OpenVZ Forum: Devel » [PATCH v5 00/14] kmem controller for memcg.

Home » Mailing lists » Devel » [PATCH v5 00/14] kmem controller for memcg.

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

Re: [PATCH v5 04/14] kmem accounting basic infrastructure [message #48498 is a reply to message #48435]

Thu, 18 October 2012 19:37

Tejun Heo
Messages: 184
Registered: November 2006

Senior Member

Hello, David.

On Wed, Oct 17, 2012 at 03:08:04PM -0700, David Rientjes wrote:
> > +static int memcg_update_kmem_limit(struct cgroup *cont, u64 val)
> > +{
> > + int ret = -EINVAL;
> > +#ifdef CONFIG_MEMCG_KMEM
> > + struct mem_cgroup *memcg = mem_cgroup_from_cont(cont);
> > + /*
> > + * For simplicity, we won't allow this to be disabled. It also can't
> > + * be changed if the cgroup has children already, or if tasks had
> > + * already joined.
> > + *
> > + * If tasks join before we set the limit, a person looking at
> > + * kmem.usage_in_bytes will have no way to determine when it took
> > + * place, which makes the value quite meaningless.
> > + *
> > + * After it first became limited, changes in the value of the limit are
> > + * of course permitted.
> > + *
> > + * Taking the cgroup_lock is really offensive, but it is so far the only
> > + * way to guarantee that no children will appear. There are plenty of
> > + * other offenders, and they should all go away. Fine grained locking
> > + * is probably the way to go here. When we are fully hierarchical, we
> > + * can also get rid of the use_hierarchy check.
>
> Not sure it's so offensive, it's a pretty standard way of ensuring that
> cont->children doesn't get manipulated in a race.

cgroup_lock is inherently one of the outermost locks as it protects
cgroup hierarchy and modifying cgroup hierarchy involves invoking
subsystem callbacks which may grab subsystem locks. Grabbing it
directly from subsystems thus creates high likelihood of creating a
dependency loop and it's nasty to break.

And I'm unsure whether making cgroup locks finer grained would help as
cpuset grabs cgroup_lock to perform actual task migration which would
require the outermost cgroup locking anyway. This one already has
showed up in a couple lockdep warnings involving static_key usages.

A couple days ago, I posted a patchset to remove cgroup_lock usage
from cgroup_freezer and at least cgroup_freezer seems like it was
aiming for the wrong behavior which led to the wrong locking behavior
requiring grabbing cgroup_lock. I can't say whether others will be
that easy tho.

Anyways, so, cgroup_lock is in the process of being unexported and I'd
really like another usage isn't added but maybe that requires larger
changes to memcg and not something which can be achieved here. Dunno.
Will think more about it.

Thanks.

--
tejun

Report message to a moderator

[Message index]

		[PATCH v5 00/14] kmem controller for memcg. By: Glauber Costa on Tue, 16 October 2012 10:16
		[PATCH v5 03/14] memcg: change defines to an enum By: Glauber Costa on Tue, 16 October 2012 10:16
		Re: [PATCH v5 03/14] memcg: change defines to an enum By: David Rientjes on Wed, 17 October 2012 21:50
		[PATCH v5 01/14] memcg: Make it possible to use the stock for more than one page. By: Glauber Costa on Tue, 16 October 2012 10:16
		Re: [PATCH v5 01/14] memcg: Make it possible to use the stock for more than one page. By: akpm on Wed, 17 October 2012 22:11
		Re: [PATCH v5 01/14] memcg: Make it possible to use the stock for more than one page. By: Glauber Costa on Thu, 18 October 2012 08:55
		[PATCH v5 04/14] kmem accounting basic infrastructure By: Glauber Costa on Tue, 16 October 2012 10:16
		Re: [PATCH v5 04/14] kmem accounting basic infrastructure By: Michal Hocko on Tue, 16 October 2012 12:14
		Re: [PATCH v5 04/14] kmem accounting basic infrastructure By: David Rientjes on Wed, 17 October 2012 22:08
		Re: [PATCH v5 04/14] kmem accounting basic infrastructure By: Glauber Costa on Thu, 18 October 2012 09:02
		Re: [PATCH v5 04/14] kmem accounting basic infrastructure By: Tejun Heo on Thu, 18 October 2012 19:47
		Re: [PATCH v5 04/14] kmem accounting basic infrastructure By: Tejun Heo on Thu, 18 October 2012 19:37
		Re: [PATCH v5 04/14] kmem accounting basic infrastructure By: akpm on Wed, 17 October 2012 22:12
		Re: [PATCH v5 04/14] kmem accounting basic infrastructure By: Glauber Costa on Thu, 18 October 2012 09:04
		[PATCH v5 11/14] memcg: allow a memcg with kmem charges to be destructed. By: Glauber Costa on Tue, 16 October 2012 10:16
		Re: [PATCH v5 11/14] memcg: allow a memcg with kmem charges to be destructed. By: akpm on Wed, 17 October 2012 22:12
		Re: [PATCH v5 11/14] memcg: allow a memcg with kmem charges to be destructed. By: Glauber Costa on Thu, 18 October 2012 09:33
		[PATCH v5 06/14] memcg: kmem controller infrastructure By: Glauber Costa on Tue, 16 October 2012 10:16
		Re: [PATCH v5 06/14] memcg: kmem controller infrastructure By: KAMEZAWA Hiroyuki on Wed, 17 October 2012 06:40
		Re: [PATCH v5 06/14] memcg: kmem controller infrastructure By: akpm on Wed, 17 October 2012 22:12
		Re: [PATCH v5 06/14] memcg: kmem controller infrastructure By: Glauber Costa on Thu, 18 October 2012 09:16
		Re: [PATCH v5 06/14] memcg: kmem controller infrastructure By: David Rientjes on Thu, 18 October 2012 22:06
		Re: [PATCH v5 06/14] memcg: kmem controller infrastructure By: Glauber Costa on Fri, 19 October 2012 09:10
		Re: [PATCH v5 06/14] memcg: kmem controller infrastructure By: David Rientjes on Fri, 19 October 2012 09:31
		Re: [PATCH v5 06/14] memcg: kmem controller infrastructure By: Glauber Costa on Fri, 19 October 2012 10:00
		Re: [PATCH v5 06/14] memcg: kmem controller infrastructure By: David Rientjes on Wed, 17 October 2012 22:37
		Re: [PATCH v5 06/14] memcg: kmem controller infrastructure By: Glauber Costa on Thu, 18 October 2012 09:23
		Re: [PATCH v5 06/14] memcg: kmem controller infrastructure By: David Rientjes on Thu, 18 October 2012 21:59
		Re: [PATCH v5 06/14] memcg: kmem controller infrastructure By: Glauber Costa on Fri, 19 October 2012 10:08
		Re: [PATCH v5 06/14] memcg: kmem controller infrastructure By: David Rientjes on Fri, 19 October 2012 20:34
		Re: [PATCH v5 06/14] memcg: kmem controller infrastructure By: Glauber Costa on Mon, 22 October 2012 12:34
		Re: [PATCH v5 06/14] memcg: kmem controller infrastructure By: Michal Hocko on Mon, 22 October 2012 12:51
		Re: [PATCH v5 06/14] memcg: kmem controller infrastructure By: Glauber Costa on Mon, 22 October 2012 12:52
		[PATCH v5 05/14] Add a __GFP_KMEMCG flag By: Glauber Costa on Tue, 16 October 2012 10:16
		Re: [PATCH v5 05/14] Add a __GFP_KMEMCG flag By: Michal Hocko on Tue, 16 October 2012 12:15
		Re: [PATCH v5 05/14] Add a __GFP_KMEMCG flag By: David Rientjes on Wed, 17 October 2012 22:09
		[PATCH v5 09/14] memcg: kmem accounting lifecycle management By: Glauber Costa on Tue, 16 October 2012 10:16
		Re: [PATCH v5 09/14] memcg: kmem accounting lifecycle management By: David Rientjes on Wed, 17 October 2012 23:28
		Re: [PATCH v5 09/14] memcg: kmem accounting lifecycle management By: Michal Hocko on Thu, 18 October 2012 06:14
		Re: [PATCH v5 09/14] memcg: kmem accounting lifecycle management By: Glauber Costa on Thu, 18 October 2012 09:42
		[PATCH v5 10/14] memcg: use static branches when code not in use By: Glauber Costa on Tue, 16 October 2012 10:16
		[PATCH v5 14/14] Add documentation about the kmem controller By: Glauber Costa on Tue, 16 October 2012 10:16
		Re: [PATCH v5 14/14] Add documentation about the kmem controller By: Michal Hocko on Tue, 16 October 2012 12:23
		Re: [PATCH v5 14/14] Add documentation about the kmem controller By: Christoph Lameter on Tue, 16 October 2012 18:25
		Re: [PATCH v5 14/14] Add documentation about the kmem controller By: Glauber Costa on Tue, 16 October 2012 19:02
		Re: [PATCH v5 14/14] Add documentation about the kmem controller By: Christoph Lameter on Tue, 16 October 2012 19:30
		Re: [PATCH v5 14/14] Add documentation about the kmem controller By: akpm on Wed, 17 October 2012 22:12
		Re: [PATCH v5 14/14] Add documentation about the kmem controller By: Glauber Costa on Thu, 18 October 2012 09:38
		[PATCH v5 13/14] protect architectures where THREAD_SIZE >= PAGE_SIZE against fork bombs By: Glauber Costa on Tue, 16 October 2012 10:16
		Re: [PATCH v5 13/14] protect architectures where THREAD_SIZE >= PAGE_SIZE against fork bombs By: akpm on Wed, 17 October 2012 22:12
		Re: [PATCH v5 13/14] protect architectures where THREAD_SIZE >= PAGE_SIZE against fork bombs By: Glauber Costa on Thu, 18 October 2012 09:37
		[PATCH v5 07/14] mm: Allocate kernel pages to the right memcg By: Glauber Costa on Tue, 16 October 2012 10:16
		Re: [PATCH v5 07/14] mm: Allocate kernel pages to the right memcg By: Christoph Lameter on Tue, 16 October 2012 15:31
		Re: [PATCH v5 07/14] mm: Allocate kernel pages to the right memcg By: Glauber Costa on Tue, 16 October 2012 18:55
		Re: [PATCH v5 07/14] mm: Allocate kernel pages to the right memcg By: akpm on Wed, 17 October 2012 22:12
		Re: [PATCH v5 07/14] mm: Allocate kernel pages to the right memcg By: Glauber Costa on Thu, 18 October 2012 09:24
		Re: [PATCH v5 07/14] mm: Allocate kernel pages to the right memcg By: akpm on Thu, 18 October 2012 20:44
		Re: [PATCH v5 07/14] mm: Allocate kernel pages to the right memcg By: Glauber Costa on Thu, 18 October 2012 11:53
		Re: [PATCH v5 07/14] mm: Allocate kernel pages to the right memcg By: David Rientjes on Wed, 17 October 2012 22:43
		Re: [PATCH v5 00/14] kmem controller for memcg. By: akpm on Wed, 17 October 2012 22:11
		Re: [PATCH v5 00/14] kmem controller for memcg. By: Glauber Costa on Thu, 18 October 2012 08:52
		Re: [PATCH v5 00/14] kmem controller for memcg. By: akpm on Thu, 18 October 2012 19:21
		Re: [PATCH v5 00/14] kmem controller for memcg. By: Glauber Costa on Fri, 19 October 2012 09:55

Previous Topic:	[PATCH RFC] sched: boost throttled entities on wakeups
Next Topic:	[PATCH v6 01/10] ipc: remove forced assignment of selected message

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Mon Aug 25 15:24:10 GMT 2025

Total time taken to generate the page: 0.11455 seconds