vmscan: check high watermark after shrink zone
authorMinchan Kim <minchan.kim@gmail.com>
Fri, 5 Mar 2010 21:41:45 +0000 (13:41 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Sat, 6 Mar 2010 19:26:24 +0000 (11:26 -0800)
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().

[akpm@linux-foundation.org: fix comment, layout]
Signed-off-by: Minchan Kim <minchan.kim@gmail.com>
Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Rik van Riel <riel@redhat.com>
Reviewed-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/vmscan.c

index c26986c85ce01b1d15a8914065a829890054c210..96ebe6608d5ce536cd0b20ac41b32ccdcc16441f 100644 (file)
@@ -2060,9 +2060,6 @@ loop_again:
                                        priority != DEF_PRIORITY)
                                continue;
 
-                       if (!zone_watermark_ok(zone, order,
-                                       high_wmark_pages(zone), end_zone, 0))
-                               all_zones_ok = 0;
                        temp_priority[i] = priority;
                        sc.nr_scanned = 0;
                        note_zone_scanning_priority(zone, priority);
@@ -2102,13 +2099,18 @@ 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.  This
+                                * means that we have a 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 (all_zones_ok)