[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20100112150152.78604b78.akpm@linux-foundation.org>
Date: Tue, 12 Jan 2010 15:01:52 -0800
From: Andrew Morton <akpm@...ux-foundation.org>
To: Minchan Kim <minchan.kim@...il.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@...fujitsu.com>,
Mel Gorman <mel@....ul.ie>, Rik van Riel <riel@...hat.com>,
linux-mm <linux-mm@...ck.org>,
lkml <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH -mmotm-2010-01-06-14-34] check high watermark after
shrink zone
On Fri, 8 Jan 2010 14:12:35 +0900
Minchan Kim <minchan.kim@...il.com> 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().
> Signed-off-by: Minchan Kim <minchan.kim@...il.com>
> CC: KOSAKI Motohiro <kosaki.motohiro@...fujitsu.com>
> CC: Mel Gorman <mel@....ul.ie>
> CC: Rik van Riel <riel@...hat.com>
> ---
> 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.
> 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.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists