| Re: [PATCH 00/23] slab+slub accounting for memcg [message #46203 is a reply to message #46185] |
Wed, 02 May 2012 15:14  |
Glauber Costa
Messages: 916
Registered: October 2011
|
Senior Member |
|
|
On 04/30/2012 06:43 PM, Suleiman Souhlal wrote:
>> I am leaving destruction of caches out of the series, although most
>> > of the infrastructure for that is here, since we did it in earlier
>> > series. This is basically because right now Kame is reworking it for
>> > user memcg, and I like the new proposed behavior a lot more. We all seemed
>> > to have agreed that reclaim is an interesting problem by itself, and
>> > is not included in this already too complicated series. Please note
>> > that this is still marked as experimental, so we have so room. A proper
>> > shrinker implementation is a hard requirement to take the kmem controller
>> > out of the experimental state.
> We will have to be careful for cache destruction.
> I found several races between allocation and destruction, in my patchset.
>
> I think we should consider doing the uncharging of kmem when
> destroying a memcg in mem_cgroup_destroy() instead of in
> pre_destroy(), because it's still possible that there are threads in
> the cgroup while pre_destroy() is being called (or for threads to be
> moved into the cgroup).
I found some problems here as well.
I am trying to work ontop of what Kamezawa posted for pre_destroy()
rework. I have one or two incorrect uncharging issues to solve, that's
actually what is holding me for posting a new version.
expected soon
|
|
|
|
OpenVZ Forum
Members
Search
Help
Register
Login
Home
