Re: [PATCH -mmotm-2010-01-06-14-34] check high watermark aftershrink zone

[Minchan Kim]

On Tue, 2010-01-12 at 15:01 -0800, Andrew Morton wrote:
> On Fri, 8 Jan 2010 14:12:35 +0900
> Minchan Kim <minchan.kim@xxxxxxxxx> wrote:
> 
> > Kswapd check that zone have enough free by zone_water_mark.
> > If any zone doesn't have enough page, it set all_zones_ok to zero.
> > all_zone_ok makes kswapd retry not sleeping.
> > 
> > I think the watermark check before shrink zone is pointless.
> > Kswapd try to shrink zone then the check is meaningul.
> > 
> > This patch move the check after shrink zone.
> 
> The changelog is rather hard to understand.  I changed it to
> 
> : Kswapd checks that zone has sufficient pages free via zone_watermark_ok().
> : 
> : If any zone doesn't have enough pages, we set all_zones_ok to zero. 
> : !all_zone_ok makes kswapd retry rather than sleeping.
> : 
> : I think the watermark check before shrink_zone() is pointless.  Only after
> : kswapd has tried to shrink the zone is the check meaningful.
> : 
> : Move the check to after the call to shrink_zone().
> 

Thanks, Andrew. 

> > Signed-off-by: Minchan Kim <minchan.kim@xxxxxxxxx>
> > CC: KOSAKI Motohiro <kosaki.motohiro@xxxxxxxxxxxxxx>
> > CC: Mel Gorman <mel@xxxxxxxxx>
> > CC: Rik van Riel <riel@xxxxxxxxxx>
> > ---
> >  mm/vmscan.c |   21 +++++++++++----------
> >  1 files changed, 11 insertions(+), 10 deletions(-)
> > 
> > diff --git a/mm/vmscan.c b/mm/vmscan.c
> > index 885207a..b81adf8 100644
> > --- a/mm/vmscan.c
> > +++ b/mm/vmscan.c
> > @@ -2057,9 +2057,6 @@ loop_again:
> >  					priority != DEF_PRIORITY)
> >  				continue;
> >  
> > -			if (!zone_watermark_ok(zone, order,
> > -					high_wmark_pages(zone), end_zone, 0))
> > -				all_zones_ok = 0;
> 
> This will make kswapd stop doing reclaim if all zones have
> zone_is_all_unreclaimable():
> 
> 			if (zone_is_all_unreclaimable(zone))
> 				continue;
> 
> This seems bad.

Do you mean zone_is_all_unreclaimable in front of if (nr_slab ==0 && ..)?

                        reclaim_state->reclaimed_slab = 0;
                        nr_slab = shrink_slab(sc.nr_scanned, GFP_KERNEL,
                                                lru_pages);
                        sc.nr_reclaimed += reclaim_state->reclaimed_slab;
                        total_scanned += sc.nr_scanned;
                        if (zone_is_all_unreclaimable(zone)) <=== 
                                continue;


Actually I think the check is pointless, too. 
We set ZONE_ALL_UNRECLAIMABLE after the check and increase next zone in
loop. 
The check is a little bit effective in just case concurrent zone
reclaim. But if we remove the check, it's one more call
zone_watermark_ok and it's okay, I think. 

In addition, we check zone_is_all_unreclaimable in start in loop
following as. 

                for (i = 0; i <= end_zone; i++) {
                        struct zone *zone = pgdat->node_zones + i; 
                        int nr_slab;
                        int nid, zid; 

                        if (!populated_zone(zone))
                                continue;

                        if (zone_is_all_unreclaimable(zone) && <===
                                        priority != DEF_PRIORITY)
                                continue;

so the check in higher priority is effective if anyone doesn't free any
page. 


> 
> >  			temp_priority[i] = priority;
> >  			sc.nr_scanned = 0;
> >  			note_zone_scanning_priority(zone, priority);
> > @@ -2099,13 +2096,17 @@ loop_again:
> >  			    total_scanned > sc.nr_reclaimed + sc.nr_reclaimed / 2)
> >  				sc.may_writepage = 1;
> >  
> > -			/*
> > -			 * We are still under min water mark. it mean we have
> > -			 * GFP_ATOMIC allocation failure risk. Hurry up!
> > -			 */
> > -			if (!zone_watermark_ok(zone, order, min_wmark_pages(zone),
> > -					      end_zone, 0))
> > -				has_under_min_watermark_zone = 1;
> > +			if (!zone_watermark_ok(zone, order,
> > +					high_wmark_pages(zone), end_zone, 0)) {
> > +				all_zones_ok = 0;
> > +				/*
> > +				 * We are still under min water mark. it mean we have
> > +				 * GFP_ATOMIC allocation failure risk. Hurry up!
> > +				 */
> > +				if (!zone_watermark_ok(zone, order, min_wmark_pages(zone),
> > +						      end_zone, 0))
> > +					has_under_min_watermark_zone = 1;
> > +			}
> >  
> 
> The vmscan.c code makes an effort to look nice in an 80-col display.

Okay. I will keep in mind. 
-- 
Kind regards,
Minchan Kim

--
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/