From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753249AbcAESFR (ORCPT ); Tue, 5 Jan 2016 13:05:17 -0500 Received: from mail-wm0-f51.google.com ([74.125.82.51]:33167 "EHLO mail-wm0-f51.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751842AbcAESFL (ORCPT ); Tue, 5 Jan 2016 13:05:11 -0500 Date: Tue, 5 Jan 2016 19:05:09 +0100 From: Michal Hocko To: Tetsuo Handa Cc: akpm@linux-foundation.org, mgorman@suse.de, rientjes@google.com, torvalds@linux-foundation.org, oleg@redhat.com, hughd@google.com, andrea@kernel.org, riel@redhat.com, linux-mm@kvack.org, linux-kernel@vger.kernel.org Subject: Re: [RFC][PATCH] sysrq: ensure manual invocation of the OOM killer under OOM livelock Message-ID: <20160105180507.GB23326@dhcp22.suse.cz> References: <201512301533.JDJ18237.QOFOMVSFtHOJLF@I-love.SAKURA.ne.jp> <20160105162246.GH15324@dhcp22.suse.cz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20160105162246.GH15324@dhcp22.suse.cz> User-Agent: Mutt/1.5.24 (2015-08-30) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue 05-01-16 17:22:46, Michal Hocko wrote: > On Wed 30-12-15 15:33:47, Tetsuo Handa wrote: [...] > > I wish for a kernel thread that does OOM-kill operation. > > Maybe we can change the OOM reaper kernel thread to do it. > > What do you think? > > I do no think a separate kernel thread would help much if the > allocations have to keep looping in the allocator. oom_reaper is a > separate kernel thread only due to locking required for the exit_mmap > path. Let me clarify what I've meant here. What you actually want is to do select_bad_process and oom_kill_process (including oom_reap_vmas) in the kernel thread context, right? That should be doable because we do not depend on the allocation context there. That would certainly save 1 kernel thread for the sysrq+f part but it would make the regular case more complicated AFAICS. We would have to handle queuing of the oom requests because multiple oom killers might be active in different allocation domains (cpusets, memcgs) so I am not so sure this would be a great win in the end. But I haven't tried to do it so I might be wrong and it will turn up being much more easier than I expect. -- Michal Hocko SUSE Labs