Date:	Wed, 29 Jul 2015 09:14:54 -0400
From:	Johannes Weiner <hannes@...xchg.org>
To:	Michal Hocko <mhocko@...nel.org>
Cc:	Tejun Heo <tj@...nel.org>,
	Andrew Morton <akpm@...ux-foundation.org>,
	Oleg Nesterov <oleg@...hat.com>,
	Vladimir Davydov <vdavydov@...allels.com>,
	Greg Thelen <gthelen@...gle.com>,
	KAMEZAWA Hiroyuki <kamezawa.hiroyu@...fujitsu.com>,
	KOSAKI Motohiro <kosaki.motohiro@...fujitsu.com>,
	linux-mm@...ck.org, LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 7/8] memcg: get rid of mm_struct::owner

On Tue, Jul 14, 2015 at 05:18:23PM +0200, Michal Hocko wrote:
> On Fri 10-07-15 16:05:33, Michal Hocko wrote:
> > JFYI: I've found some more issues while hamerring this more.
> 
> OK so the main issue is quite simple but I have completely missed it when
> thinking about the patch before. clone(CLONE_VM) without CLONE_THREAD is
> really nasty and it will easily lockup the machine with preempt. disabled
> for ever. It goes like this:
> taskA (in memcg A)
>   taskB = clone(CLONE_VM)
> 				taskB
> 				  A -> B	# Both tasks charge to B now
> 				  exit()	# No tasks in B -> can be
> 				  		# offlined now
> 				css_offline()
>   mem_cgroup_try_charge
>     get_mem_cgroup_from_mm
>       rcu_read_lock()
>       do {
>       } while css_tryget_online(mm->memcg)	# will never succeed
>       rcu_read_unlock()
> 
> taskA and taskB are basically independent entities wrt. the life
> cycle (unlike threads which are bound to the group leader). The
> previous code handles this by re-ownering during exit by the monster
> mm_update_next_owner.
>
> I can see the following options without reintroducing reintroducing
> some form of mm_update_next_owner:
> 
> 1) Do not allow offlining a cgroup if we have active users in it.  This
> would require a callback from the cgroup core to the subsystem called if
> there are no active tasks tracked by the cgroup core. Tracking on the memcg
> side doesn't sound terribly hard - just mark a mm_struct as an alien and
> count the number of aliens during the move in mem_cgroup. mm_drop_memcg
> then drops the counter. We could end up with EBUSY cgroup without any
> visible tasks which is a bit awkward.

You couldn't remove the group, and you wouldn't know which task needs
to move to get the mm out of there. That's no good.

> 2) update get_mem_cgroup_from_mm and others to fallback to the parent
> memcg if the current one is offline. This would be in line with charge
> reparenting we used to do. I cannot say I would like this because it
> allows for easy runaway to the root memcg if the hierarchy is not
> configured cautiously. The code would be also quite tricky because each
> direct consumer of mm->memcg would have to be aware of this. This is
> awkward.

In the unified hierarchy, there won't be tasks inside intermediate
nodes, so reparenting would lead to surprising behavior.

> 3) fail mem_cgroup_can_attach if we are trying to migrate a task sharing
> mm_struct with a process outside of the tset. If I understand the
> tset properly this would require all the sharing tasks to be migrated
> together and we would never end up with task_css != &task->mm->css.
> __cgroup_procs_write doesn't seem to support multi pid move currently
> AFAICS, though. cgroup_migrate_add_src, however, seems to be intended
> for this purpose so this should be doable. Without that support we would
> basically disallow migrating these tasks - I wouldn't object if you ask
> me.

I'd prefer not adding controller-specific failure modes for attaching,
and this too would lead to very non-obvious behavior.

> Do you see other options? From the above three options the 3rd one
> sounds the most sane to me and the 1st quite easy to implement. Both will
> require some cgroup core work though. But maybe we would be good enough
> with 3rd option without supporting moving schizophrenic tasks and that
> would be reduced to memcg code.

A modified form of 1) would be to track the mms referring to a memcg
but during offline search the process tree for a matching task. This
is heavy-handed, but it's a rare case and this work would be done in
the cgroup removal path rather than during task exit. This is stolen
from the current mm_update_next_owner():

list_for_each_entry(mm, memcg->mms, memcg_list) {
    for_each_process(g) {
        if (g->flags & PF_KTHREAD)
            continue;
        for_each_thread(g, c) {
            if (c->mm == mm)
                goto assign;
            if (c->mm)
                break;
        }
    }
assign:
    memcg = mem_cgroup_from_task(c);
    mm->memcg = memcg;
    list_move(&mm->memcg_list, &memcg->mms);
}

(plus appropriate synchronization, of course)
--
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/

Powered by blists - more mailing lists