Re: [patch v2 3/6] memcg: rework mem_cgroup_iter to use cgroup iterators

[Glauber Costa]

On 11/28/2012 01:33 PM, Michal Hocko wrote:
> On Wed 28-11-12 13:23:57, Glauber Costa wrote:
>> On 11/28/2012 01:17 PM, Michal Hocko wrote:
>>> On Wed 28-11-12 17:47:59, KAMEZAWA Hiroyuki wrote:
>>>> (2012/11/27 3:47), Michal Hocko wrote:
>>> [...]
>>>>> +		/*
>>>>> +		 * Even if we found a group we have to make sure it is alive.
>>>>> +		 * css && !memcg means that the groups should be skipped and
>>>>> +		 * we should continue the tree walk.
>>>>> +		 * last_visited css is safe to use because it is protected by
>>>>> +		 * css_get and the tree walk is rcu safe.
>>>>> +		 */
>>>>> +		if (css == &root->css || (css && css_tryget(css)))
>>>>> +			memcg = mem_cgroup_from_css(css);
>>>>
>>>> Could you note that this iterator will never visit dangling(removed)
>>>> memcg, somewhere ?
>>>
>>> OK, I can add it to the function comment but the behavior hasn't changed
>>> so I wouldn't like to confuse anybody.
>>>
>>>> Hmm, I'm not sure but it may be trouble at shrkinking dangling
>>>> kmem_cache(slab).
>>>
>>> We do not shrink slab at all. 
>>
>> yet. However...
>>
>>> Those objects that are in a dead memcg
>>> wait for their owner tho release them which will make the dangling group
>>> eventually go away
>>>
>>>>
>>>> Costa, how do you think ?
>>>>
>>
>> In general, I particularly believe it is a good idea to skip dead memcgs
>> in the iterator. I don't anticipate any problems with shrinking at all.
> 
> We even cannot iterate dead ones because their cgroups are gone and so
> you do not have any way to iterate. So either make them alive by css_get
> or we cannot iterate them.
> 

We are in full agreement.


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