From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754459AbbIRQ5l (ORCPT ); Fri, 18 Sep 2015 12:57:41 -0400 Received: from mx1.redhat.com ([209.132.183.28]:43741 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753171AbbIRQ5k (ORCPT ); Fri, 18 Sep 2015 12:57:40 -0400 Date: Fri, 18 Sep 2015 18:54:42 +0200 From: Oleg Nesterov To: Tetsuo Handa Cc: cl@linux.com, kwalker@redhat.com, akpm@linux-foundation.org, mhocko@suse.cz, rientjes@google.com, hannes@cmpxchg.org, vdavydov@parallels.com, linux-mm@kvack.org, linux-kernel@vger.kernel.org, skozina@redhat.com Subject: Re: [PATCH] mm/oom_kill.c: don't kill TASK_UNINTERRUPTIBLE tasks Message-ID: <20150918165441.GA20665@redhat.com> References: <1442512783-14719-1-git-send-email-kwalker@redhat.com> <20150917192204.GA2728@redhat.com> <20150918162423.GA18136@redhat.com> <201509190139.GJH48908.QMSFJLFtOHOVFO@I-love.SAKURA.ne.jp> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <201509190139.GJH48908.QMSFJLFtOHOVFO@I-love.SAKURA.ne.jp> User-Agent: Mutt/1.5.18 (2008-05-17) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 09/19, Tetsuo Handa wrote: > > Oleg Nesterov wrote: > > To simplify the discussion lets ignore PF_FROZEN, this is another issue. > > > > I am not sure this change is enough, we need to ensure that > > select_bad_process() won't pick the same task (or its sub-thread) again. > > SysRq-f is sometimes unusable because it continues choosing the same thread. > oom_kill_process() should not choose a thread which already has TIF_MEMDIE. So I was right, this is really not enough... > I think we need to rewrite oom_kill_process(). Heh. I can only ack the intent and wish you good luck ;) > > And perhaps something like > > > > wait_event_timeout(oom_victims_wait, !oom_victims, > > configurable_timeout); > > > > before select_bad_process() makes sense? > > I think you should not sleep for long with oom_lock mutex held. > http://marc.info/?l=linux-mm&m=143031212312459 Yes, yes, sure, I didn't mean we should wait under oom_lock. Oleg.