Re: 3.13-rc breaks MEMCG_SWAP

[Michal Hocko <mhocko@suse.cz>]

On Tue 17-12-13 07:29:26, Tejun Heo wrote:
> Hello, Li.
> 
> On Tue, Dec 17, 2013 at 11:13:39AM +0800, Li Zefan wrote:
> > diff --git a/kernel/cgroup.c b/kernel/cgroup.c
> > index c36d906..769b5bb 100644
> > --- a/kernel/cgroup.c
> > +++ b/kernel/cgroup.c
> > @@ -868,6 +868,15 @@ static void cgroup_diput(struct dentry *dentry, struct inode *inode)
> >  		struct cgroup *cgrp = dentry->d_fsdata;
> >  
> >  		BUG_ON(!(cgroup_is_dead(cgrp)));
> > +
> > +		/*
> > +		 * We should remove the cgroup object from idr before its
> > +		 * grace period starts, so we won't be looking up a cgroup
> > +		 * while the cgroup is being freed.
> > +		 */
> 
> Let's remove this comment and instead comment that this is to be made
> per-css.  I mixed up the lifetime rules of the cgroup and css and
> thought css_from_id() should fail once css is confirmed to be offline,
> so the above comment.  It looks like we'll eventually have to move
> cgrp->id to css->id (just simple per-ss idr) as the two objects'
> lifetime rules will be completely separate.  Other than that, looks
> good to me.

Yeah, please remove it. It made me think that idr_remove cannot be
pulled to later and that's why I ruled out css based solution from the
beginning.
-- 
Michal Hocko
SUSE Labs