mm: memcontrol: fix excessive complexity in memory.stat reporting

We've seen memory.stat reads in top-level cgroups take up to fourteen
seconds during a userspace bug that created tens of thousands of ghost
cgroups pinned by lingering page cache.

Even with a more reasonable number of cgroups, aggregating memory.stat
is unnecessarily heavy.  The complexity is this:

	nr_cgroups * nr_stat_items * nr_possible_cpus

where the stat items are ~70 at this point.  With 128 cgroups and 128
CPUs - decent, not enormous setups - reading the top-level memory.stat
has to aggregate over a million per-cpu counters.  This doesn't scale.

Instead of spreading the source of truth across all CPUs, use the
per-cpu counters merely to batch updates to shared atomic counters.

This is the same as the per-cpu stocks we use for charging memory to the
shared atomic page_counters, and also the way the global vmstat counters
are implemented.

Vmstat has elaborate spilling thresholds that depend on the number of
CPUs, amount of memory, and memory pressure - carefully balancing the
cost of counter updates with the amount of per-cpu error.  That's
because the vmstat counters are system-wide, but also used for decisions
inside the kernel (e.g.  NR_FREE_PAGES in the allocator).  Neither is
true for the memory controller.

Use the same static batch size we already use for page_counter updates
during charging.  The per-cpu error in the stats will be 128k, which is
an acceptable ratio of cores to memory accounting granularity.

[hannes@cmpxchg.org: fix warning in __this_cpu_xchg() calls]
  Link: http://lkml.kernel.org/r/20171201135750.GB8097@cmpxchg.org
Link: http://lkml.kernel.org/r/20171103153336.24044-3-hannes@cmpxchg.org
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com>
Cc: Michal Hocko <mhocko@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

1 files changed, 62 insertions, 34 deletions

diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h
index 1ffc54ac4cc9..882046863581 100644
--- a/include/linux/memcontrol.h
+++ b/include/linux/memcontrol.h
@@ -108,7 +108,10 @@ struct lruvec_stat {
  */
 struct mem_cgroup_per_node {
 	struct lruvec		lruvec;
-	struct lruvec_stat __percpu *lruvec_stat;
+
+	struct lruvec_stat __percpu *lruvec_stat_cpu;
+	atomic_long_t		lruvec_stat[NR_VM_NODE_STAT_ITEMS];
+
 	unsigned long		lru_zone_size[MAX_NR_ZONES][NR_LRU_LISTS];
 
 	struct mem_cgroup_reclaim_iter	iter[DEF_PRIORITY + 1];
@@ -227,10 +230,10 @@ struct mem_cgroup {
 	spinlock_t		move_lock;
 	struct task_struct	*move_lock_task;
 	unsigned long		move_lock_flags;
-	/*
-	 * percpu counter.
-	 */
-	struct mem_cgroup_stat_cpu __percpu *stat;
+
+	struct mem_cgroup_stat_cpu __percpu *stat_cpu;
+	atomic_long_t		stat[MEMCG_NR_STAT];
+	atomic_long_t		events[MEMCG_NR_EVENTS];
 
 	unsigned long		socket_pressure;
 
@@ -265,6 +268,12 @@ struct mem_cgroup {
 	/* WARNING: nodeinfo must be the last member here */
 };
 
+/*
+ * size of first charge trial. "32" comes from vmscan.c's magic value.
+ * TODO: maybe necessary to use big numbers in big irons.
+ */
+#define MEMCG_CHARGE_BATCH 32U
+
 extern struct mem_cgroup *root_mem_cgroup;
 
 static inline bool mem_cgroup_disabled(void)
@@ -485,32 +494,38 @@ void unlock_page_memcg(struct page *page);
 static inline unsigned long memcg_page_state(struct mem_cgroup *memcg,
 					     int idx)
 {
-	long val = 0;
-	int cpu;
-
-	for_each_possible_cpu(cpu)
-		val += per_cpu(memcg->stat->count[idx], cpu);
-
-	if (val < 0)
-		val = 0;
-
-	return val;
+	long x = atomic_long_read(&memcg->stat[idx]);
+#ifdef CONFIG_SMP
+	if (x < 0)
+		x = 0;
+#endif
+	return x;
 }
 
 /* idx can be of type enum memcg_stat_item or node_stat_item */
 static inline void __mod_memcg_state(struct mem_cgroup *memcg,
 				     int idx, int val)
 {
-	if (!mem_cgroup_disabled())
-		__this_cpu_add(memcg->stat->count[idx], val);
+	long x;
+
+	if (mem_cgroup_disabled())
+		return;
+
+	x = val + __this_cpu_read(memcg->stat_cpu->count[idx]);
+	if (unlikely(abs(x) > MEMCG_CHARGE_BATCH)) {
+		atomic_long_add(x, &memcg->stat[idx]);
+		x = 0;
+	}
+	__this_cpu_write(memcg->stat_cpu->count[idx], x);
 }
 
 /* idx can be of type enum memcg_stat_item or node_stat_item */
 static inline void mod_memcg_state(struct mem_cgroup *memcg,
 				   int idx, int val)
 {
-	if (!mem_cgroup_disabled())
-		this_cpu_add(memcg->stat->count[idx], val);
+	preempt_disable();
+	__mod_memcg_state(memcg, idx, val);
+	preempt_enable();
 }
 
 /**
@@ -548,26 +563,25 @@ static inline unsigned long lruvec_page_state(struct lruvec *lruvec,
 					      enum node_stat_item idx)
 {
 	struct mem_cgroup_per_node *pn;
-	long val = 0;
-	int cpu;
+	long x;
 
 	if (mem_cgroup_disabled())
 		return node_page_state(lruvec_pgdat(lruvec), idx);
 
 	pn = container_of(lruvec, struct mem_cgroup_per_node, lruvec);
-	for_each_possible_cpu(cpu)
-		val += per_cpu(pn->lruvec_stat->count[idx], cpu);
-
-	if (val < 0)
-		val = 0;
-
-	return val;
+	x = atomic_long_read(&pn->lruvec_stat[idx]);
+#ifdef CONFIG_SMP
+	if (x < 0)
+		x = 0;
+#endif
+	return x;
 }
 
 static inline void __mod_lruvec_state(struct lruvec *lruvec,
 				      enum node_stat_item idx, int val)
 {
 	struct mem_cgroup_per_node *pn;
+	long x;
 
 	/* Update node */
 	__mod_node_page_state(lruvec_pgdat(lruvec), idx, val);
@@ -581,7 +595,12 @@ static inline void __mod_lruvec_state(struct lruvec *lruvec,
 	__mod_memcg_state(pn->memcg, idx, val);
 
 	/* Update lruvec */
-	__this_cpu_add(pn->lruvec_stat->count[idx], val);
+	x = val + __this_cpu_read(pn->lruvec_stat_cpu->count[idx]);
+	if (unlikely(abs(x) > MEMCG_CHARGE_BATCH)) {
+		atomic_long_add(x, &pn->lruvec_stat[idx]);
+		x = 0;
+	}
+	__this_cpu_write(pn->lruvec_stat_cpu->count[idx], x);
 }
 
 static inline void mod_lruvec_state(struct lruvec *lruvec,
@@ -624,16 +643,25 @@ unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order,
 static inline void __count_memcg_events(struct mem_cgroup *memcg,
 					int idx, unsigned long count)
 {
-	if (!mem_cgroup_disabled())
-		__this_cpu_add(memcg->stat->events[idx], count);
+	unsigned long x;
+
+	if (mem_cgroup_disabled())
+		return;
+
+	x = count + __this_cpu_read(memcg->stat_cpu->events[idx]);
+	if (unlikely(x > MEMCG_CHARGE_BATCH)) {
+		atomic_long_add(x, &memcg->events[idx]);
+		x = 0;
+	}
+	__this_cpu_write(memcg->stat_cpu->events[idx], x);
 }
 
-/* idx can be of type enum memcg_event_item or vm_event_item */
 static inline void count_memcg_events(struct mem_cgroup *memcg,
 				      int idx, unsigned long count)
 {
-	if (!mem_cgroup_disabled())
-		this_cpu_add(memcg->stat->events[idx], count);
+	preempt_disable();
+	__count_memcg_events(memcg, idx, count);
+	preempt_enable();
 }
 
 /* idx can be of type enum memcg_event_item or vm_event_item */