From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756957Ab0KPUlh (ORCPT ); Tue, 16 Nov 2010 15:41:37 -0500 Received: from smtp1.linux-foundation.org ([140.211.169.13]:55813 "EHLO smtp1.linux-foundation.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755199Ab0KPUlg (ORCPT ); Tue, 16 Nov 2010 15:41:36 -0500 Date: Tue, 16 Nov 2010 12:41:17 -0800 From: Andrew Morton To: Daisuke Nishimura Cc: LKML , linux-mm , Balbir Singh , KAMEZAWA Hiroyuki , stable@kernel.org Subject: Re: [BUGFIX] memcg: avoid deadlock between move charge and try_charge() Message-Id: <20101116124117.64608b66.akpm@linux-foundation.org> In-Reply-To: <20101116191748.d6645376.nishimura@mxp.nes.nec.co.jp> References: <20101116191748.d6645376.nishimura@mxp.nes.nec.co.jp> X-Mailer: Sylpheed 2.4.8 (GTK+ 2.12.9; x86_64-pc-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, 16 Nov 2010 19:17:48 +0900 Daisuke Nishimura wrote: > From: Daisuke Nishimura > > __mem_cgroup_try_charge() can be called under down_write(&mmap_sem)(e.g. > mlock does it). This means it can cause deadlock if it races with move charge: > > Ex.1) > move charge | try charge > --------------------------------------+------------------------------ > mem_cgroup_can_attach() | down_write(&mmap_sem) > mc.moving_task = current | .. > mem_cgroup_precharge_mc() | __mem_cgroup_try_charge() > mem_cgroup_count_precharge() | prepare_to_wait() > down_read(&mmap_sem) | if (mc.moving_task) > -> cannot aquire the lock | -> true > | schedule() > > Ex.2) > move charge | try charge > --------------------------------------+------------------------------ > mem_cgroup_can_attach() | > mc.moving_task = current | > mem_cgroup_precharge_mc() | > mem_cgroup_count_precharge() | > down_read(&mmap_sem) | > .. | > up_read(&mmap_sem) | > | down_write(&mmap_sem) > mem_cgroup_move_task() | .. > mem_cgroup_move_charge() | __mem_cgroup_try_charge() > down_read(&mmap_sem) | prepare_to_wait() > -> cannot aquire the lock | if (mc.moving_task) > | -> true > | schedule() > > To avoid this deadlock, we do all the move charge works (both can_attach() and > attach()) under one mmap_sem section. > And after this patch, we set/clear mc.moving_task outside mc.lock, because we > use the lock only to check mc.from/to. > > Signed-off-by: Daisuke Nishimura I put this in the send-to-Linus-in-about-a-week queue. > Cc: The patch doesn't apply well to 2.6.36 so if we do want it backported then please prepare a tested backport for the -stable guys? Thanks.