[stable][BUGFIX][PATCH v3] memcg: avoid oom-killing innocent taskin case of use_hierarchy

[Daisuke Nishimura]

On Mon, 4 Jan 2010 14:28:19 -0800
Greg KH <greg@xxxxxxxxx> wrote:

> On Thu, Dec 17, 2009 at 09:47:24AM +0900, Daisuke Nishimura wrote:
> > Stable team.
> > 
> > Cay you pick this up for 2.6.32.y(and 2.6.31.y if it will be released) ?
> > 
> > This is a for-stable version of a bugfix patch that corresponds to the
> > upstream commmit d31f56dbf8bafaacb0c617f9a6f137498d5c7aed.
> 
> I've applied it to the .32-stable tree, but it does not apply to .31.
> Care to provide a version of the patch for that kernel if you want it
> applied there?
> 
hmm, strange. I can apply it onto 2.6.31.9. It might conflict with other patches
in 2.6.31.y queue ?
Anyway, I've attached the patch that is rebased on 2.6.31.9. Please tell me if you
have any problem with it.

v3: rebased on 2.6.31.9
===