Re: [PATCH] Prevent shrinking of active anon lru list in case of no swap space

[Minchan Kim <minchan.kim@gmail.com>]

On Thu, 14 May 2009 16:19:52 +0900 (JST)
KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> wrote:

> > 
> > Now shrink_active_list is called several places.
> > But if we don't have a swap space, we can't reclaim anon pages.
> > So, we don't need deactivating anon pages in anon lru list.
> > 
> > Signed-off-by: Minchan Kim <minchan.kim@gmail.com>
> > Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
> > Cc: Johannes Weiner <hannes@cmpxchg.org>
> > Cc: Rik van Riel <riel@redhat.com>
> 
> hm, The analysis seems right. but
> 
> In general branch should put into non frequency path.
> why caller modification as following patch is wrong?

Yes. You're right. Firstly I though that. 
But there are a few place where inactve_anon_is_low is called.
In addition to it, we always have to call with (nr_swap_pages <= 0) with it. 
So alone inactive_anon_is_low is lost function's meaning.

I think it's meaningless. 
But We can improve one function wrapper.

/* I can't be remind proper function name. Can you suggest me ? */
if (do_we_have_to_shrink_list(zone, &sc))


int do_we_have_to_shrink_list(...)
{
   if (inactive_anon_is_low(zone, &sc) && (nr_swap_pages <= 0))
   
}

What do you think ?

> 
>                         /*
>                          * Do some background aging of the anon list, to give
>                          * pages a chance to be referenced before reclaiming.
>                          */
> -                        if (inactive_anon_is_low(zone, &sc))
> +                        if (inactive_anon_is_low(zone, &sc) && (nr_swap_pages <= 0))
>                                 shrink_active_list(SWAP_CLUSTER_MAX, zone,
>                                                         &sc, priority, 0);
> 
> 
> 
> 
> > ---
> >  mm/vmscan.c |    6 ++++++
> >  1 files changed, 6 insertions(+), 0 deletions(-)
> > 
> > diff --git a/mm/vmscan.c b/mm/vmscan.c
> > index 2f9d555..e4d71f4 100644
> > --- a/mm/vmscan.c
> > +++ b/mm/vmscan.c
> > @@ -1238,6 +1238,12 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone,
> >  	enum lru_list lru;
> >  	struct zone_reclaim_stat *reclaim_stat = get_reclaim_stat(zone, sc);
> >  
> > +	/* 
> > +	 * we can't shrink anon list in case of no swap space.
> > +	 */
> > +	if (file == 0 && nr_swap_pages <= 0)
> > +		return;
> > +
> >
> >  	lru_add_drain();
> >  	spin_lock_irq(&zone->lru_lock);
> >  	pgmoved = sc->isolate_pages(nr_pages, &l_hold, &pgscanned, sc->order,
> > -- 
> > 1.5.4.3
> > 
> > 
> > -- 
> > Kinds Regards
> > Minchan Kim
> 
> 
> 


-- 
Kinds Regards
Minchan Kim

--
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>