From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757703Ab3ANSmo (ORCPT ); Mon, 14 Jan 2013 13:42:44 -0500 Received: from mail-qa0-f50.google.com ([209.85.216.50]:37846 "EHLO mail-qa0-f50.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755896Ab3ANSmm (ORCPT ); Mon, 14 Jan 2013 13:42:42 -0500 Date: Mon, 14 Jan 2013 10:42:36 -0800 From: Tejun Heo To: Li Zefan Cc: Colin Cross , Mike Galbraith , LKML , Cgroups Subject: Re: [PATCH 1/2] cgroup: remove synchronize_rcu() from cgroup_attach_{task|proc}() Message-ID: <20130114184236.GF27942@mtj.dyndns.org> References: <50F3CE8E.5000402@huawei.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <50F3CE8E.5000402@huawei.com> 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 Hey, Li. On Mon, Jan 14, 2013 at 05:23:26PM +0800, Li Zefan wrote: > These 2 syncronize_rcu()s make attaching a task to a cgroup > quite slow, and it can't be ignored in some situations. > > A real case from Colin Cross: Android uses cgroups heavily to > manage thread priorities, putting threads in a background group > with reduced cpu.shares when they are not visible to the user, > and in a foreground group when they are. Some RPCs from foreground > threads to background threads will temporarily move the background > thread into the foreground group for the duration of the RPC. > This results in many calls to cgroup_attach_task. > > In cgroup_attach_task() it's task->cgroups that is protected by RCU, > and put_css_set() calls kfree_rcu() to free it. > > If we remove this synchronize_rcu(), there can be threads in RCU-read > sections accessing their old cgroup via current->cgroups with > concurrent rmdir operation, but this is safe. Yeah, these synchronize_rcu()s are utterly confused. synchornize_rcu() necessarily have to come between two operations to guarantee that the changes made before it are visible to all rcu readers before proceeding beyond it. Here, synchornize_rcu() are at the end of attach operations with nothing beyond it. Duh.... its only effect would be delaying completion of write(2) to sysfs tasks/procs files until all rcu readers see the change, which is utterly irrelevant. It doesn't mean anything. Applying to cgroup/for-3.9. Thanks. -- tejun