Subject: + mm-memcontrol-factor-out-reclaim-iterator-loading-and-updating.patch added to -mm tree
To: hannes@xxxxxxxxxxx,glommer@xxxxxxxxxxxxx,kamezawa.hiroyu@xxxxxxxxxxxxxx,mhocko@xxxxxxx,tj@xxxxxxxxxx
From: akpm@xxxxxxxxxxxxxxxxxxxx
Date: Wed, 05 Jun 2013 16:05:18 -0700
The patch titled
Subject: mm: memcontrol: factor out reclaim iterator loading and updating
has been added to the -mm tree. Its filename is
mm-memcontrol-factor-out-reclaim-iterator-loading-and-updating.patch
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/SubmitChecklist when testing your code ***
The -mm tree is included into linux-next and is updated
there every 3-4 working days
------------------------------------------------------
From: Johannes Weiner <hannes@xxxxxxxxxxx>
Subject: mm: memcontrol: factor out reclaim iterator loading and updating
mem_cgroup_iter() is too hard to follow. Factor out the lockless reclaim
iterator loading and updating so it's easier to follow the big picture.
Also document the iterator invalidation mechanism a bit more extensively.
Signed-off-by: Johannes Weiner <hannes@xxxxxxxxxxx>
Reported-by: Tejun Heo <tj@xxxxxxxxxx>
Cc: Michal Hocko <mhocko@xxxxxxx>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>
Cc: Glauber Costa <glommer@xxxxxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
mm/memcontrol.c | 86 ++++++++++++++++++++++++++++++----------------
1 file changed, 57 insertions(+), 29 deletions(-)
diff -puN mm/memcontrol.c~mm-memcontrol-factor-out-reclaim-iterator-loading-and-updating mm/memcontrol.c
--- a/mm/memcontrol.c~mm-memcontrol-factor-out-reclaim-iterator-loading-and-updating
+++ a/mm/memcontrol.c
@@ -1148,6 +1148,58 @@ skip_node:
return NULL;
}
+static void mem_cgroup_iter_invalidate(struct mem_cgroup *root)
+{
+ /*
+ * When a group in the hierarchy below root is destroyed, the
+ * hierarchy iterator can no longer be trusted since it might
+ * have pointed to the destroyed group. Invalidate it.
+ */
+ atomic_inc(&root->dead_count);
+}
+
+static struct mem_cgroup *
+mem_cgroup_iter_load(struct mem_cgroup_reclaim_iter *iter,
+ struct mem_cgroup *root,
+ int *sequence)
+{
+ struct mem_cgroup *position = NULL;
+ /*
+ * A cgroup destruction happens in two stages: offlining and
+ * release. They are separated by a RCU grace period.
+ *
+ * If the iterator is valid, we may still race with an
+ * offlining. The RCU lock ensures the object won't be
+ * released, tryget will fail if we lost the race.
+ */
+ *sequence = atomic_read(&root->dead_count);
+ if (iter->last_dead_count == *sequence) {
+ smp_rmb();
+ position = iter->last_visited;
+ if (position && !css_tryget(&position->css))
+ position = NULL;
+ }
+ return position;
+}
+
+static void mem_cgroup_iter_update(struct mem_cgroup_reclaim_iter *iter,
+ struct mem_cgroup *last_visited,
+ struct mem_cgroup *new_position,
+ int sequence)
+{
+ if (last_visited)
+ css_put(&last_visited->css);
+ /*
+ * We store the sequence count from the time @last_visited was
+ * loaded successfully instead of rereading it here so that we
+ * don't lose destruction events in between. We could have
+ * raced with the destruction of @new_position after all.
+ */
+ iter->last_visited = new_position;
+ smp_wmb();
+ iter->last_dead_count = sequence;
+}
+
/**
* mem_cgroup_iter - iterate over memory cgroup hierarchy
* @root: hierarchy root
@@ -1171,7 +1223,6 @@ struct mem_cgroup *mem_cgroup_iter(struc
{
struct mem_cgroup *memcg = NULL;
struct mem_cgroup *last_visited = NULL;
- unsigned long uninitialized_var(dead_count);
if (mem_cgroup_disabled())
return NULL;
@@ -1191,6 +1242,7 @@ struct mem_cgroup *mem_cgroup_iter(struc
rcu_read_lock();
while (!memcg) {
struct mem_cgroup_reclaim_iter *uninitialized_var(iter);
+ int uninitialized_var(seq);
if (reclaim) {
int nid = zone_to_nid(reclaim->zone);
@@ -1204,37 +1256,13 @@ struct mem_cgroup *mem_cgroup_iter(struc
goto out_unlock;
}
- /*
- * If the dead_count mismatches, a destruction
- * has happened or is happening concurrently.
- * If the dead_count matches, a destruction
- * might still happen concurrently, but since
- * we checked under RCU, that destruction
- * won't free the object until we release the
- * RCU reader lock. Thus, the dead_count
- * check verifies the pointer is still valid,
- * css_tryget() verifies the cgroup pointed to
- * is alive.
- */
- dead_count = atomic_read(&root->dead_count);
- if (dead_count == iter->last_dead_count) {
- smp_rmb();
- last_visited = iter->last_visited;
- if (last_visited &&
- !css_tryget(&last_visited->css))
- last_visited = NULL;
- }
+ last_visited = mem_cgroup_iter_load(iter, root, &seq);
}
memcg = __mem_cgroup_iter_next(root, last_visited);
if (reclaim) {
- if (last_visited)
- css_put(&last_visited->css);
-
- iter->last_visited = memcg;
- smp_wmb();
- iter->last_dead_count = dead_count;
+ mem_cgroup_iter_update(iter, last_visited, memcg, seq);
if (!memcg)
iter->generation++;
@@ -6317,14 +6345,14 @@ static void mem_cgroup_invalidate_reclai
struct mem_cgroup *parent = memcg;
while ((parent = parent_mem_cgroup(parent)))
- atomic_inc(&parent->dead_count);
+ mem_cgroup_iter_invalidate(parent);
/*
* if the root memcg is not hierarchical we have to check it
* explicitely.
*/
if (!root_mem_cgroup->use_hierarchy)
- atomic_inc(&root_mem_cgroup->dead_count);
+ mem_cgroup_iter_invalidate(root_mem_cgroup);
}
static void mem_cgroup_css_offline(struct cgroup *cont)
_
Patches currently in -mm which might be from hannes@xxxxxxxxxxx are
memcg-dont-initialize-kmem-cache-destroying-work-for-root-caches.patch
swap-avoid-read_swap_cache_async-race-to-deadlock-while-waiting-on-discard-i-o-completion.patch
mm-memcontrol-fix-lockless-reclaim-hierarchy-iterator.patch
mm-memcontrol-factor-out-reclaim-iterator-loading-and-updating.patch
mm-memcg-dont-take-task_lock-in-task_in_mem_cgroup.patch
mm-vmscan-limit-the-number-of-pages-kswapd-reclaims-at-each-priority.patch
mm-vmscan-obey-proportional-scanning-requirements-for-kswapd.patch
mm-vmscan-flatten-kswapd-priority-loop.patch
mm-vmscan-decide-whether-to-compact-the-pgdat-based-on-reclaim-progress.patch
mm-vmscan-do-not-allow-kswapd-to-scan-at-maximum-priority.patch
mm-vmscan-have-kswapd-writeback-pages-based-on-dirty-pages-encountered-not-priority.patch
mm-vmscan-block-kswapd-if-it-is-encountering-pages-under-writeback.patch
mm-vmscan-block-kswapd-if-it-is-encountering-pages-under-writeback-fix.patch
mm-vmscan-check-if-kswapd-should-writepage-once-per-pgdat-scan.patch
mm-vmscan-move-logic-from-balance_pgdat-to-kswapd_shrink_zone.patch
mm-vmscan-stall-page-reclaim-and-writeback-pages-based-on-dirty-writepage-pages-encountered-v3.patch
mm-vmscan-stall-page-reclaim-after-a-list-of-pages-have-been-processed-v3.patch
mm-vmscan-set-zone-flags-before-blocking.patch
mm-vmscan-move-direct-reclaim-wait_iff_congested-into-shrink_list.patch
mm-vmscan-treat-pages-marked-for-immediate-reclaim-as-zone-congestion.patch
mm-vmscan-take-page-buffers-dirty-and-locked-state-into-account-v3.patch
fs-nfs-inform-the-vm-about-pages-being-committed-or-unstable.patch
memcg-update-todo-list-in-documentation.patch
mm-add-tracepoints-for-lru-activation-and-insertions.patch
mm-pagevec-defer-deciding-what-lru-to-add-a-page-to-until-pagevec-drain-time.patch
mm-activate-pagelru-pages-on-mark_page_accessed-if-page-is-on-local-pagevec.patch
mm-remove-lru-parameter-from-__pagevec_lru_add-and-remove-parts-of-pagevec-api.patch
mm-remove-lru-parameter-from-__lru_cache_add-and-lru_cache_add_lru.patch
memcg-kconfig-info-update.patch
mm-kill-free_all_bootmem_node.patch
memcg-debugging-facility-to-access-dangling-memcgs.patch
fs-bump-inode-and-dentry-counters-to-long.patch
super-fix-calculation-of-shrinkable-objects-for-small-numbers.patch
dcache-convert-dentry_statnr_unused-to-per-cpu-counters.patch
dentry-move-to-per-sb-lru-locks.patch
dcache-remove-dentries-from-lru-before-putting-on-dispose-list.patch
mm-new-shrinker-api.patch
shrinker-convert-superblock-shrinkers-to-new-api.patch
list-add-a-new-lru-list-type.patch
inode-convert-inode-lru-list-to-generic-lru-list-code.patch
dcache-convert-to-use-new-lru-list-infrastructure.patch
list_lru-per-node-list-infrastructure.patch
shrinker-add-node-awareness.patch
vmscan-per-node-deferred-work.patch
list_lru-per-node-api.patch
fs-convert-inode-and-dentry-shrinking-to-be-node-aware.patch
xfs-convert-buftarg-lru-to-generic-code.patch
xfs-rework-buffer-dispose-list-tracking.patch
xfs-convert-dquot-cache-lru-to-list_lru.patch
fs-convert-fs-shrinkers-to-new-scan-count-api.patch
drivers-convert-shrinkers-to-new-count-scan-api.patch
i915-bail-out-earlier-when-shrinker-cannot-acquire-mutex.patch
shrinker-convert-remaining-shrinkers-to-count-scan-api.patch
hugepage-convert-huge-zero-page-shrinker-to-new-shrinker-api.patch
shrinker-kill-old-shrink-api.patch
vmscan-also-shrink-slab-in-memcg-pressure.patch
memcglist_lru-duplicate-lrus-upon-kmemcg-creation.patch
lru-add-an-element-to-a-memcg-list.patch
list_lru-per-memcg-walks.patch
memcg-per-memcg-kmem-shrinking.patch
memcg-scan-cache-objects-hierarchically.patch
vmscan-take-at-least-one-pass-with-shrinkers.patch
super-targeted-memcg-reclaim.patch
memcg-move-initialization-to-memcg-creation.patch
vmpressure-in-kernel-notifications.patch
memcg-reap-dead-memcgs-upon-global-memory-pressure.patch
mm-memmap_init_zone-performance-improvement.patch
debugging-keep-track-of-page-owners-fix-2-fix-fix-fix.patch
--
To unsubscribe from this list: send the line "unsubscribe mm-commits" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
