On 06/03, Ben Blum wrote:
quoted text
>
> On Wed, Jun 02, 2010 at 10:58:55PM +0200, Oleg Nesterov wrote:
> > Hmm. The usage of ->thread_group in ->can_attach() methods doesn't
> > look safe to me... but currently bool threadgroup is always false.
>
> I recall putting a rcu_read_lock() around that part and being assured
> that made it safe. But I don't remember the details. Maybe taking
> tasklist_lock is necessary?

rcu_read_lock() is not enough, see another email I sent.

Once again.

	rcu_read_lock()
	list_for_each_rcu(tsk->thread_group)

assumes that at least tsk->thread_group->next can't point to nowhere,
this is not true. This memory can go away _before_ we take rcu lock.

Oleg.

--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/