Re: [PATCH 4/4] memcg: simplify LRU handling by new rule

[Michal Hocko]

On Wed 21-12-11 09:09:41, KAMEZAWA Hiroyuki wrote:
> On Tue, 20 Dec 2011 17:16:15 +0100
> Michal Hocko <mhocko@xxxxxxx> wrote:
> 
> > On Wed 14-12-11 16:52:26, KAMEZAWA Hiroyuki wrote:
> > > From: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>
> > > 
> > > Now, at LRU handling, memory cgroup needs to do complicated works
> > > to see valid pc->mem_cgroup, which may be overwritten.
> > > 
> > > This patch is for relaxing the protocol. This patch guarantees
> > >    - when pc->mem_cgroup is overwritten, page must not be on LRU.
> > 
> > How the patch guarantees that? I do not see any enforcement. In fact we
> > depend on the previous patches, don't we.
> > 
> 
> Ah, yes. We depends on previous patch series.
> 
> 
> > > 
> > > By this, LRU routine can believe pc->mem_cgroup and don't need to
> > > check bits on pc->flags. This new rule may adds small overheads to
> > > swapin. But in most case, lru handling gets faster.
> > > 
> > > After this patch, PCG_ACCT_LRU bit is obsolete and removed.
> > 
> > It makes things much more simpler. I just think it needs a better
> > description.
> > 
> 
> O.K.
> 
> 99% of memcg charging are done by following call path.
> 
>    - alloc_page() -> charge() -> map/enter radix-tree -> add to LRU.
> 
> We need some special case cares.
> 
>    - SwapCache - newly allocated/fully unmapped pages are added to LRU
>                  before charge.
>      => handled by previous patch.
>    - FUSE      - unused pages are reused.
>      => handled by previous patch.
> 
>    - move_account
>      => we do isolate_page().
> 
> Now, we can guarantee pc->mem_cgroup is set when page is not added to
> LRU or under zone->lru_lock + isolate from LRU.
> 
> I'll add some Documenation to...memcg_debug.txt

Yes, much better.
Btw. I forgot to add
Acked-by: Michal Hocko <mhocko@xxxxxxx>

Thanks
-- 
Michal Hocko
SUSE Labs
SUSE LINUX s.r.o.
Lihovarska 1060/12
190 00 Praha 9    
Czech Republic
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/