Hi,
On Fri, 26 Feb 2010 13:50:04 +0900
Minchan Kim <minchan.kim@xxxxxxxxx> wrote:
> > Hm ? I don't read the whole thread but can_attach() is called under
> > cgroup_mutex(). So, it doesn't need to use RCU.
>
> Vivek mentioned memcg is protected by RCU if I understand his intention right.
> So I commented that without enough knowledge of memcg.
> After your comment, I dive into the code.
>
> Just out of curiosity.
>
> Really, memcg is protected by RCU?
yes. All cgroup subsystem is protected by RCU.
> I think most of RCU around memcg is for protecting task_struct and
> cgroup_subsys_state.
> The memcg is protected by cgroup_mutex as you mentioned.
> Am I missing something?
There are several levels of protections.
cgroup subsystem's ->destroy() call back is finally called by
As this.
768 synchronize_rcu();
769
770 mutex_lock(&cgroup_mutex);
771 /*
772 * Release the subsystem state objects.
773 */
774 for_each_subsys(cgrp->root, ss)
775 ss->destroy(ss, cgrp);
776
777 cgrp->root->number_of_cgroups--;
778 mutex_unlock(&cgroup_mutex);
Before here,
- there are no tasks under this cgroup (cgroup's refcnt is 0)
&& cgroup is marked as REMOVED.
Then, this access
rcu_read_lock();
mem = mem_cgroup_from_task(task);
if (css_tryget(mem->css)) <===============checks cgroup refcnt
....
rcu_read_unlock()
is O.K.
And, it's graranteed that we don't have to do complicated fine-grain check
if cgroup_mutex() is held.
Because cgroup_mutex() is system-wide heavy lock, this refcnt+RCU trick is
used and works quite well.
Thanks,
-Kame
_______________________________________________
Containers mailing list
Containers@xxxxxxxxxxxxxxxxxxxxxxxxxx
https://lists.linux-foundation.org/mailman/listinfo/containers
