mm: fix check for reclaimable pages in PF_MEMALLOC reclaim throttling
authorJohannes Weiner <hannes@cmpxchg.org>
Wed, 3 May 2017 21:51:54 +0000 (14:51 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Wed, 3 May 2017 22:52:07 +0000 (15:52 -0700)
PF_MEMALLOC direct reclaimers get throttled on a node when the sum of
all free pages in each zone fall below half the min watermark.  During
the summation, we want to exclude zones that don't have reclaimables.
Checking the same pgdat over and over again doesn't make sense.

Fixes: 599d0c954f91 ("mm, vmscan: move LRU lists to node")
Link: http://lkml.kernel.org/r/20170228214007.5621-3-hannes@cmpxchg.org
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Hillf Danton <hillf.zj@alibaba-inc.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Jia He <hejianet@gmail.com>
Cc: Mel Gorman <mgorman@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/vmscan.c

index 667644e..52832be 100644 (file)
@@ -2835,8 +2835,10 @@ static bool allow_direct_reclaim(pg_data_t *pgdat)
 
        for (i = 0; i <= ZONE_NORMAL; i++) {
                zone = &pgdat->node_zones[i];
-               if (!managed_zone(zone) ||
-                   pgdat_reclaimable_pages(pgdat) == 0)
+               if (!managed_zone(zone))
+                       continue;
+
+               if (!zone_reclaimable_pages(zone))
                        continue;
 
                pfmemalloc_reserve += min_wmark_pages(zone);