Can t->mm change between pevious line and next line?

Good point, thanks. I'll add READ_ONCE() to ensure gcc won't re-load
t->mm again.

> > @@ -530,7 +541,7 @@ void oom_kill_process(struct oom_control *oc, struct task_struct *p,
> >  		list_for_each_entry(child, &t->children, sibling) {
> >  			unsigned int child_points;
> >
> > -			if (child->mm == p->mm)
> > +			if (process_has_mm(child, p->mm))
> >  				continue;
>
> We hold read_lock(&tasklist_lock) but not rcu_read_lock().
> Is for_each_thread() safe without rcu_read_lock()?

Yes, for_each_thread() is rcu and/or tasklist_lock safe.

Oleg.