[Devel,rh7,1/2] ms/mm: vmscan: reduce size of inactive file list

Submitted by Andrey Ryabinin on July 12, 2017, 3:01 p.m.

Details

Message ID 20170712150125.16276-1-aryabinin@virtuozzo.com
State New
Series "Series without cover letter"
Headers show

Commit Message

Andrey Ryabinin July 12, 2017, 3:01 p.m.
From: Rik van Riel <riel@redhat.com>

The inactive file list should still be large enough to contain readahead
windows and freshly written file data, but it no longer is the only
source for detecting multiple accesses to file pages.  The workingset
refault measurement code causes recently evicted file pages that get
accessed again after a shorter interval to be promoted directly to the
active list.

With that mechanism in place, we can afford to (on a larger system)
dedicate more memory to the active file list, so we can actually cache
more of the frequently used file pages in memory, and not have them
pushed out by streaming writes, once-used streaming file reads, etc.

This can help things like database workloads, where only half the page
cache can currently be used to cache the database working set.  This
patch automatically increases that fraction on larger systems, using the
same ratio that has already been used for anonymous memory.

[hannes@cmpxchg.org: cgroup-awareness]
Signed-off-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Reported-by: Andres Freund <andres@anarazel.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

https://jira.sw.ru/browse/PSBM-68029
Signed-off-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
---
 include/linux/memcontrol.h |   7 ---
 mm/page_alloc.c            |  44 -------------------
 mm/vmscan.c                | 106 ++++++++++++++++++---------------------------
 3 files changed, 42 insertions(+), 115 deletions(-)

Patch hide | download patch | download mbox

diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h
index 681f3201e86..f6747e4c506 100644
--- a/include/linux/memcontrol.h
+++ b/include/linux/memcontrol.h
@@ -96,7 +96,6 @@  void mem_cgroup_iter_break(struct mem_cgroup *, struct mem_cgroup *);
 /*
  * For memory reclaim.
  */
-int mem_cgroup_inactive_anon_is_low(struct lruvec *lruvec);
 bool mem_cgroup_dcache_is_low(struct mem_cgroup *memcg);
 bool mem_cgroup_low(struct mem_cgroup *root, struct mem_cgroup *memcg);
 bool mem_cgroup_cleancache_disabled(struct page *page);
@@ -316,12 +315,6 @@  static inline void mem_cgroup_put(struct mem_cgroup *memcg)
 {
 }
 
-static inline int
-mem_cgroup_inactive_anon_is_low(struct lruvec *lruvec)
-{
-	return 1;
-}
-
 static inline bool mem_cgroup_dcache_is_low(struct mem_cgroup *memcg)
 {
 	return false;
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 260775672c5..538bb8b4129 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -6384,49 +6384,6 @@  void setup_per_zone_wmarks(void)
 }
 
 /*
- * The inactive anon list should be small enough that the VM never has to
- * do too much work, but large enough that each inactive page has a chance
- * to be referenced again before it is swapped out.
- *
- * The inactive_anon ratio is the target ratio of ACTIVE_ANON to
- * INACTIVE_ANON pages on this zone's LRU, maintained by the
- * pageout code. A zone->inactive_ratio of 3 means 3:1 or 25% of
- * the anonymous pages are kept on the inactive list.
- *
- * total     target    max
- * memory    ratio     inactive anon
- * -------------------------------------
- *   10MB       1         5MB
- *  100MB       1        50MB
- *    1GB       3       250MB
- *   10GB      10       0.9GB
- *  100GB      31         3GB
- *    1TB     101        10GB
- *   10TB     320        32GB
- */
-static void __meminit calculate_zone_inactive_ratio(struct zone *zone)
-{
-	unsigned int gb, ratio;
-
-	/* Zone size in gigabytes */
-	gb = zone->managed_pages >> (30 - PAGE_SHIFT);
-	if (gb)
-		ratio = int_sqrt(10 * gb);
-	else
-		ratio = 1;
-
-	zone->inactive_ratio = ratio;
-}
-
-static void __meminit setup_per_zone_inactive_ratio(void)
-{
-	struct zone *zone;
-
-	for_each_zone(zone)
-		calculate_zone_inactive_ratio(zone);
-}
-
-/*
  * Initialise min_free_kbytes.
  *
  * For small machines we want it small (128k min).  For large machines
@@ -6471,7 +6428,6 @@  int __meminit init_per_zone_wmark_min(void)
 	setup_per_zone_wmarks();
 	refresh_zone_stat_thresholds();
 	setup_per_zone_lowmem_reserve();
-	setup_per_zone_inactive_ratio();
 	return 0;
 }
 module_init(init_per_zone_wmark_min)
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 7b24537448b..26e76201b3f 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -1812,86 +1812,64 @@  static void shrink_active_list(unsigned long nr_to_scan,
 	KSTAT_PERF_LEAVE(refill_inact);
 }
 
-#ifdef CONFIG_SWAP
-static int inactive_anon_is_low_global(struct zone *zone)
-{
-	unsigned long active, inactive;
-
-	active = zone_page_state(zone, NR_ACTIVE_ANON);
-	inactive = zone_page_state(zone, NR_INACTIVE_ANON);
 
-	if (inactive * zone->inactive_ratio < active)
-		return 1;
-
-	return 0;
-}
-
-/**
- * inactive_anon_is_low - check if anonymous pages need to be deactivated
- * @lruvec: LRU vector to check
+/*
+ * The inactive anon list should be small enough that the VM never has
+ * to do too much work.
  *
- * Returns true if the zone does not have enough inactive anon pages,
- * meaning some active anon pages need to be deactivated.
- */
-static int inactive_anon_is_low(struct lruvec *lruvec)
-{
-	/*
-	 * If we don't have swap space, anonymous page deactivation
-	 * is pointless.
-	 */
-	if (!total_swap_pages)
-		return 0;
-
-	if (!mem_cgroup_disabled())
-		return mem_cgroup_inactive_anon_is_low(lruvec);
-
-	return inactive_anon_is_low_global(lruvec_zone(lruvec));
-}
-#else
-static inline int inactive_anon_is_low(struct lruvec *lruvec)
-{
-	return 0;
-}
-#endif
-
-/**
- * inactive_file_is_low - check if file pages need to be deactivated
- * @lruvec: LRU vector to check
+ * The inactive file list should be small enough to leave most memory
+ * to the established workingset on the scan-resistant active list,
+ * but large enough to avoid thrashing the aggregate readahead window.
  *
- * When the system is doing streaming IO, memory pressure here
- * ensures that active file pages get deactivated, until more
- * than half of the file pages are on the inactive list.
+ * Both inactive lists should also be large enough that each inactive
+ * page has a chance to be referenced again before it is reclaimed.
  *
- * Once we get to that situation, protect the system's working
- * set from being evicted by disabling active file page aging.
+ * The inactive_ratio is the target ratio of ACTIVE to INACTIVE pages
+ * on this LRU, maintained by the pageout code. A zone->inactive_ratio
+ * of 3 means 3:1 or 25% of the pages are kept on the inactive list.
  *
- * This uses a different ratio than the anonymous pages, because
- * the page cache uses a use-once replacement algorithm.
+ * total     target    max
+ * memory    ratio     inactive
+ * -------------------------------------
+ *   10MB       1         5MB
+ *  100MB       1        50MB
+ *    1GB       3       250MB
+ *   10GB      10       0.9GB
+ *  100GB      31         3GB
+ *    1TB     101        10GB
+ *   10TB     320        32GB
  */
-static int inactive_file_is_low(struct lruvec *lruvec)
+static int inactive_list_is_low(struct lruvec *lruvec, bool file)
 {
+	unsigned long inactive_ratio;
 	unsigned long inactive;
 	unsigned long active;
+	unsigned long gb;
 
-	inactive = get_lru_size(lruvec, LRU_INACTIVE_FILE);
-	active = get_lru_size(lruvec, LRU_ACTIVE_FILE);
+	/*
+	 * If we don't have swap space, anonymous page deactivation
+	 * is pointless.
+	 */
+	if (!file && !total_swap_pages)
+		return false;
 
-	return active > inactive;
-}
+	inactive = get_lru_size(lruvec, file * LRU_FILE);
+	active = get_lru_size(lruvec, file * LRU_FILE + LRU_ACTIVE);
 
-static int inactive_list_is_low(struct lruvec *lruvec, enum lru_list lru)
-{
-	if (is_file_lru(lru))
-		return inactive_file_is_low(lruvec);
+	gb = (inactive + active) >> (30 - PAGE_SHIFT);
+	if (gb)
+		inactive_ratio = int_sqrt(10 * gb);
 	else
-		return inactive_anon_is_low(lruvec);
+		inactive_ratio = 1;
+
+	return inactive * inactive_ratio < active;
 }
 
 static unsigned long shrink_list(enum lru_list lru, unsigned long nr_to_scan,
 				 struct lruvec *lruvec, struct scan_control *sc)
 {
 	if (is_active_lru(lru)) {
-		if (inactive_list_is_low(lruvec, lru))
+		if (inactive_list_is_low(lruvec, is_file_lru(lru)))
 			shrink_active_list(nr_to_scan, lruvec, sc, lru);
 		return 0;
 	}
@@ -2056,7 +2034,7 @@  static void get_scan_count(struct lruvec *lruvec, struct scan_control *sc,
 	 * There is enough inactive page cache, do not reclaim
 	 * anything from the anonymous working set right now.
 	 */
-	if (!inactive_file_is_low(lruvec) &&
+	if (!inactive_list_is_low(lruvec, true) &&
 	    get_lru_size(lruvec, LRU_INACTIVE_FILE) >> sc->priority > 0) {
 		scan_balance = SCAN_FILE;
 		goto out;
@@ -2280,7 +2258,7 @@  static void shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc,
 	 * Even if we did not try to evict anon pages at all, we want to
 	 * rebalance the anon lru active/inactive ratio.
 	 */
-	if (inactive_anon_is_low(lruvec))
+	if (inactive_list_is_low(lruvec, false))
 		shrink_active_list(SWAP_CLUSTER_MAX, lruvec,
 				   sc, LRU_ACTIVE_ANON);
 
@@ -2975,7 +2953,7 @@  static void age_active_anon(struct zone *zone, struct scan_control *sc)
 	do {
 		struct lruvec *lruvec = mem_cgroup_zone_lruvec(zone, memcg);
 
-		if (inactive_anon_is_low(lruvec))
+		if (inactive_list_is_low(lruvec, false))
 			shrink_active_list(SWAP_CLUSTER_MAX, lruvec,
 					   sc, LRU_ACTIVE_ANON);