linux-kernel - Re: [PATCH][BUGFIX] cgroups: more safe tasklist locking in cgroup_attach

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20110815231942.GC29942@somewhere>
Date:	Tue, 16 Aug 2011 01:19:44 +0200
From:	Frederic Weisbecker <fweisbec@...il.com>
To:	Ben Blum <bblum@...rew.cmu.edu>
Cc:	Oleg Nesterov <oleg@...hat.com>, NeilBrown <neilb@...e.de>,
	paulmck@...ux.vnet.ibm.com, Paul Menage <menage@...gle.com>,
	Li Zefan <lizf@...fujitsu.com>,
	containers@...ts.linux-foundation.org,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	Andrew Morton <akpm@...ux-foundation.org>
Subject: Re: [PATCH][BUGFIX] cgroups: more safe tasklist locking in
 cgroup_attach_proc

On Mon, Aug 15, 2011 at 07:09:00PM -0400, Ben Blum wrote:
> On Mon, Aug 15, 2011 at 07:04:15PM -0400, Ben Blum wrote:
> > On Tue, Aug 16, 2011 at 12:50:06AM +0200, Frederic Weisbecker wrote:
> > > On Mon, Aug 15, 2011 at 08:49:57PM +0200, Oleg Nesterov wrote:
> > > > On 07/29, Ben Blum wrote:
> > > > >
> > > > > According to this thread - https://lkml.org/lkml/2011/7/27/243 - RCU is
> > > > > not sufficient to guarantee the tasklist is stable w.r.t. de_thread and
> > > > > exit. Taking tasklist_lock for reading, instead of rcu_read_lock,
> > > > > ensures proper exclusion.
> > > > 
> > > > Yes.
> > > > 
> > > > So far I still think we should fix while_each_thread() so that it works
> > > > under rcu_read_lock() "as exepected", I'll try to think more.
> > > > 
> > > > But whatever we do with while_each_thread(), this can't help
> > > > cgroup_attach_proc(), it needs the locking.
> > > > 
> > > > > -	rcu_read_lock();
> > > > > +	read_lock(&tasklist_lock);
> > > > >  	if (!thread_group_leader(leader)) {
> > > > 
> > > > Agreed, this should work.
> > > > 
> > > > But can't we avoid the global list? thread_group_leader() or not, we do
> > > > not really care. We only need to ensure we can safely find all threads.
> > > > 
> > > > How about the patch below?
> > > > 
> > > > 
> > > > With or without this/your patch this leader can die right after we
> > > > drop the lock. ss->can_attach(leader) and ss->attach(leader) look
> > > > suspicious. If a sub-thread execs, this task_struct has nothing to
> > > > do with the threadgroup.
> > > > 
> > > > 
> > > > 
> > > > Also. This is off-topic, but... Why cgroup_attach_proc() and
> > > > cgroup_attach_task() do ->attach_task() + cgroup_task_migrate()
> > > > in the different order? cgroup_attach_proc() looks wrong even
> > > > if currently doesn't matter.
> > > 
> > > Right. As we concluded in our off-list discussion, if there
> > > is no strong reason for that, I'm going to fix that in my task
> > > counter patchset because there it really matters. If we can't
> > > migrate the thread because it has already exited, we really
> > > don't want to call ->attach_task() but rather cancel_attach_task().
> > > 
> > > Thanks.
> > > 
> > 
> > Yes. Um, this must have been a mistake on my part. The lines of code
> > should be the other way around. It should be done in a separate bugfix
> > patch, though, so it goes through faster...
> > 
> > -- Ben
> > 
> 
> also, there is no cancel_attach_task, afaict. cancel_attach is only for
> memcg, for the whole operation at once - unless you are changing this,
> in which case feel free to modify on top of the patch I'm about to send
> out.

Yeah right, I forgot it's something I'm adding with the task counter subsystem.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/