Re: [patch 10/11] mm: make per-memcg LRU lists exclusive

[Michal Hocko <mhocko@suse.cz>]

On Wed 21-09-11 17:47:45, Johannes Weiner wrote:
> On Wed, Sep 21, 2011 at 05:24:58PM +0200, Michal Hocko wrote:
> > On Mon 12-09-11 12:57:27, Johannes Weiner wrote:
[...]
> > > @@ -934,115 +954,123 @@ EXPORT_SYMBOL(mem_cgroup_count_vm_event);
> > >   * When moving account, the page is not on LRU. It's isolated.
> > >   */
> > >  
> > > -struct page *mem_cgroup_lru_to_page(struct zone *zone, struct mem_cgroup *mem,
> > > -				    enum lru_list lru)
> > > +/**
> > > + * mem_cgroup_lru_add_list - account for adding an lru page and return lruvec
> > > + * @zone: zone of the page
> > > + * @page: the page
> > > + * @lru: current lru
> > > + *
> > > + * This function accounts for @page being added to @lru, and returns
> > > + * the lruvec for the given @zone and the memcg @page is charged to.
> > > + *
> > > + * The callsite is then responsible for physically linking the page to
> > > + * the returned lruvec->lists[@lru].
> > > + */
> > > +struct lruvec *mem_cgroup_lru_add_list(struct zone *zone, struct page *page,
> > > +				       enum lru_list lru)
> > 
> > I know that names are alway tricky but what about mem_cgroup_acct_lru_add?
> > Analogously for mem_cgroup_lru_del_list, mem_cgroup_lru_del and
> > mem_cgroup_lru_move_lists.
> 
> Hmm, but it doesn't just lru-account, it also looks up the right
> lruvec for the caller to link the page to, so it's not necessarily an
> improvement, although I agree that the name could be better.

Sorry, I do not have any better idea. I would just like if the name
didn't suggest that we actually modify the list.

> 
> > > @@ -3615,11 +3593,11 @@ unsigned long mem_cgroup_soft_limit_reclaim(struct zone *zone, int order,
> > >  static int mem_cgroup_force_empty_list(struct mem_cgroup *memcg,
> > >  				int node, int zid, enum lru_list lru)
> > >  {
> > > -	struct zone *zone;
> > >  	struct mem_cgroup_per_zone *mz;
> > > -	struct page_cgroup *pc, *busy;
> > >  	unsigned long flags, loop;
> > >  	struct list_head *list;
> > > +	struct page *busy;
> > > +	struct zone *zone;
> > 
> > Any specific reason to move zone declaration down here? Not that it
> > matters much. Just curious.
> 
> I find this arrangement more readable, I believe Ingo Molnar called it
> the reverse christmas tree once :-).  Longest lines first, then sort
> lines of equal length alphabetically.
> 
> And since it was basically complete, except for @zone, I just HAD to!

:)

> 
> > > @@ -3639,16 +3618,16 @@ static int mem_cgroup_force_empty_list(struct mem_cgroup *memcg,
> > >  			spin_unlock_irqrestore(&zone->lru_lock, flags);
> > >  			break;
> > >  		}
> > > -		pc = list_entry(list->prev, struct page_cgroup, lru);
> > > -		if (busy == pc) {
> > > -			list_move(&pc->lru, list);
> > > +		page = list_entry(list->prev, struct page, lru);
> > > +		if (busy == page) {
> > > +			list_move(&page->lru, list);
> > >  			busy = NULL;
> > >  			spin_unlock_irqrestore(&zone->lru_lock, flags);
> > >  			continue;
> > >  		}
> > >  		spin_unlock_irqrestore(&zone->lru_lock, flags);
> > >  
> > > -		page = lookup_cgroup_page(pc);
> > > +		pc = lookup_page_cgroup(page);
> > 
> > lookup_page_cgroup might return NULL so we probably want BUG_ON(!pc)
> > here. We are not very consistent about checking the return value,
> > though.
> 
> I think this is a myth and we should remove all those checks.  How can
> pages circulate in userspace before they are fully onlined and their
> page_cgroup buddies allocated?  In this case: how would they have been
> charged in the first place and sit on a list without a list_head? :-)

Yes, that is right. This should never happen (last famous words). I can
imagine that a memory offlinening bug could cause issues.

Anyway the more appropriate way to handle that would BUG_ON directly in
lookup_page_cgroup.

-- 
Michal Hocko
SUSE Labs
SUSE LINUX s.r.o.
Lihovarska 1060/12
190 00 Praha 9    
Czech Republic