[RHEL8,COMMIT] ms/mm: vmscan: detect file thrashing at the reclaim root

Submitted by Konstantin Khorenko on April 2, 2020, 1:03 p.m.

Details

Message ID 202004021303.032D34u6002894@finist_co8.work.ct
State New
Series "Series without cover letter"
Headers show

Commit Message

Konstantin Khorenko April 2, 2020, 1:03 p.m.
The commit is pushed to "branch-rh8-4.18.0-80.1.2.vz8.3.x-ovz" and will appear at https://src.openvz.org/scm/ovz/vzkernel.git
after rh8-4.18.0-80.1.2.vz8.3.4
------>
commit a2532b57037e382689b6f0cb2cf20460f473dd3e
Author: Johannes Weiner <hannes@cmpxchg.org>
Date:   Thu Apr 2 16:03:04 2020 +0300

    ms/mm: vmscan: detect file thrashing at the reclaim root
    
    We use refault information to determine whether the cache workingset is
    stable or transitioning, and dynamically adjust the inactive:active file
    LRU ratio so as to maximize protection from one-off cache during stable
    periods, and minimize IO during transitions.
    
    With cgroups and their nested LRU lists, we currently don't do this
    correctly.  While recursive cgroup reclaim establishes a relative LRU
    order among the pages of all involved cgroups, refaults only affect the
    local LRU order in the cgroup in which they are occuring.  As a result,
    cache transitions can take longer in a cgrouped system as the active pages
    of sibling cgroups aren't challenged when they should be.
    
    [ Right now, this is somewhat theoretical, because the siblings, under
      continued regular reclaim pressure, should eventually run out of
      inactive pages - and since inactive:active *size* balancing is also
      done on a cgroup-local level, we will challenge the active pages
      eventually in most cases. But the next patch will move that relative
      size enforcement to the reclaim root as well, and then this patch
      here will be necessary to propagate refault pressure to siblings. ]
    
    This patch moves refault detection to the root of reclaim.  Instead of
    remembering the cgroup owner of an evicted page, remember the cgroup that
    caused the reclaim to happen.  When refaults later occur, they'll
    correctly influence the cross-cgroup LRU order that reclaim follows.
    
    I.e.  if global reclaim kicked out pages in some subgroup A/B/C, the
    refault of those pages will challenge the global LRU order, and not just
    the local order down inside C.
    
    [hannes@cmpxchg.org:  use page_memcg() instead of another lookup]
      Link: http://lkml.kernel.org/r/20191115160722.GA309754@cmpxchg.org
    Link: http://lkml.kernel.org/r/20191107205334.158354-3-hannes@cmpxchg.org
    Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
    Reviewed-by: Suren Baghdasaryan <surenb@google.com>
    Cc: Andrey Ryabinin <aryabinin@virtuozzo.com>
    Cc: Michal Hocko <mhocko@suse.com>
    Cc: Rik van Riel <riel@surriel.com>
    Cc: Shakeel Butt <shakeelb@google.com>
    Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
    Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
    
    (cherry picked from commit b910718a948a9120d90faf632b33ed23c70e266a)
    Signed-off-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
---
 include/linux/memcontrol.h |  5 ++++
 include/linux/swap.h       |  2 +-
 mm/vmscan.c                | 32 ++++++++++------------
 mm/workingset.c            | 68 ++++++++++++++++++++++++++++++++++++----------
 4 files changed, 74 insertions(+), 33 deletions(-)

Patch hide | download patch | download mbox

diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h
index 77616612c540..43f9020c392c 100644
--- a/include/linux/memcontrol.h
+++ b/include/linux/memcontrol.h
@@ -891,6 +891,11 @@  static inline struct lruvec *mem_cgroup_page_lruvec(struct page *page,
 	return &pgdat->__lruvec;
 }
 
+static inline struct mem_cgroup *parent_mem_cgroup(struct mem_cgroup *memcg)
+{
+	return NULL;
+}
+
 static inline bool mm_match_cgroup(struct mm_struct *mm,
 		struct mem_cgroup *memcg)
 {
diff --git a/include/linux/swap.h b/include/linux/swap.h
index 84e7f5576c35..1c23a9eb4141 100644
--- a/include/linux/swap.h
+++ b/include/linux/swap.h
@@ -299,7 +299,7 @@  struct vma_swap_readahead {
 };
 
 /* linux/mm/workingset.c */
-void *workingset_eviction(struct page *page);
+void *workingset_eviction(struct page *page, struct mem_cgroup *target_memcg);
 void workingset_refault(struct page *page, void *shadow);
 void workingset_activation(struct page *page);
 
diff --git a/mm/vmscan.c b/mm/vmscan.c
index cab7157e3b8b..670ee25e16dc 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -666,7 +666,7 @@  static pageout_t pageout(struct page *page, struct address_space *mapping,
  * gets returned with a refcount of 0.
  */
 static int __remove_mapping(struct address_space *mapping, struct page *page,
-			    bool reclaimed)
+			    bool reclaimed, struct mem_cgroup *target_memcg)
 {
 	unsigned long flags;
 	int refcount;
@@ -741,7 +741,7 @@  static int __remove_mapping(struct address_space *mapping, struct page *page,
 		 */
 		if (reclaimed && page_is_file_cache(page) &&
 		    !mapping_exiting(mapping) && !dax_mapping(mapping))
-			shadow = workingset_eviction(page);
+			shadow = workingset_eviction(page, target_memcg);
 		__delete_from_page_cache(page, shadow);
 		xa_unlock_irqrestore(&mapping->i_pages, flags);
 
@@ -764,7 +764,7 @@  static int __remove_mapping(struct address_space *mapping, struct page *page,
  */
 int remove_mapping(struct address_space *mapping, struct page *page)
 {
-	if (__remove_mapping(mapping, page, false)) {
+	if (__remove_mapping(mapping, page, false, NULL)) {
 		/*
 		 * Unfreezing the refcount with 1 rather than 2 effectively
 		 * drops the pagecache ref for us without requiring another
@@ -1237,7 +1237,8 @@  static unsigned long shrink_page_list(struct list_head *page_list,
 
 			count_vm_event(PGLAZYFREED);
 			count_memcg_page_event(page, PGLAZYFREED);
-		} else if (!mapping || !__remove_mapping(mapping, page, true))
+		} else if (!mapping || !__remove_mapping(mapping, page, true,
+							 sc->target_mem_cgroup))
 			goto keep_locked;
 
 		unlock_page(page);
@@ -1995,6 +1996,7 @@  static bool inactive_list_is_low(struct lruvec *lruvec, bool file,
 	enum lru_list inactive_lru = file * LRU_FILE;
 	unsigned long inactive, active;
 	unsigned long inactive_ratio;
+	struct lruvec *target_lruvec;
 	unsigned long refaults;
 	unsigned long gb;
 
@@ -2006,8 +2008,9 @@  static bool inactive_list_is_low(struct lruvec *lruvec, bool file,
 	 * is being established. Disable active list protection to get
 	 * rid of the stale workingset quickly.
 	 */
-	refaults = lruvec_page_state(lruvec, WORKINGSET_ACTIVATE);
-	if (file && actual_reclaim && lruvec->refaults != refaults) {
+	target_lruvec = mem_cgroup_lruvec(sc->target_mem_cgroup, pgdat);
+	refaults = lruvec_page_state(target_lruvec, WORKINGSET_ACTIVATE);
+	if (file && actual_reclaim && target_lruvec->refaults != refaults) {
 		inactive_ratio = 0;
 	} else {
 		gb = (inactive + active) >> (30 - PAGE_SHIFT);
@@ -2760,19 +2763,14 @@  static void shrink_zones(struct zonelist *zonelist, struct scan_control *sc)
 	sc->gfp_mask = orig_mask;
 }
 
-static void snapshot_refaults(struct mem_cgroup *root_memcg, pg_data_t *pgdat)
+static void snapshot_refaults(struct mem_cgroup *target_memcg, pg_data_t *pgdat)
 {
-	struct mem_cgroup *memcg;
-
-	memcg = mem_cgroup_iter(root_memcg, NULL, NULL);
-	do {
-		unsigned long refaults;
-		struct lruvec *lruvec;
+	struct lruvec *target_lruvec;
+	unsigned long refaults;
 
-		lruvec = mem_cgroup_lruvec(memcg, pgdat);
-		refaults = lruvec_page_state(lruvec, WORKINGSET_ACTIVATE);
-		lruvec->refaults = refaults;
-	} while ((memcg = mem_cgroup_iter(root_memcg, memcg, NULL)));
+	target_lruvec = mem_cgroup_lruvec(target_memcg, pgdat);
+	refaults = lruvec_page_state(target_lruvec, WORKINGSET_ACTIVATE);
+	target_lruvec->refaults = refaults;
 }
 
 /*
diff --git a/mm/workingset.c b/mm/workingset.c
index 931dc2343073..4c053b353c35 100644
--- a/mm/workingset.c
+++ b/mm/workingset.c
@@ -214,28 +214,54 @@  static void unpack_shadow(void *shadow, int *memcgidp, pg_data_t **pgdat,
 	*workingsetp = workingset;
 }
 
+static void advance_inactive_age(struct mem_cgroup *memcg, pg_data_t *pgdat)
+{
+	/*
+	 * Reclaiming a cgroup means reclaiming all its children in a
+	 * round-robin fashion. That means that each cgroup has an LRU
+	 * order that is composed of the LRU orders of its child
+	 * cgroups; and every page has an LRU position not just in the
+	 * cgroup that owns it, but in all of that group's ancestors.
+	 *
+	 * So when the physical inactive list of a leaf cgroup ages,
+	 * the virtual inactive lists of all its parents, including
+	 * the root cgroup's, age as well.
+	 */
+	do {
+		struct lruvec *lruvec;
+
+		lruvec = mem_cgroup_lruvec(memcg, pgdat);
+		atomic_long_inc(&lruvec->inactive_age);
+	} while (memcg && (memcg = parent_mem_cgroup(memcg)));
+}
+
+
 /**
  * workingset_eviction - note the eviction of a page from memory
+ * @target_memcg: the cgroup that is causing the reclaim
  * @page: the page being evicted
  *
  * Returns a shadow entry to be stored in @page->mapping->i_pages in place
  * of the evicted @page so that a later refault can be detected.
  */
-void *workingset_eviction(struct page *page)
+void *workingset_eviction(struct page *page, struct mem_cgroup *target_memcg)
 {
 	struct pglist_data *pgdat = page_pgdat(page);
-	struct mem_cgroup *memcg = page_memcg(page);
-	int memcgid = mem_cgroup_id(memcg);
 	unsigned long eviction;
 	struct lruvec *lruvec;
+	int memcgid;
 
 	/* Page is fully exclusive and pins page->mem_cgroup */
 	VM_BUG_ON_PAGE(PageLRU(page), page);
 	VM_BUG_ON_PAGE(page_count(page), page);
 	VM_BUG_ON_PAGE(!PageLocked(page), page);
 
-	lruvec = mem_cgroup_lruvec(memcg, pgdat);
-	eviction = atomic_long_inc_return(&lruvec->inactive_age);
+	advance_inactive_age(page_memcg(page), pgdat);
+
+	lruvec = mem_cgroup_lruvec(target_memcg, pgdat);
+	/* XXX: target_memcg can be NULL, go through lruvec */
+	memcgid = mem_cgroup_id(lruvec_memcg(lruvec));
+	eviction = atomic_long_read(&lruvec->inactive_age);
 	return pack_shadow(memcgid, pgdat, eviction, PageWorkingset(page));
 }
 
@@ -245,10 +271,13 @@  void *workingset_eviction(struct page *page)
  * @shadow: shadow entry of the evicted page
  *
  * Calculates and evaluates the refault distance of the previously
- * evicted page in the context of the node it was allocated in.
+ * evicted page in the context of the node and the memcg whose memory
+ * pressure caused the eviction.
  */
 void workingset_refault(struct page *page, void *shadow)
 {
+	struct mem_cgroup *eviction_memcg;
+	struct lruvec *eviction_lruvec;
 	unsigned long refault_distance;
 	struct pglist_data *pgdat;
 	unsigned long active_file;
@@ -278,12 +307,12 @@  void workingset_refault(struct page *page, void *shadow)
 	 * would be better if the root_mem_cgroup existed in all
 	 * configurations instead.
 	 */
-	memcg = mem_cgroup_from_id(memcgid);
-	if (!mem_cgroup_disabled() && !memcg)
+	eviction_memcg = mem_cgroup_from_id(memcgid);
+	if (!mem_cgroup_disabled() && !eviction_memcg)
 		goto out;
-	lruvec = mem_cgroup_lruvec(memcg, pgdat);
-	refault = atomic_long_read(&lruvec->inactive_age);
-	active_file = lruvec_lru_size(lruvec, LRU_ACTIVE_FILE, MAX_NR_ZONES);
+	eviction_lruvec = mem_cgroup_lruvec(eviction_memcg, pgdat);
+	refault = atomic_long_read(&eviction_lruvec->inactive_age);
+	active_file = lruvec_page_state(eviction_lruvec, NR_ACTIVE_FILE);
 
 	/*
 	 * Calculate the refault distance
@@ -303,6 +332,17 @@  void workingset_refault(struct page *page, void *shadow)
 	 */
 	refault_distance = (refault - eviction) & EVICTION_MASK;
 
+	/*
+	 * The activation decision for this page is made at the level
+	 * where the eviction occurred, as that is where the LRU order
+	 * during page reclaim is being determined.
+	 *
+	 * However, the cgroup that will own the page is the one that
+	 * is actually experiencing the refault event.
+	 */
+	memcg = page_memcg(page);
+	lruvec = mem_cgroup_lruvec(memcg, pgdat);
+
 	inc_lruvec_state(lruvec, WORKINGSET_REFAULT);
 
 	/*
@@ -314,7 +354,7 @@  void workingset_refault(struct page *page, void *shadow)
 		goto out;
 
 	SetPageActive(page);
-	atomic_long_inc(&lruvec->inactive_age);
+	advance_inactive_age(memcg, pgdat);
 	inc_lruvec_state(lruvec, WORKINGSET_ACTIVATE);
 
 	/* Page was active prior to eviction */
@@ -333,7 +373,6 @@  void workingset_refault(struct page *page, void *shadow)
 void workingset_activation(struct page *page)
 {
 	struct mem_cgroup *memcg;
-	struct lruvec *lruvec;
 
 	rcu_read_lock();
 	/*
@@ -346,8 +385,7 @@  void workingset_activation(struct page *page)
 	memcg = page_memcg_rcu(page);
 	if (!mem_cgroup_disabled() && !memcg)
 		goto out;
-	lruvec = mem_cgroup_lruvec(memcg, page_pgdat(page));
-	atomic_long_inc(&lruvec->inactive_age);
+	advance_inactive_age(memcg, page_pgdat(page));
 out:
 	rcu_read_unlock();
 }