From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S933321Ab3BLQNg (ORCPT ); Tue, 12 Feb 2013 11:13:36 -0500 Received: from cantor2.suse.de ([195.135.220.15]:59079 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932173Ab3BLQNf (ORCPT ); Tue, 12 Feb 2013 11:13:35 -0500 Date: Tue, 12 Feb 2013 17:13:32 +0100 From: Michal Hocko To: Johannes Weiner Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org, KAMEZAWA Hiroyuki , Ying Han , Tejun Heo , Glauber Costa , Li Zefan Subject: Re: [PATCH v3 4/7] memcg: remove memcg from the reclaim iterators Message-ID: <20130212161332.GI4863@dhcp22.suse.cz> References: <20130208193318.GA15951@cmpxchg.org> <20130211151649.GD19922@dhcp22.suse.cz> <20130211175619.GC13218@cmpxchg.org> <20130211192929.GB29000@dhcp22.suse.cz> <20130211195824.GB15951@cmpxchg.org> <20130211212756.GC29000@dhcp22.suse.cz> <20130211223943.GC15951@cmpxchg.org> <20130212095419.GB4863@dhcp22.suse.cz> <20130212151002.GD15951@cmpxchg.org> <20130212154330.GG4863@dhcp22.suse.cz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20130212154330.GG4863@dhcp22.suse.cz> User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue 12-02-13 16:43:30, Michal Hocko wrote: [...] The example was not complete: > Wait a moment. But what prevents from the following race? > > rcu_read_lock() cgroup_next_descendant_pre css_tryget(css); memcg = mem_cgroup_from_css(css) atomic_add(CSS_DEACT_BIAS, &css->refcnt) > mem_cgroup_css_offline(memcg) We should be safe if we did synchronize_rcu() before root->dead_count++, no? Because then we would have a guarantee that if css_tryget(memcg) suceeded then we wouldn't race with dead_count++ it triggered. > root->dead_count++ > iter->last_dead_count = root->dead_count > iter->last_visited = memcg > // final > css_put(memcg); > // last_visited is still valid > rcu_read_unlock() > [...] > // next iteration > rcu_read_lock() > iter->last_dead_count == root->dead_count > // KABOOM > > The race window between dead_count++ and css_put is quite big but that > is not important because that css_put can happen anytime before we start > the next iteration and take rcu_read_lock. -- Michal Hocko SUSE Labs From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from psmtp.com (na3sys010amx165.postini.com [74.125.245.165]) by kanga.kvack.org (Postfix) with SMTP id 055F86B0002 for ; Tue, 12 Feb 2013 11:13:35 -0500 (EST) Date: Tue, 12 Feb 2013 17:13:32 +0100 From: Michal Hocko Subject: Re: [PATCH v3 4/7] memcg: remove memcg from the reclaim iterators Message-ID: <20130212161332.GI4863@dhcp22.suse.cz> References: <20130208193318.GA15951@cmpxchg.org> <20130211151649.GD19922@dhcp22.suse.cz> <20130211175619.GC13218@cmpxchg.org> <20130211192929.GB29000@dhcp22.suse.cz> <20130211195824.GB15951@cmpxchg.org> <20130211212756.GC29000@dhcp22.suse.cz> <20130211223943.GC15951@cmpxchg.org> <20130212095419.GB4863@dhcp22.suse.cz> <20130212151002.GD15951@cmpxchg.org> <20130212154330.GG4863@dhcp22.suse.cz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20130212154330.GG4863@dhcp22.suse.cz> Sender: owner-linux-mm@kvack.org List-ID: To: Johannes Weiner Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org, KAMEZAWA Hiroyuki , Ying Han , Tejun Heo , Glauber Costa , Li Zefan On Tue 12-02-13 16:43:30, Michal Hocko wrote: [...] The example was not complete: > Wait a moment. But what prevents from the following race? > > rcu_read_lock() cgroup_next_descendant_pre css_tryget(css); memcg = mem_cgroup_from_css(css) atomic_add(CSS_DEACT_BIAS, &css->refcnt) > mem_cgroup_css_offline(memcg) We should be safe if we did synchronize_rcu() before root->dead_count++, no? Because then we would have a guarantee that if css_tryget(memcg) suceeded then we wouldn't race with dead_count++ it triggered. > root->dead_count++ > iter->last_dead_count = root->dead_count > iter->last_visited = memcg > // final > css_put(memcg); > // last_visited is still valid > rcu_read_unlock() > [...] > // next iteration > rcu_read_lock() > iter->last_dead_count == root->dead_count > // KABOOM > > The race window between dead_count++ and css_put is quite big but that > is not important because that css_put can happen anytime before we start > the next iteration and take rcu_read_lock. -- Michal Hocko SUSE Labs -- 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