From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail202.messagelabs.com (mail202.messagelabs.com [216.82.254.227]) by kanga.kvack.org (Postfix) with ESMTP id 728766B01C7 for ; Fri, 26 Mar 2010 18:08:16 -0400 (EDT) Date: Fri, 26 Mar 2010 15:08:05 -0700 From: Andrew Morton Subject: Re: [PATCH] oom killer: break from infinite loop Message-Id: <20100326150805.f5853d1c.akpm@linux-foundation.org> In-Reply-To: <1269447905-5939-1-git-send-email-anfei.zhou@gmail.com> References: <1269447905-5939-1-git-send-email-anfei.zhou@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-linux-mm@kvack.org To: Anfei Zhou Cc: rientjes@google.com, kosaki.motohiro@jp.fujitsu.com, nishimura@mxp.nes.nec.co.jp, kamezawa.hiroyu@jp.fujitsu.com, linux-mm@kvack.org, linux-kernel@vger.kernel.org, Oleg Nesterov List-ID: On Thu, 25 Mar 2010 00:25:05 +0800 Anfei Zhou wrote: > In multi-threading environment, if the current task(A) have got > the mm->mmap_sem semaphore, and the thread(B) in the same process > is selected to be oom killed, because they shares the same semaphore, > thread B can not really be killed. So __alloc_pages_slowpath turns > to be a infinite loop. Here set all the threads in the group to > TIF_MEMDIE, it gets a chance to break and exit. > > Signed-off-by: Anfei Zhou > --- > mm/oom_kill.c | 4 ++++ > 1 files changed, 4 insertions(+), 0 deletions(-) > > diff --git a/mm/oom_kill.c b/mm/oom_kill.c > index 9b223af..aab9892 100644 > --- a/mm/oom_kill.c > +++ b/mm/oom_kill.c > @@ -381,6 +381,8 @@ static void dump_header(struct task_struct *p, gfp_t gfp_mask, int order, > */ > static void __oom_kill_task(struct task_struct *p, int verbose) > { > + struct task_struct *t; > + > if (is_global_init(p)) { > WARN_ON(1); > printk(KERN_WARNING "tried to kill init!\n"); > @@ -412,6 +414,8 @@ static void __oom_kill_task(struct task_struct *p, int verbose) > */ > p->rt.time_slice = HZ; > set_tsk_thread_flag(p, TIF_MEMDIE); > + for (t = next_thread(p); t != p; t = next_thread(t)) > + set_tsk_thread_flag(t, TIF_MEMDIE); > > force_sig(SIGKILL, p); Don't we need some sort of locking while walking that ring? Unintuitively it appears to be spin_lock_irq(&p->sighand->siglock). -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: email@kvack.org