Re: [PATCH v2 06/11] memcg: kmem controller infrastructure

On Wed 15-08-12 18:01:51, Glauber Costa wrote:

On 08/15/2012 05:09 PM, Michal Hocko wrote:

quoted

On Wed 15-08-12 13:42:24, Glauber Costa wrote:
[...]

quoted

quoted

quoted

+
+	ret = 0;
+
+	if (!memcg)
+		return ret;
+
+	_memcg = memcg;
+	ret = __mem_cgroup_try_charge(NULL, gfp, delta / PAGE_SIZE,
+	    &_memcg, may_oom);

This is really dangerous because atomic allocation which seem to be
possible could result in deadlocks because of the reclaim. 

Can you elaborate on how this would happen?

Say you have an atomic allocation and we hit the limit so we get either
to reclaim which can sleep or to oom which can sleep as well (depending
on the oom_control).

I see now, you seem to be right.

No I am not because it seems that I am really blind these days...
We were doing this in mem_cgroup_do_charge for ages:
	if (!(gfp_mask & __GFP_WAIT))
                return CHARGE_WOULDBLOCK;

/me goes to hide and get with further feedback with a clean head.

Sorry about that.
-- 
Michal Hocko
SUSE Labs