OpenVZ Forum: Devel » [PATCH 00/23] slab+slub accounting for memcg

Home » Mailing lists » Devel » [PATCH 00/23] slab+slub accounting for memcg

Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend

Re: [PATCH 17/23] kmem controller charge/uncharge infrastructure [message #46061 is a reply to message #46052]

Tue, 24 April 2012 20:21

David Rientjes
Messages: 59
Registered: November 2006

Member

On Tue, 24 Apr 2012, Frederic Weisbecker wrote:

> > This seems horribly inconsistent with memcg charging of user memory since
> > it charges to p->mm->owner and you're charging to p. So a thread attached
> > to a memcg can charge user memory to one memcg while charging slab to
> > another memcg?
>
> Charging to the thread rather than the process seem to me the right behaviour:
> you can have two threads of a same process attached to different cgroups.
>
> Perhaps it is the user memory memcg that needs to be fixed?
>

No, because memory is represented by mm_struct, not task_struct, so you
must charge to p->mm->owner to allow for moving threads amongst memcgs
later for memory.move_charge_at_immigrate. You shouldn't be able to
charge two different memcgs for memory represented by a single mm.

> > > +
> > > + if (!mem_cgroup_kmem_enabled(memcg))
> > > + goto out;
> > > +
> > > + mem_cgroup_get(memcg);
> > > + ret = memcg_charge_kmem(memcg, gfp, size) == 0;
> > > + if (ret)
> > > + mem_cgroup_put(memcg);
> > > +out:
> > > + rcu_read_unlock();
> > > + return ret;
> > > +}
> > > +EXPORT_SYMBOL(__mem_cgroup_charge_kmem);
> > > +
> > > +void __mem_cgroup_uncharge_kmem(size_t size)
> > > +{
> > > + struct mem_cgroup *memcg;
> > > +
> > > + rcu_read_lock();
> > > + memcg = mem_cgroup_from_task(current);
> > > +
> > > + if (!mem_cgroup_kmem_enabled(memcg))
> > > + goto out;
> > > +
> > > + mem_cgroup_put(memcg);
> > > + memcg_uncharge_kmem(memcg, size);
> > > +out:
> > > + rcu_read_unlock();
> > > +}
> > > +EXPORT_SYMBOL(__mem_cgroup_uncharge_kmem);

Report message to a moderator

[Message index]

		[PATCH 00/23] slab+slub accounting for memcg By: Glauber Costa on Fri, 20 April 2012 21:57
		[PATCH 02/23] slub: always get the cache from its page in kfree By: Glauber Costa on Fri, 20 April 2012 21:57
		[PATCH 04/23] memcg: Make it possible to use the stock for more than one page. By: Glauber Costa on Fri, 20 April 2012 21:57
		Re: [PATCH 04/23] memcg: Make it possible to use the stock for more than one page. By: KAMEZAWA Hiroyuki on Wed, 25 April 2012 00:59
		[PATCH 03/23] slab: rename gfpflags to allocflags By: Glauber Costa on Fri, 20 April 2012 21:57
		[PATCH 07/23] change defines to an enum By: Glauber Costa on Fri, 20 April 2012 21:57
		Re: [PATCH 07/23] change defines to an enum By: KAMEZAWA Hiroyuki on Wed, 25 April 2012 01:18
		[PATCH 05/23] memcg: Reclaim when more than one page needed. By: Glauber Costa on Fri, 20 April 2012 21:57
		Re: [PATCH 05/23] memcg: Reclaim when more than one page needed. By: KAMEZAWA Hiroyuki on Wed, 25 April 2012 01:16
		[PATCH 09/23] kmem slab accounting basic infrastructure By: Glauber Costa on Fri, 20 April 2012 21:57
		Re: [PATCH 09/23] kmem slab accounting basic infrastructure By: KAMEZAWA Hiroyuki on Wed, 25 April 2012 01:32
		Re: [PATCH 09/23] kmem slab accounting basic infrastructure By: Glauber Costa on Wed, 25 April 2012 14:38
		Re: [PATCH 09/23] kmem slab accounting basic infrastructure By: KAMEZAWA Hiroyuki on Thu, 26 April 2012 00:08
		Re: [PATCH 09/23] kmem slab accounting basic infrastructure By: Suleiman Souhlal on Mon, 30 April 2012 19:33
		Re: [PATCH 09/23] kmem slab accounting basic infrastructure By: Glauber Costa on Wed, 02 May 2012 15:15
		[PATCH 08/23] don't force return value checking in res_counter_charge_nofail By: Glauber Costa on Fri, 20 April 2012 21:57
		Re: [PATCH 08/23] don't force return value checking in res_counter_charge_nofail By: KAMEZAWA Hiroyuki on Wed, 25 April 2012 01:28
		[PATCH 06/23] slab: use obj_size field of struct kmem_cache when not debugging By: Glauber Costa on Fri, 20 April 2012 21:57
		[PATCH 01/23] slub: don't create a copy of the name string in kmem_cache_create By: Glauber Costa on Fri, 20 April 2012 21:57
		[PATCH 10/23] slab/slub: struct memcg_params By: Glauber Costa on Fri, 20 April 2012 21:57
		Re: [PATCH 10/23] slab/slub: struct memcg_params By: Suleiman Souhlal on Mon, 30 April 2012 19:42
		[PATCH 11/23] slub: consider a memcg parameter in kmem_create_cache By: Glauber Costa on Fri, 20 April 2012 21:57
		Re: [PATCH 11/23] slub: consider a memcg parameter in kmem_create_cache By: Frederic Weisbecker on Tue, 24 April 2012 14:03
		Re: [PATCH 11/23] slub: consider a memcg parameter in kmem_create_cache By: Glauber Costa on Tue, 24 April 2012 14:27
		Re: [PATCH 11/23] slub: consider a memcg parameter in kmem_create_cache By: KAMEZAWA Hiroyuki on Wed, 25 April 2012 01:38
		Re: [PATCH 11/23] slub: consider a memcg parameter in kmem_create_cache By: Glauber Costa on Wed, 25 April 2012 14:37
		Re: [PATCH 11/23] slub: consider a memcg parameter in kmem_create_cache By: Suleiman Souhlal on Mon, 30 April 2012 19:51
		Re: [PATCH 11/23] slub: consider a memcg parameter in kmem_create_cache By: Glauber Costa on Wed, 02 May 2012 15:18
		[PATCH 12/23] slab: pass memcg parameter to kmem_cache_create By: Glauber Costa on Sun, 22 April 2012 23:53
		Re: [PATCH 12/23] slab: pass memcg parameter to kmem_cache_create By: Suleiman Souhlal on Mon, 30 April 2012 19:54
		[PATCH 21/23] memcg: Track all the memcg children of a kmem_cache. By: Glauber Costa on Sun, 22 April 2012 23:53
		[PATCH 13/23] slub: create duplicate cache By: Glauber Costa on Sun, 22 April 2012 23:53
		Re: [PATCH 13/23] slub: create duplicate cache By: Frederic Weisbecker on Tue, 24 April 2012 14:18
		Re: [PATCH 13/23] slub: create duplicate cache By: Glauber Costa on Tue, 24 April 2012 14:37
		Re: [PATCH 13/23] slub: create duplicate cache By: Frederic Weisbecker on Thu, 26 April 2012 13:10
		Re: [PATCH 13/23] slub: create duplicate cache By: Suleiman Souhlal on Mon, 30 April 2012 20:15
		[PATCH 15/23] slab: create duplicate cache By: Glauber Costa on Sun, 22 April 2012 23:53
		[PATCH 14/23] slub: provide kmalloc_no_account By: Glauber Costa on Sun, 22 April 2012 23:53
		[PATCH 22/23] memcg: Per-memcg memory.kmem.slabinfo file. By: Glauber Costa on Sun, 22 April 2012 23:53
		[PATCH 16/23] slab: provide kmalloc_no_account By: Glauber Costa on Sun, 22 April 2012 23:53
		Re: [PATCH 16/23] slab: provide kmalloc_no_account By: KAMEZAWA Hiroyuki on Wed, 25 April 2012 01:44
		Re: [PATCH 16/23] slab: provide kmalloc_no_account By: Glauber Costa on Wed, 25 April 2012 14:29
		Re: [PATCH 16/23] slab: provide kmalloc_no_account By: KAMEZAWA Hiroyuki on Thu, 26 April 2012 00:13
		[PATCH 18/23] slub: charge allocation to a memcg By: Glauber Costa on Sun, 22 April 2012 23:53
		[PATCH 17/23] kmem controller charge/uncharge infrastructure By: Glauber Costa on Sun, 22 April 2012 23:53
		Re: [PATCH 17/23] kmem controller charge/uncharge infrastructure By: David Rientjes on Mon, 23 April 2012 22:25
		Re: [PATCH 17/23] kmem controller charge/uncharge infrastructure By: Frederic Weisbecker on Tue, 24 April 2012 14:22
		Re: [PATCH 17/23] kmem controller charge/uncharge infrastructure By: Glauber Costa on Tue, 24 April 2012 14:40
		Re: [PATCH 17/23] kmem controller charge/uncharge infrastructure By: David Rientjes on Tue, 24 April 2012 20:25
		Re: [PATCH 17/23] kmem controller charge/uncharge infrastructure By: Glauber Costa on Tue, 24 April 2012 21:36
		Re: [PATCH 17/23] kmem controller charge/uncharge infrastructure By: David Rientjes on Tue, 24 April 2012 22:54
		Re: [PATCH 17/23] kmem controller charge/uncharge infrastructure By: Glauber Costa on Wed, 25 April 2012 14:43
		Re: [PATCH 17/23] kmem controller charge/uncharge infrastructure By: David Rientjes on Tue, 24 April 2012 20:21
		Re: [PATCH 17/23] kmem controller charge/uncharge infrastructure By: Frederic Weisbecker on Fri, 27 April 2012 11:38
		Re: [PATCH 17/23] kmem controller charge/uncharge infrastructure By: David Rientjes on Fri, 27 April 2012 18:13
		Re: [PATCH 17/23] kmem controller charge/uncharge infrastructure By: KAMEZAWA Hiroyuki on Wed, 25 April 2012 01:56
		Re: [PATCH 17/23] kmem controller charge/uncharge infrastructure By: Glauber Costa on Wed, 25 April 2012 14:44
		Re: [PATCH 17/23] kmem controller charge/uncharge infrastructure By: Frederic Weisbecker on Fri, 27 April 2012 12:22
		Re: [PATCH 17/23] kmem controller charge/uncharge infrastructure By: Suleiman Souhlal on Mon, 30 April 2012 20:56
		Re: [PATCH 17/23] kmem controller charge/uncharge infrastructure By: Glauber Costa on Wed, 02 May 2012 15:34
		[PATCH 19/23] slab: per-memcg accounting of slab caches By: Glauber Costa on Sun, 22 April 2012 23:53
		Re: [PATCH 19/23] slab: per-memcg accounting of slab caches By: Suleiman Souhlal on Mon, 30 April 2012 21:25
		Re: [PATCH 19/23] slab: per-memcg accounting of slab caches By: Glauber Costa on Wed, 02 May 2012 15:40
		[PATCH 23/23] slub: create slabinfo file for memcg By: Glauber Costa on Sun, 22 April 2012 23:53
		[PATCH 20/23] memcg: disable kmem code when not in use. By: Glauber Costa on Sun, 22 April 2012 23:53
		Re: [PATCH 00/23] slab+slub accounting for memcg By: Glauber Costa on Sun, 22 April 2012 23:59
		[PATCH 0/3] A few fixes for '[PATCH 00/23] slab+slub accounting for memcg' series By: Anton Vorontsov on Mon, 30 April 2012 09:59
		[PATCH 1/3] slab: Proper off-slabs handling when duplicating caches By: Anton Vorontsov on Mon, 30 April 2012 10:01
		[PATCH 2/3] slab: Fix imbalanced rcu locking By: Anton Vorontsov on Mon, 30 April 2012 10:01
		[PATCH 3/3] slab: Get rid of mem_cgroup_put_kmem_cache() By: Anton Vorontsov on Mon, 30 April 2012 10:02

Previous Topic:	[PATCH 00/23] slab+slub accounting for memcg
Next Topic:	[PATCH v3 0/2] SUNRPC: separate per-net data creation from service

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Wed Jun 25 11:30:44 GMT 2025

Total time taken to generate the page: 0.03436 seconds