mm, vmscan: only wakeup kswapd once per node for the requested classzone
authorMel Gorman <mgorman@techsingularity.net>
Thu, 28 Jul 2016 22:46:26 +0000 (15:46 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Thu, 28 Jul 2016 23:07:41 +0000 (16:07 -0700)
kswapd is woken when zones are below the low watermark but the wakeup
decision is not taking the classzone into account.  Now that reclaim is
node-based, it is only required to wake kswapd once per node and only if
all zones are unbalanced for the requested classzone.

Note that one node might be checked multiple times if the zonelist is
ordered by node because there is no cheap way of tracking what nodes
have already been visited.  For zone-ordering, each node should be
checked only once.

Link: http://lkml.kernel.org/r/1467970510-21195-22-git-send-email-mgorman@techsingularity.net
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Hillf Danton <hillf.zj@alibaba-inc.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Rik van Riel <riel@surriel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/page_alloc.c
mm/vmscan.c

index 0f92e04b58dbd19081e7604a838f48a3d3a1f69c..a34d9fcf1339b64fa41fc7b9d2bd3859a1a76aee 100644 (file)
@@ -3367,10 +3367,14 @@ static void wake_all_kswapds(unsigned int order, const struct alloc_context *ac)
 {
        struct zoneref *z;
        struct zone *zone;
+       pg_data_t *last_pgdat = NULL;
 
        for_each_zone_zonelist_nodemask(zone, z, ac->zonelist,
-                                               ac->high_zoneidx, ac->nodemask)
-               wakeup_kswapd(zone, order, ac_classzone_idx(ac));
+                                       ac->high_zoneidx, ac->nodemask) {
+               if (last_pgdat != zone->zone_pgdat)
+                       wakeup_kswapd(zone, order, ac_classzone_idx(ac));
+               last_pgdat = zone->zone_pgdat;
+       }
 }
 
 static inline unsigned int
index 9b61a55b6e3802ac0308a224c9353ee3cd520852..31edd7776289a12a6c3952a4aae4b1a007d41eac 100644 (file)
@@ -3421,6 +3421,7 @@ kswapd_try_sleep:
 void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx)
 {
        pg_data_t *pgdat;
+       int z;
 
        if (!populated_zone(zone))
                return;
@@ -3432,8 +3433,16 @@ void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx)
        pgdat->kswapd_order = max(pgdat->kswapd_order, order);
        if (!waitqueue_active(&pgdat->kswapd_wait))
                return;
-       if (zone_balanced(zone, order, 0))
-               return;
+
+       /* Only wake kswapd if all zones are unbalanced */
+       for (z = 0; z <= classzone_idx; z++) {
+               zone = pgdat->node_zones + z;
+               if (!populated_zone(zone))
+                       continue;
+
+               if (zone_balanced(zone, order, classzone_idx))
+                       return;
+       }
 
        trace_mm_vmscan_wakeup_kswapd(pgdat->node_id, zone_idx(zone), order);
        wake_up_interruptible(&pgdat->kswapd_wait);