From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753421AbaBGNpg (ORCPT ); Fri, 7 Feb 2014 08:45:36 -0500 Received: from cantor2.suse.de ([195.135.220.15]:38883 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751950AbaBGNpe (ORCPT ); Fri, 7 Feb 2014 08:45:34 -0500 Date: Fri, 7 Feb 2014 14:45:33 +0100 From: Michal Hocko To: Hugh Dickins Cc: Tejun Heo , Filipe Brandenburger , Li Zefan , Andrew Morton , Johannes Weiner , Greg Thelen , Michel Lespinasse , Markus Blank-Burian , Shawn Bohrer , cgroups@vger.kernel.org, linux-kernel@vger.kernel.org, linux-mm@kvack.org Subject: Re: [PATCH] cgroup: use an ordered workqueue for cgroup destruction Message-ID: <20140207134533.GC5121@dhcp22.suse.cz> References: MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: 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 Thu 06-02-14 15:56:01, Hugh Dickins wrote: > Sometimes the cleanup after memcg hierarchy testing gets stuck in > mem_cgroup_reparent_charges(), unable to bring non-kmem usage down to 0. > > There may turn out to be several causes, but a major cause is this: the > workitem to offline parent can get run before workitem to offline child; > parent's mem_cgroup_reparent_charges() circles around waiting for the > child's pages to be reparented to its lrus, but it's holding cgroup_mutex > which prevents the child from reaching its mem_cgroup_reparent_charges(). > > Just use an ordered workqueue for cgroup_destroy_wq. Hmm, interesting. Markus has seen hangs even with mem_cgroup_css_offline and the referenced cgroup fixes, maybe this is the the right one finally. > Fixes: e5fca243abae ("cgroup: use a dedicated workqueue for cgroup destruction") > Suggested-by: Filipe Brandenburger > Signed-off-by: Hugh Dickins > Cc: stable@vger.kernel.org # 3.10+ Reviewed-by: Michal Hocko e5fca243abae was marked for 3.9 stable but I do not see it in the Greg's 3.9 stable branch so 3.10+ seems to be sufficient. > --- > > kernel/cgroup.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > --- 3.14-rc1/kernel/cgroup.c 2014-02-02 18:49:07.737302111 -0800 > +++ linux/kernel/cgroup.c 2014-02-06 15:20:35.548904965 -0800 > @@ -4845,12 +4845,12 @@ static int __init cgroup_wq_init(void) > /* > * There isn't much point in executing destruction path in > * parallel. Good chunk is serialized with cgroup_mutex anyway. > - * Use 1 for @max_active. > + * Must be ordered to make sure parent is offlined after children. > * > * We would prefer to do this in cgroup_init() above, but that > * is called before init_workqueues(): so leave this until after. > */ > - cgroup_destroy_wq = alloc_workqueue("cgroup_destroy", 0, 1); > + cgroup_destroy_wq = alloc_ordered_workqueue("cgroup_destroy", 0); > BUG_ON(!cgroup_destroy_wq); > > /* -- Michal Hocko SUSE Labs