From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753483AbZCSJUn (ORCPT ); Thu, 19 Mar 2009 05:20:43 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1754367AbZCSJUf (ORCPT ); Thu, 19 Mar 2009 05:20:35 -0400 Received: from viefep11-int.chello.at ([62.179.121.31]:11191 "EHLO viefep11-int.chello.at" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754117AbZCSJUe (ORCPT ); Thu, 19 Mar 2009 05:20:34 -0400 X-SourceIP: 213.93.53.227 Subject: Re: [PATCH -tip] cpuacct: Make cpuacct hierarchy walk in cpuacct_charge() safe when rcupreempt is used. From: Peter Zijlstra To: bharata@linux.vnet.ibm.com Cc: Li Zefan , linux-kernel@vger.kernel.org, Dhaval Giani , Balbir Singh , Paul Menage , Ingo Molnar , KAMEZAWA Hiroyuki In-Reply-To: <20090317073649.GH3314@in.ibm.com> References: <20090317061754.GD3314@in.ibm.com> <49BF42FB.4030103@cn.fujitsu.com> <20090317073649.GH3314@in.ibm.com> Content-Type: text/plain Date: Thu, 19 Mar 2009 10:20:21 +0100 Message-Id: <1237454421.7867.27.camel@twins> Mime-Version: 1.0 X-Mailer: Evolution 2.24.3 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, 2009-03-17 at 13:06 +0530, Bharata B Rao wrote: > On Tue, Mar 17, 2009 at 02:28:11PM +0800, Li Zefan wrote: > > Bharata B Rao wrote: > > > cpuacct: Make cpuacct hierarchy walk in cpuacct_charge() safe when > > > rcupreempt is used. > > > > > > cpuacct_charge() obtains task's ca and does a hierarchy walk upwards. > > > This can race with the task's movement between cgroups. This race > > > can cause an access to freed ca pointer in cpuacct_charge(). This will not > > > > Actually it can also end up access invalid tsk->cgroups. ;) > > > > get tsk->cgroups (cg) > > (move tsk to another cgroup) or (tsk exiting) > > -> kfree(tsk->cgroups) > > get cg->subsys[..] > > Ok :) Here is the patch again with updated description. > > cpuacct: Make cpuacct hierarchy walk in cpuacct_charge() safe when > rcupreempt is used. > > cpuacct_charge() obtains task's ca and does a hierarchy walk upwards. > This can race with the task's movement between cgroups. This race > can cause an access to freed ca pointer in cpuacct_charge() or access > to invalid cgroups pointer of the task. This will not happen with rcu or > tree rcu as cpuacct_charge() is called with preemption disabled. However if > rcupreempt is used, the race is seen. Thanks to Li Zefan for explaining this. > > Fix this race by explicitly protecting ca and the hierarchy walk with > rcu_read_lock(). > > Signed-off-by: Bharata B Rao I would ditch the comment, it doesn't add anything. The simple rule is: if you want RCU-safe, use rcu_read_lock(). preempt/irq disable isn't sufficient -- hasn't been for a long long while. After that, Acked-by: Peter Zijlstra > --- > kernel/sched.c | 8 ++++++++ > 1 file changed, 8 insertions(+) > > --- a/kernel/sched.c > +++ b/kernel/sched.c > @@ -9891,6 +9891,13 @@ static void cpuacct_charge(struct task_s > return; > > cpu = task_cpu(tsk); > + > + /* > + * preemption is already disabled here, but to be safe with > + * rcupreempt, take rcu_read_lock(). This protects ca and > + * hence the hierarchy walk. > + */ > + rcu_read_lock(); > ca = task_ca(tsk); > > do { > @@ -9898,6 +9905,7 @@ static void cpuacct_charge(struct task_s > *cpuusage += cputime; > ca = ca->parent; > } while (ca); > + rcu_read_unlock(); > } > > struct cgroup_subsys cpuacct_subsys = {