---

yaml
---
r: 176268
b: refs/heads/master
c: f50de2d
h: refs/heads/master
v: v3

[refs]

@@ -1,2 +1,2 @@
 ---
-refs/heads/master: 273f047e36d83179573dc7e3a8af6aceaa8c599e
+refs/heads/master: f50de2d3811081957156b5d736778799379c29de

trunk/include/linux/vmstat.h

@@ -40,6 +40,7 @@ enum vm_event_item { PGPGIN, PGPGOUT, PSWPIN, PSWPOUT,
 		PGSCAN_ZONE_RECLAIM_FAILED,
 #endif
 		PGINODESTEAL, SLABS_SCANNED, KSWAPD_STEAL, KSWAPD_INODESTEAL,
+		KSWAPD_PREMATURE_FAST, KSWAPD_PREMATURE_SLOW,
 		PAGEOUTRUN, ALLOCSTALL, PGROTATED,
 #ifdef CONFIG_HUGETLB_PAGE
 		HTLB_BUDDY_PGALLOC, HTLB_BUDDY_PGALLOC_FAIL,

trunk/mm/vmscan.c

@@ -1904,6 +1904,24 @@ unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *mem_cont,
 }
 #endif
 
+/* is kswapd sleeping prematurely? */
+static int sleeping_prematurely(int order, long remaining)
+{
+	struct zone *zone;
+
+	/* If a direct reclaimer woke kswapd within HZ/10, it's premature */
+	if (remaining)
+		return 1;
+
+	/* If after HZ/10, a zone is below the high mark, it's premature */
+	for_each_populated_zone(zone)
+		if (!zone_watermark_ok(zone, order, high_wmark_pages(zone),
+								0, 0))
+			return 1;
+
+	return 0;
+}
+
 /*
  * For kswapd, balance_pgdat() will work across all this node's zones until
  * they are all at high_wmark_pages(zone).
@@ -2185,8 +2203,30 @@ static int kswapd(void *p)
 			 */
 			order = new_order;
 		} else {
-			if (!freezing(current) && !kthread_should_stop())
-				schedule();
+			if (!freezing(current) && !kthread_should_stop()) {
+				long remaining = 0;
+
+				/* Try to sleep for a short interval */
+				if (!sleeping_prematurely(order, remaining)) {
+					remaining = schedule_timeout(HZ/10);
+					finish_wait(&pgdat->kswapd_wait, &wait);
+					prepare_to_wait(&pgdat->kswapd_wait, &wait, TASK_INTERRUPTIBLE);
+				}
+
+				/*
+				 * After a short sleep, check if it was a
+				 * premature sleep. If not, then go fully
+				 * to sleep until explicitly woken up
+				 */
+				if (!sleeping_prematurely(order, remaining))
+					schedule();
+				else {
+					if (remaining)
+						count_vm_event(KSWAPD_PREMATURE_FAST);
+					else
+						count_vm_event(KSWAPD_PREMATURE_SLOW);
+				}
+			}
 
 			order = pgdat->kswapd_max_order;
 		}

trunk/mm/vmstat.c

@@ -683,6 +683,8 @@ static const char * const vmstat_text[] = {
 	"slabs_scanned",
 	"kswapd_steal",
 	"kswapd_inodesteal",
+	"kswapd_slept_prematurely_fast",
+	"kswapd_slept_prematurely_slow",
 	"pageoutrun",
 	"allocstall",