Re: [patch v2] mm: memcontrol: convert reclaim iterator to simple css refcounting

[Michal Hocko <mhocko@suse.cz>]

On Wed 24-09-14 13:16:53, Johannes Weiner wrote:
> On Wed, Sep 24, 2014 at 06:47:39PM +0200, Michal Hocko wrote:
> > On Fri 19-09-14 17:28:43, Johannes Weiner wrote:
[...]
> > > -		memcg = __mem_cgroup_iter_next(root, last_visited);
> > > +		do {
> > > +			pos = ACCESS_ONCE(mz->reclaim_iter[priority]);
> > > +		} while (pos && !css_tryget(&pos->css));
> > 
> > This is a bit confusing. AFAIU css_tryget fails only when the current
> > ref count is zero already. When do we keep cached memcg with zero count
> > behind? We always do css_get after cmpxchg.
> > 
> > Hmm, there is a small window between cmpxchg and css_get when we store
> > the current memcg into the reclaim_iter[priority]. If the current memcg
> > is root then we do not take any css reference before cmpxchg and so it
> > might drop down to zero in the mean time so other CPU might see zero I
> > guess. But I do not see how css_get after cmpxchg on such css works.
> > I guess I should go and check the css reference counting again.
> 
> It's not about root or the newly stored memcg, it's that you might
> read the position right before it's replaced and css_put(), at which

OK, got it

	CPU0					CPU1
pos = reclaim_iter[priority]
					cmpxchg(reclaim_iter[priority], pos, memcg)
					css_put(pos)	# -> 0
css_tryget(pos)

Thanks!
-- 
Michal Hocko
SUSE Labs

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>