[PATCH v2 0/3] percpu: memcg memory accounting rework

[PATCH v2 0/3] percpu: memcg memory accounting rework

[Roman Gushchin]

This patchset reworks memcg-based accounting of the percpu memory:
instead of using two types of chunks (memcg-aware and non-memcg-aware)
it treats all chunks as memcg-aware. An extra memory used for objcg
arrays will is compensated by a better chunks utilization, so the
total memory footprint should be the same or smaller.

First two patches are preparational changes and cleanups on the memcg side.
The third one is percpu accounting rework.

v2:
  - make all chunks memcg-aware instead of on-demand objcg allocation, by Dennis

v1:
  https://lkml.org/lkml/2021/5/11/1343


Roman Gushchin (3):
  mm, memcg: mark cgroup_memory_nosocket, nokmem and noswap as
    __ro_after_init
  mm, memcg: introduce mem_cgroup_kmem_disabled()
  percpu: rework memcg accounting

 include/linux/memcontrol.h |   5 ++
 mm/memcontrol.c            |  11 ++-
 mm/percpu-internal.h       |  52 +--------------
 mm/percpu-km.c             |   5 +-
 mm/percpu-stats.c          |  46 +++++--------
 mm/percpu-vm.c             |  11 ++-
 mm/percpu.c                | 133 +++++++++++++++----------------------
 7 files changed, 88 insertions(+), 175 deletions(-)

-- 
2.31.1

[PATCH v2 1/3] mm, memcg: mark cgroup_memory_nosocket, nokmem and noswap as __ro_after_init

[Roman Gushchin]

cgroup_memory_nosocket, cgroup_memory_nokmem and cgroup_memory_noswap
are initialized during the kernel initialization and never change
their value afterwards.

cgroup_memory_nosocket, cgroup_memory_nokmem are written only from
cgroup_memory(), which is marked as __init.

cgroup_memory_noswap is written from setup_swap_account() and
mem_cgroup_swap_init(), both are marked as __init.

Mark all three variables as __ro_after_init.

Signed-off-by: Roman Gushchin <guro@fb.com>
---
 mm/memcontrol.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index e064ac0d850a..e6203ee24a11 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -80,14 +80,14 @@ struct mem_cgroup *root_mem_cgroup __read_mostly;
 DEFINE_PER_CPU(struct mem_cgroup *, int_active_memcg);
 
 /* Socket memory accounting disabled? */
-static bool cgroup_memory_nosocket;
+static bool cgroup_memory_nosocket __ro_after_init;
 
 /* Kernel memory accounting disabled? */
-static bool cgroup_memory_nokmem;
+static bool cgroup_memory_nokmem __ro_after_init;
 
 /* Whether the swap controller is active */
 #ifdef CONFIG_MEMCG_SWAP
-bool cgroup_memory_noswap __read_mostly;
+bool cgroup_memory_noswap __ro_after_init;
 #else
 #define cgroup_memory_noswap		1
 #endif
-- 
2.31.1

[PATCH v2 2/3] mm, memcg: introduce mem_cgroup_kmem_disabled()

[Roman Gushchin]

Introduce a new mem_cgroup_kmem_disabled() helper, similar to
mem_cgroup_disabled(), to check whether the kernel memory accounting
is off. A user could disable it using a boot option to eliminate
some associated costs.

The helper can be used outside of memcontrol.c to dynamically disable
the kmem-related code. The returned value is stable after the kernel
initialization is finished.

Signed-off-by: Roman Gushchin <guro@fb.com>
---
 include/linux/memcontrol.h | 5 +++++
 mm/memcontrol.c            | 5 +++++
 2 files changed, 10 insertions(+)

diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h
index 0c04d39a7967..8ef51c58f470 100644
--- a/include/linux/memcontrol.h
+++ b/include/linux/memcontrol.h
@@ -1583,6 +1583,7 @@ static inline void memcg_set_shrinker_bit(struct mem_cgroup *memcg,
 #endif
 
 #ifdef CONFIG_MEMCG_KMEM
+bool mem_cgroup_kmem_disabled(void);
 int __memcg_kmem_charge_page(struct page *page, gfp_t gfp, int order);
 void __memcg_kmem_uncharge_page(struct page *page, int order);
 
@@ -1636,6 +1637,10 @@ static inline int memcg_cache_id(struct mem_cgroup *memcg)
 struct mem_cgroup *mem_cgroup_from_obj(void *p);
 
 #else
+static inline bool mem_cgroup_kmem_disabled(void)
+{
+	return true;
+}
 
 static inline int memcg_kmem_charge_page(struct page *page, gfp_t gfp,
 					 int order)
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index e6203ee24a11..1fa9b00ec71d 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -255,6 +255,11 @@ struct cgroup_subsys_state *vmpressure_to_css(struct vmpressure *vmpr)
 #ifdef CONFIG_MEMCG_KMEM
 extern spinlock_t css_set_lock;
 
+bool mem_cgroup_kmem_disabled(void)
+{
+	return cgroup_memory_nokmem;
+}
+
 static int __memcg_kmem_charge(struct mem_cgroup *memcg, gfp_t gfp,
 			       unsigned int nr_pages);
 static void __memcg_kmem_uncharge(struct mem_cgroup *memcg,
-- 
2.31.1

[PATCH v2 3/3] percpu: rework memcg accounting

[Roman Gushchin]

The current implementation of the memcg accounting of the percpu
memory is based on the idea of having two separate sets of chunks for
accounted and non-accounted memory. This approach has an advantage
of not wasting any extra memory for memcg data for non-accounted
chunks, however it complicates the code and leads to a higher chunks
number due to a lower chunk utilization.

Instead of having two chunk types it's possible to declare all* chunks
memcg-aware unless the kernel memory accounting is disabled globally
by a boot option. The size of objcg_array is usually small in
comparison to chunks themselves (it obviously depends on the number of
CPUs), so even if some chunk will have no accounted allocations, the
memory waste isn't significant and will likely be compensated by
a higher chunk utilization. Also, with time more and more percpu
allocations will likely become accounted.

* The first chunk is initialized before the memory cgroup subsystem,
  so we don't know for sure whether we need to allocate obj_cgroups.
  Because it's small, let's make it free for use. Then we don't need
  to allocate obj_cgroups for it.

Signed-off-by: Roman Gushchin <guro@fb.com>
---
 mm/percpu-internal.h |  52 +----------------
 mm/percpu-km.c       |   5 +-
 mm/percpu-stats.c    |  46 +++++----------
 mm/percpu-vm.c       |  11 ++--
 mm/percpu.c          | 133 +++++++++++++++++--------------------------
 5 files changed, 75 insertions(+), 172 deletions(-)

diff --git a/mm/percpu-internal.h b/mm/percpu-internal.h
index 10604dce806f..b6dc22904088 100644
--- a/mm/percpu-internal.h
+++ b/mm/percpu-internal.h
@@ -5,25 +5,6 @@
 #include <linux/types.h>
 #include <linux/percpu.h>
 
-/*
- * There are two chunk types: root and memcg-aware.
- * Chunks of each type have separate slots list.
- *
- * Memcg-aware chunks have an attached vector of obj_cgroup pointers, which is
- * used to store memcg membership data of a percpu object.  Obj_cgroups are
- * ref-counted pointers to a memory cgroup with an ability to switch dynamically
- * to the parent memory cgroup.  This allows to reclaim a deleted memory cgroup
- * without reclaiming of all outstanding objects, which hold a reference at it.
- */
-enum pcpu_chunk_type {
-	PCPU_CHUNK_ROOT,
-#ifdef CONFIG_MEMCG_KMEM
-	PCPU_CHUNK_MEMCG,
-#endif
-	PCPU_NR_CHUNK_TYPES,
-	PCPU_FAIL_ALLOC = PCPU_NR_CHUNK_TYPES
-};
-
 /*
  * pcpu_block_md is the metadata block struct.
  * Each chunk's bitmap is split into a number of full blocks.
@@ -91,7 +72,7 @@ extern struct list_head *pcpu_chunk_lists;
 extern int pcpu_nr_slots;
 extern int pcpu_sidelined_slot;
 extern int pcpu_to_depopulate_slot;
-extern int pcpu_nr_empty_pop_pages[];
+extern int pcpu_nr_empty_pop_pages;
 
 extern struct pcpu_chunk *pcpu_first_chunk;
 extern struct pcpu_chunk *pcpu_reserved_chunk;
@@ -132,37 +113,6 @@ static inline int pcpu_chunk_map_bits(struct pcpu_chunk *chunk)
 	return pcpu_nr_pages_to_map_bits(chunk->nr_pages);
 }
 
-#ifdef CONFIG_MEMCG_KMEM
-static inline enum pcpu_chunk_type pcpu_chunk_type(struct pcpu_chunk *chunk)
-{
-	if (chunk->obj_cgroups)
-		return PCPU_CHUNK_MEMCG;
-	return PCPU_CHUNK_ROOT;
-}
-
-static inline bool pcpu_is_memcg_chunk(enum pcpu_chunk_type chunk_type)
-{
-	return chunk_type == PCPU_CHUNK_MEMCG;
-}
-
-#else
-static inline enum pcpu_chunk_type pcpu_chunk_type(struct pcpu_chunk *chunk)
-{
-	return PCPU_CHUNK_ROOT;
-}
-
-static inline bool pcpu_is_memcg_chunk(enum pcpu_chunk_type chunk_type)
-{
-	return false;
-}
-#endif
-
-static inline struct list_head *pcpu_chunk_list(enum pcpu_chunk_type chunk_type)
-{
-	return &pcpu_chunk_lists[pcpu_nr_slots *
-				 pcpu_is_memcg_chunk(chunk_type)];
-}
-
 #ifdef CONFIG_PERCPU_STATS
 
 #include <linux/spinlock.h>
diff --git a/mm/percpu-km.c b/mm/percpu-km.c
index c84a9f781a6c..c9d529dc7651 100644
--- a/mm/percpu-km.c
+++ b/mm/percpu-km.c
@@ -44,8 +44,7 @@ static void pcpu_depopulate_chunk(struct pcpu_chunk *chunk,
 	/* nada */
 }
 
-static struct pcpu_chunk *pcpu_create_chunk(enum pcpu_chunk_type type,
-					    gfp_t gfp)
+static struct pcpu_chunk *pcpu_create_chunk(gfp_t gfp)
 {
 	const int nr_pages = pcpu_group_sizes[0] >> PAGE_SHIFT;
 	struct pcpu_chunk *chunk;
@@ -53,7 +52,7 @@ static struct pcpu_chunk *pcpu_create_chunk(enum pcpu_chunk_type type,
 	unsigned long flags;
 	int i;
 
-	chunk = pcpu_alloc_chunk(type, gfp);
+	chunk = pcpu_alloc_chunk(gfp);
 	if (!chunk)
 		return NULL;
 
diff --git a/mm/percpu-stats.c b/mm/percpu-stats.c
index 2125981acfb9..c6bd092ff7a3 100644
--- a/mm/percpu-stats.c
+++ b/mm/percpu-stats.c
@@ -34,15 +34,11 @@ static int find_max_nr_alloc(void)
 {
 	struct pcpu_chunk *chunk;
 	int slot, max_nr_alloc;
-	enum pcpu_chunk_type type;
 
 	max_nr_alloc = 0;
-	for (type = 0; type < PCPU_NR_CHUNK_TYPES; type++)
-		for (slot = 0; slot < pcpu_nr_slots; slot++)
-			list_for_each_entry(chunk, &pcpu_chunk_list(type)[slot],
-					    list)
-				max_nr_alloc = max(max_nr_alloc,
-						   chunk->nr_alloc);
+	for (slot = 0; slot < pcpu_nr_slots; slot++)
+		list_for_each_entry(chunk, &pcpu_chunk_lists[slot], list)
+			max_nr_alloc = max(max_nr_alloc, chunk->nr_alloc);
 
 	return max_nr_alloc;
 }
@@ -133,9 +129,6 @@ static void chunk_map_stats(struct seq_file *m, struct pcpu_chunk *chunk,
 	P("cur_min_alloc", cur_min_alloc);
 	P("cur_med_alloc", cur_med_alloc);
 	P("cur_max_alloc", cur_max_alloc);
-#ifdef CONFIG_MEMCG_KMEM
-	P("memcg_aware", pcpu_is_memcg_chunk(pcpu_chunk_type(chunk)));
-#endif
 	seq_putc(m, '\n');
 }
 
@@ -144,8 +137,6 @@ static int percpu_stats_show(struct seq_file *m, void *v)
 	struct pcpu_chunk *chunk;
 	int slot, max_nr_alloc;
 	int *buffer;
-	enum pcpu_chunk_type type;
-	int nr_empty_pop_pages;
 
 alloc_buffer:
 	spin_lock_irq(&pcpu_lock);
@@ -166,10 +157,6 @@ static int percpu_stats_show(struct seq_file *m, void *v)
 		goto alloc_buffer;
 	}
 
-	nr_empty_pop_pages = 0;
-	for (type = 0; type < PCPU_NR_CHUNK_TYPES; type++)
-		nr_empty_pop_pages += pcpu_nr_empty_pop_pages[type];
-
 #define PL(X)								\
 	seq_printf(m, "  %-20s: %12lld\n", #X, (long long int)pcpu_stats_ai.X)
 
@@ -201,7 +188,7 @@ static int percpu_stats_show(struct seq_file *m, void *v)
 	PU(nr_max_chunks);
 	PU(min_alloc_size);
 	PU(max_alloc_size);
-	P("empty_pop_pages", nr_empty_pop_pages);
+	P("empty_pop_pages", pcpu_nr_empty_pop_pages);
 	seq_putc(m, '\n');
 
 #undef PU
@@ -215,20 +202,17 @@ static int percpu_stats_show(struct seq_file *m, void *v)
 		chunk_map_stats(m, pcpu_reserved_chunk, buffer);
 	}
 
-	for (type = 0; type < PCPU_NR_CHUNK_TYPES; type++) {
-		for (slot = 0; slot < pcpu_nr_slots; slot++) {
-			list_for_each_entry(chunk, &pcpu_chunk_list(type)[slot],
-					    list) {
-				if (chunk == pcpu_first_chunk)
-					seq_puts(m, "Chunk: <- First Chunk\n");
-				else if (slot == pcpu_to_depopulate_slot)
-					seq_puts(m, "Chunk (to_depopulate)\n");
-				else if (slot == pcpu_sidelined_slot)
-					seq_puts(m, "Chunk (sidelined):\n");
-				else
-					seq_puts(m, "Chunk:\n");
-				chunk_map_stats(m, chunk, buffer);
-			}
+	for (slot = 0; slot < pcpu_nr_slots; slot++) {
+		list_for_each_entry(chunk, &pcpu_chunk_lists[slot], list) {
+			if (chunk == pcpu_first_chunk)
+				seq_puts(m, "Chunk: <- First Chunk\n");
+			else if (slot == pcpu_to_depopulate_slot)
+				seq_puts(m, "Chunk (to_depopulate)\n");
+			else if (slot == pcpu_sidelined_slot)
+				seq_puts(m, "Chunk (sidelined):\n");
+			else
+				seq_puts(m, "Chunk:\n");
+			chunk_map_stats(m, chunk, buffer);
 		}
 	}
 
diff --git a/mm/percpu-vm.c b/mm/percpu-vm.c
index c75f6f24f2d5..e59566b949d1 100644
--- a/mm/percpu-vm.c
+++ b/mm/percpu-vm.c
@@ -328,13 +328,12 @@ static void pcpu_depopulate_chunk(struct pcpu_chunk *chunk,
 	pcpu_free_pages(chunk, pages, page_start, page_end);
 }
 
-static struct pcpu_chunk *pcpu_create_chunk(enum pcpu_chunk_type type,
-					    gfp_t gfp)
+static struct pcpu_chunk *pcpu_create_chunk(gfp_t gfp)
 {
 	struct pcpu_chunk *chunk;
 	struct vm_struct **vms;
 
-	chunk = pcpu_alloc_chunk(type, gfp);
+	chunk = pcpu_alloc_chunk(gfp);
 	if (!chunk)
 		return NULL;
 
@@ -403,7 +402,7 @@ static bool pcpu_should_reclaim_chunk(struct pcpu_chunk *chunk)
 	 * chunk, move it to the to_depopulate list.
 	 */
 	return ((chunk->isolated && chunk->nr_empty_pop_pages) ||
-		(pcpu_nr_empty_pop_pages[pcpu_chunk_type(chunk)] >
-		 PCPU_EMPTY_POP_PAGES_HIGH + chunk->nr_empty_pop_pages &&
-		chunk->nr_empty_pop_pages >= chunk->nr_pages / 4));
+		(pcpu_nr_empty_pop_pages > PCPU_EMPTY_POP_PAGES_HIGH +
+		 chunk->nr_empty_pop_pages &&
+		 chunk->nr_empty_pop_pages >= chunk->nr_pages / 4));
 }
diff --git a/mm/percpu.c b/mm/percpu.c
index 3135e56ce8d4..e77e3fe45b38 100644
--- a/mm/percpu.c
+++ b/mm/percpu.c
@@ -179,10 +179,10 @@ struct list_head *pcpu_chunk_lists __ro_after_init; /* chunk list slots */
 static LIST_HEAD(pcpu_map_extend_chunks);
 
 /*
- * The number of empty populated pages by chunk type, protected by pcpu_lock.
+ * The number of empty populated pages, protected by pcpu_lock.
  * The reserved chunk doesn't contribute to the count.
  */
-int pcpu_nr_empty_pop_pages[PCPU_NR_CHUNK_TYPES];
+int pcpu_nr_empty_pop_pages;
 
 /*
  * The number of populated pages in use by the allocator, protected by
@@ -532,13 +532,10 @@ static void __pcpu_chunk_move(struct pcpu_chunk *chunk, int slot,
 			      bool move_front)
 {
 	if (chunk != pcpu_reserved_chunk) {
-		struct list_head *pcpu_slot;
-
-		pcpu_slot = pcpu_chunk_list(pcpu_chunk_type(chunk));
 		if (move_front)
-			list_move(&chunk->list, &pcpu_slot[slot]);
+			list_move(&chunk->list, &pcpu_chunk_lists[slot]);
 		else
-			list_move_tail(&chunk->list, &pcpu_slot[slot]);
+			list_move_tail(&chunk->list, &pcpu_chunk_lists[slot]);
 	}
 }
 
@@ -574,27 +571,22 @@ static void pcpu_chunk_relocate(struct pcpu_chunk *chunk, int oslot)
 
 static void pcpu_isolate_chunk(struct pcpu_chunk *chunk)
 {
-	enum pcpu_chunk_type type = pcpu_chunk_type(chunk);
-	struct list_head *pcpu_slot = pcpu_chunk_list(type);
-
 	lockdep_assert_held(&pcpu_lock);
 
 	if (!chunk->isolated) {
 		chunk->isolated = true;
-		pcpu_nr_empty_pop_pages[type] -= chunk->nr_empty_pop_pages;
+		pcpu_nr_empty_pop_pages -= chunk->nr_empty_pop_pages;
 	}
-	list_move(&chunk->list, &pcpu_slot[pcpu_to_depopulate_slot]);
+	list_move(&chunk->list, &pcpu_chunk_lists[pcpu_to_depopulate_slot]);
 }
 
 static void pcpu_reintegrate_chunk(struct pcpu_chunk *chunk)
 {
-	enum pcpu_chunk_type type = pcpu_chunk_type(chunk);
-
 	lockdep_assert_held(&pcpu_lock);
 
 	if (chunk->isolated) {
 		chunk->isolated = false;
-		pcpu_nr_empty_pop_pages[type] += chunk->nr_empty_pop_pages;
+		pcpu_nr_empty_pop_pages += chunk->nr_empty_pop_pages;
 		pcpu_chunk_relocate(chunk, -1);
 	}
 }
@@ -612,7 +604,7 @@ static inline void pcpu_update_empty_pages(struct pcpu_chunk *chunk, int nr)
 {
 	chunk->nr_empty_pop_pages += nr;
 	if (chunk != pcpu_reserved_chunk && !chunk->isolated)
-		pcpu_nr_empty_pop_pages[pcpu_chunk_type(chunk)] += nr;
+		pcpu_nr_empty_pop_pages += nr;
 }
 
 /*
@@ -1405,7 +1397,7 @@ static struct pcpu_chunk * __init pcpu_alloc_first_chunk(unsigned long tmp_addr,
 		      alloc_size);
 
 #ifdef CONFIG_MEMCG_KMEM
-	/* first chunk isn't memcg-aware */
+	/* first chunk is free to use */
 	chunk->obj_cgroups = NULL;
 #endif
 	pcpu_init_md_blocks(chunk);
@@ -1447,7 +1439,7 @@ static struct pcpu_chunk * __init pcpu_alloc_first_chunk(unsigned long tmp_addr,
 	return chunk;
 }
 
-static struct pcpu_chunk *pcpu_alloc_chunk(enum pcpu_chunk_type type, gfp_t gfp)
+static struct pcpu_chunk *pcpu_alloc_chunk(gfp_t gfp)
 {
 	struct pcpu_chunk *chunk;
 	int region_bits;
@@ -1476,7 +1468,7 @@ static struct pcpu_chunk *pcpu_alloc_chunk(enum pcpu_chunk_type type, gfp_t gfp)
 		goto md_blocks_fail;
 
 #ifdef CONFIG_MEMCG_KMEM
-	if (pcpu_is_memcg_chunk(type)) {
+	if (!mem_cgroup_kmem_disabled()) {
 		chunk->obj_cgroups =
 			pcpu_mem_zalloc(pcpu_chunk_map_bits(chunk) *
 					sizeof(struct obj_cgroup *), gfp);
@@ -1589,8 +1581,7 @@ static int pcpu_populate_chunk(struct pcpu_chunk *chunk,
 			       int page_start, int page_end, gfp_t gfp);
 static void pcpu_depopulate_chunk(struct pcpu_chunk *chunk,
 				  int page_start, int page_end);
-static struct pcpu_chunk *pcpu_create_chunk(enum pcpu_chunk_type type,
-					    gfp_t gfp);
+static struct pcpu_chunk *pcpu_create_chunk(gfp_t gfp);
 static void pcpu_destroy_chunk(struct pcpu_chunk *chunk);
 static struct page *pcpu_addr_to_page(void *addr);
 static int __init pcpu_verify_alloc_info(const struct pcpu_alloc_info *ai);
@@ -1633,25 +1624,25 @@ static struct pcpu_chunk *pcpu_chunk_addr_search(void *addr)
 }
 
 #ifdef CONFIG_MEMCG_KMEM
-static enum pcpu_chunk_type pcpu_memcg_pre_alloc_hook(size_t size, gfp_t gfp,
-						     struct obj_cgroup **objcgp)
+static bool pcpu_memcg_pre_alloc_hook(size_t size, gfp_t gfp,
+				      struct obj_cgroup **objcgp)
 {
 	struct obj_cgroup *objcg;
 
 	if (!memcg_kmem_enabled() || !(gfp & __GFP_ACCOUNT))
-		return PCPU_CHUNK_ROOT;
+		return true;
 
 	objcg = get_obj_cgroup_from_current();
 	if (!objcg)
-		return PCPU_CHUNK_ROOT;
+		return true;
 
 	if (obj_cgroup_charge(objcg, gfp, size * num_possible_cpus())) {
 		obj_cgroup_put(objcg);
-		return PCPU_FAIL_ALLOC;
+		return false;
 	}
 
 	*objcgp = objcg;
-	return PCPU_CHUNK_MEMCG;
+	return true;
 }
 
 static void pcpu_memcg_post_alloc_hook(struct obj_cgroup *objcg,
@@ -1661,7 +1652,7 @@ static void pcpu_memcg_post_alloc_hook(struct obj_cgroup *objcg,
 	if (!objcg)
 		return;
 
-	if (chunk) {
+	if (likely(chunk && chunk->obj_cgroups)) {
 		chunk->obj_cgroups[off >> PCPU_MIN_ALLOC_SHIFT] = objcg;
 
 		rcu_read_lock();
@@ -1678,10 +1669,12 @@ static void pcpu_memcg_free_hook(struct pcpu_chunk *chunk, int off, size_t size)
 {
 	struct obj_cgroup *objcg;
 
-	if (!pcpu_is_memcg_chunk(pcpu_chunk_type(chunk)))
+	if (unlikely(!chunk->obj_cgroups))
 		return;
 
 	objcg = chunk->obj_cgroups[off >> PCPU_MIN_ALLOC_SHIFT];
+	if (!objcg)
+		return;
 	chunk->obj_cgroups[off >> PCPU_MIN_ALLOC_SHIFT] = NULL;
 
 	obj_cgroup_uncharge(objcg, size * num_possible_cpus());
@@ -1695,10 +1688,10 @@ static void pcpu_memcg_free_hook(struct pcpu_chunk *chunk, int off, size_t size)
 }
 
 #else /* CONFIG_MEMCG_KMEM */
-static enum pcpu_chunk_type
+static bool
 pcpu_memcg_pre_alloc_hook(size_t size, gfp_t gfp, struct obj_cgroup **objcgp)
 {
-	return PCPU_CHUNK_ROOT;
+	return true;
 }
 
 static void pcpu_memcg_post_alloc_hook(struct obj_cgroup *objcg,
@@ -1733,8 +1726,6 @@ static void __percpu *pcpu_alloc(size_t size, size_t align, bool reserved,
 	gfp_t pcpu_gfp;
 	bool is_atomic;
 	bool do_warn;
-	enum pcpu_chunk_type type;
-	struct list_head *pcpu_slot;
 	struct obj_cgroup *objcg = NULL;
 	static int warn_limit = 10;
 	struct pcpu_chunk *chunk, *next;
@@ -1770,10 +1761,8 @@ static void __percpu *pcpu_alloc(size_t size, size_t align, bool reserved,
 		return NULL;
 	}
 
-	type = pcpu_memcg_pre_alloc_hook(size, gfp, &objcg);
-	if (unlikely(type == PCPU_FAIL_ALLOC))
+	if (unlikely(!pcpu_memcg_pre_alloc_hook(size, gfp, &objcg)))
 		return NULL;
-	pcpu_slot = pcpu_chunk_list(type);
 
 	if (!is_atomic) {
 		/*
@@ -1812,7 +1801,7 @@ static void __percpu *pcpu_alloc(size_t size, size_t align, bool reserved,
 restart:
 	/* search through normal chunks */
 	for (slot = pcpu_size_to_slot(size); slot <= pcpu_free_slot; slot++) {
-		list_for_each_entry_safe(chunk, next, &pcpu_slot[slot], list) {
+		list_for_each_entry_safe(chunk, next, &pcpu_chunk_lists[slot], list) {
 			off = pcpu_find_block_fit(chunk, bits, bit_align,
 						  is_atomic);
 			if (off < 0) {
@@ -1841,8 +1830,8 @@ static void __percpu *pcpu_alloc(size_t size, size_t align, bool reserved,
 		goto fail;
 	}
 
-	if (list_empty(&pcpu_slot[pcpu_free_slot])) {
-		chunk = pcpu_create_chunk(type, pcpu_gfp);
+	if (list_empty(&pcpu_chunk_lists[pcpu_free_slot])) {
+		chunk = pcpu_create_chunk(pcpu_gfp);
 		if (!chunk) {
 			err = "failed to allocate new chunk";
 			goto fail;
@@ -1886,7 +1875,7 @@ static void __percpu *pcpu_alloc(size_t size, size_t align, bool reserved,
 		mutex_unlock(&pcpu_alloc_mutex);
 	}
 
-	if (pcpu_nr_empty_pop_pages[type] < PCPU_EMPTY_POP_PAGES_LOW)
+	if (pcpu_nr_empty_pop_pages < PCPU_EMPTY_POP_PAGES_LOW)
 		pcpu_schedule_balance_work();
 
 	/* clear the areas and return address relative to base address */
@@ -1985,18 +1974,16 @@ void __percpu *__alloc_reserved_percpu(size_t size, size_t align)
 
 /**
  * pcpu_balance_free - manage the amount of free chunks
- * @type: chunk type
  * @empty_only: free chunks only if there are no populated pages
  *
  * If empty_only is %false, reclaim all fully free chunks regardless of the
  * number of populated pages.  Otherwise, only reclaim chunks that have no
  * populated pages.
  */
-static void pcpu_balance_free(enum pcpu_chunk_type type, bool empty_only)
+static void pcpu_balance_free(bool empty_only)
 {
 	LIST_HEAD(to_free);
-	struct list_head *pcpu_slot = pcpu_chunk_list(type);
-	struct list_head *free_head = &pcpu_slot[pcpu_free_slot];
+	struct list_head *free_head = &pcpu_chunk_lists[pcpu_free_slot];
 	struct pcpu_chunk *chunk, *next;
 
 	/*
@@ -2035,7 +2022,6 @@ static void pcpu_balance_free(enum pcpu_chunk_type type, bool empty_only)
 
 /**
  * pcpu_balance_populated - manage the amount of populated pages
- * @type: chunk type
  *
  * Maintain a certain amount of populated pages to satisfy atomic allocations.
  * It is possible that this is called when physical memory is scarce causing
@@ -2043,11 +2029,10 @@ static void pcpu_balance_free(enum pcpu_chunk_type type, bool empty_only)
  * allocation causes the failure as it is possible that requests can be
  * serviced from already backed regions.
  */
-static void pcpu_balance_populated(enum pcpu_chunk_type type)
+static void pcpu_balance_populated(void)
 {
 	/* gfp flags passed to underlying allocators */
 	const gfp_t gfp = GFP_KERNEL | __GFP_NORETRY | __GFP_NOWARN;
-	struct list_head *pcpu_slot = pcpu_chunk_list(type);
 	struct pcpu_chunk *chunk;
 	int slot, nr_to_pop, ret;
 
@@ -2068,7 +2053,7 @@ static void pcpu_balance_populated(enum pcpu_chunk_type type)
 		pcpu_atomic_alloc_failed = false;
 	} else {
 		nr_to_pop = clamp(PCPU_EMPTY_POP_PAGES_HIGH -
-				  pcpu_nr_empty_pop_pages[type],
+				  pcpu_nr_empty_pop_pages,
 				  0, PCPU_EMPTY_POP_PAGES_HIGH);
 	}
 
@@ -2079,7 +2064,7 @@ static void pcpu_balance_populated(enum pcpu_chunk_type type)
 			break;
 
 		spin_lock_irq(&pcpu_lock);
-		list_for_each_entry(chunk, &pcpu_slot[slot], list) {
+		list_for_each_entry(chunk, &pcpu_chunk_lists[slot], list) {
 			nr_unpop = chunk->nr_pages - chunk->nr_populated;
 			if (nr_unpop)
 				break;
@@ -2111,7 +2096,7 @@ static void pcpu_balance_populated(enum pcpu_chunk_type type)
 
 	if (nr_to_pop) {
 		/* ran out of chunks to populate, create a new one and retry */
-		chunk = pcpu_create_chunk(type, gfp);
+		chunk = pcpu_create_chunk(gfp);
 		if (chunk) {
 			spin_lock_irq(&pcpu_lock);
 			pcpu_chunk_relocate(chunk, -1);
@@ -2123,7 +2108,6 @@ static void pcpu_balance_populated(enum pcpu_chunk_type type)
 
 /**
  * pcpu_reclaim_populated - scan over to_depopulate chunks and free empty pages
- * @type: chunk type
  *
  * Scan over chunks in the depopulate list and try to release unused populated
  * pages back to the system.  Depopulated chunks are sidelined to prevent
@@ -2133,9 +2117,8 @@ static void pcpu_balance_populated(enum pcpu_chunk_type type)
  * Each chunk is scanned in the reverse order to keep populated pages close to
  * the beginning of the chunk.
  */
-static void pcpu_reclaim_populated(enum pcpu_chunk_type type)
+static void pcpu_reclaim_populated(void)
 {
-	struct list_head *pcpu_slot = pcpu_chunk_list(type);
 	struct pcpu_chunk *chunk;
 	struct pcpu_block_md *block;
 	int i, end;
@@ -2149,8 +2132,8 @@ static void pcpu_reclaim_populated(enum pcpu_chunk_type type)
 	 * other accessor is the free path which only returns area back to the
 	 * allocator not touching the populated bitmap.
 	 */
-	while (!list_empty(&pcpu_slot[pcpu_to_depopulate_slot])) {
-		chunk = list_first_entry(&pcpu_slot[pcpu_to_depopulate_slot],
+	while (!list_empty(&pcpu_chunk_lists[pcpu_to_depopulate_slot])) {
+		chunk = list_first_entry(&pcpu_chunk_lists[pcpu_to_depopulate_slot],
 					 struct pcpu_chunk, list);
 		WARN_ON(chunk->immutable);
 
@@ -2164,8 +2147,7 @@ static void pcpu_reclaim_populated(enum pcpu_chunk_type type)
 				break;
 
 			/* reintegrate chunk to prevent atomic alloc failures */
-			if (pcpu_nr_empty_pop_pages[type] <
-			    PCPU_EMPTY_POP_PAGES_HIGH) {
+			if (pcpu_nr_empty_pop_pages < PCPU_EMPTY_POP_PAGES_HIGH) {
 				pcpu_reintegrate_chunk(chunk);
 				goto restart;
 			}
@@ -2205,7 +2187,7 @@ static void pcpu_reclaim_populated(enum pcpu_chunk_type type)
 			pcpu_reintegrate_chunk(chunk);
 		else
 			list_move(&chunk->list,
-				  &pcpu_slot[pcpu_sidelined_slot]);
+				  &pcpu_chunk_lists[pcpu_sidelined_slot]);
 	}
 
 	spin_unlock_irq(&pcpu_lock);
@@ -2221,8 +2203,6 @@ static void pcpu_reclaim_populated(enum pcpu_chunk_type type)
  */
 static void pcpu_balance_workfn(struct work_struct *work)
 {
-	enum pcpu_chunk_type type;
-
 	/*
 	 * pcpu_balance_free() is called twice because the first time we may
 	 * trim pages in the active pcpu_nr_empty_pop_pages which may cause us
@@ -2230,14 +2210,12 @@ static void pcpu_balance_workfn(struct work_struct *work)
 	 * to move fully free chunks to the active list to be freed if
 	 * appropriate.
 	 */
-	for (type = 0; type < PCPU_NR_CHUNK_TYPES; type++) {
-		mutex_lock(&pcpu_alloc_mutex);
-		pcpu_balance_free(type, false);
-		pcpu_reclaim_populated(type);
-		pcpu_balance_populated(type);
-		pcpu_balance_free(type, true);
-		mutex_unlock(&pcpu_alloc_mutex);
-	}
+	mutex_lock(&pcpu_alloc_mutex);
+	pcpu_balance_free(false);
+	pcpu_reclaim_populated();
+	pcpu_balance_populated();
+	pcpu_balance_free(true);
+	mutex_unlock(&pcpu_alloc_mutex);
 }
 
 /**
@@ -2256,7 +2234,6 @@ void free_percpu(void __percpu *ptr)
 	unsigned long flags;
 	int size, off;
 	bool need_balance = false;
-	struct list_head *pcpu_slot;
 
 	if (!ptr)
 		return;
@@ -2272,8 +2249,6 @@ void free_percpu(void __percpu *ptr)
 
 	size = pcpu_free_area(chunk, off);
 
-	pcpu_slot = pcpu_chunk_list(pcpu_chunk_type(chunk));
-
 	pcpu_memcg_free_hook(chunk, off, size);
 
 	/*
@@ -2284,7 +2259,7 @@ void free_percpu(void __percpu *ptr)
 	if (!chunk->isolated && chunk->free_bytes == pcpu_unit_size) {
 		struct pcpu_chunk *pos;
 
-		list_for_each_entry(pos, &pcpu_slot[pcpu_free_slot], list)
+		list_for_each_entry(pos, &pcpu_chunk_lists[pcpu_free_slot], list)
 			if (pos != chunk) {
 				need_balance = true;
 				break;
@@ -2592,7 +2567,6 @@ void __init pcpu_setup_first_chunk(const struct pcpu_alloc_info *ai,
 	int map_size;
 	unsigned long tmp_addr;
 	size_t alloc_size;
-	enum pcpu_chunk_type type;
 
 #define PCPU_SETUP_BUG_ON(cond)	do {					\
 	if (unlikely(cond)) {						\
@@ -2716,17 +2690,14 @@ void __init pcpu_setup_first_chunk(const struct pcpu_alloc_info *ai,
 	pcpu_to_depopulate_slot = pcpu_free_slot + 1;
 	pcpu_nr_slots = pcpu_to_depopulate_slot + 1;
 	pcpu_chunk_lists = memblock_alloc(pcpu_nr_slots *
-					  sizeof(pcpu_chunk_lists[0]) *
-					  PCPU_NR_CHUNK_TYPES,
+					  sizeof(pcpu_chunk_lists[0]),
 					  SMP_CACHE_BYTES);
 	if (!pcpu_chunk_lists)
 		panic("%s: Failed to allocate %zu bytes\n", __func__,
-		      pcpu_nr_slots * sizeof(pcpu_chunk_lists[0]) *
-		      PCPU_NR_CHUNK_TYPES);
+		      pcpu_nr_slots * sizeof(pcpu_chunk_lists[0]));
 
-	for (type = 0; type < PCPU_NR_CHUNK_TYPES; type++)
-		for (i = 0; i < pcpu_nr_slots; i++)
-			INIT_LIST_HEAD(&pcpu_chunk_list(type)[i]);
+	for (i = 0; i < pcpu_nr_slots; i++)
+		INIT_LIST_HEAD(&pcpu_chunk_lists[i]);
 
 	/*
 	 * The end of the static region needs to be aligned with the
@@ -2763,7 +2734,7 @@ void __init pcpu_setup_first_chunk(const struct pcpu_alloc_info *ai,
 
 	/* link the first chunk in */
 	pcpu_first_chunk = chunk;
-	pcpu_nr_empty_pop_pages[PCPU_CHUNK_ROOT] = pcpu_first_chunk->nr_empty_pop_pages;
+	pcpu_nr_empty_pop_pages = pcpu_first_chunk->nr_empty_pop_pages;
 	pcpu_chunk_relocate(pcpu_first_chunk, -1);
 
 	/* include all regions of the first chunk */
-- 
2.31.1

Re: [PATCH v2 0/3] percpu: memcg memory accounting rework

[Dennis Zhou]

Hello,

On Wed, Jun 02, 2021 at 06:09:28PM -0700, Roman Gushchin wrote:
> This patchset reworks memcg-based accounting of the percpu memory:
> instead of using two types of chunks (memcg-aware and non-memcg-aware)
> it treats all chunks as memcg-aware. An extra memory used for objcg
> arrays will is compensated by a better chunks utilization, so the
> total memory footprint should be the same or smaller.
> 
> First two patches are preparational changes and cleanups on the memcg side.
> The third one is percpu accounting rework.
> 
> v2:
>   - make all chunks memcg-aware instead of on-demand objcg allocation, by Dennis
> 
> v1:
>   https://lkml.org/lkml/2021/5/11/1343
> 
> 
> Roman Gushchin (3):
>   mm, memcg: mark cgroup_memory_nosocket, nokmem and noswap as
>     __ro_after_init
>   mm, memcg: introduce mem_cgroup_kmem_disabled()
>   percpu: rework memcg accounting
> 
>  include/linux/memcontrol.h |   5 ++
>  mm/memcontrol.c            |  11 ++-
>  mm/percpu-internal.h       |  52 +--------------
>  mm/percpu-km.c             |   5 +-
>  mm/percpu-stats.c          |  46 +++++--------
>  mm/percpu-vm.c             |  11 ++-
>  mm/percpu.c                | 133 +++++++++++++++----------------------
>  7 files changed, 88 insertions(+), 175 deletions(-)
> 
> -- 
> 2.31.1
> 

I think I'm pretty happy with this approach. If there is any issue with
the delta of memcg on but kmem accounting off, we can tackle that
separately. I'll need another day or so to review the last patch a
little more carefully though.

Andrew do you mind if I route the first two in my tree too?

Thanks,
Dennis

Re: [PATCH v2 0/3] percpu: memcg memory accounting rework

[Andrew Morton]

On Fri, 4 Jun 2021 16:10:46 +0000 Dennis Zhou <dennis@kernel.org> wrote:

> I think I'm pretty happy with this approach. If there is any issue with
> the delta of memcg on but kmem accounting off, we can tackle that
> separately. I'll need another day or so to review the last patch a
> little more carefully though.
> 
> Andrew do you mind if I route the first two in my tree too?

Please do.

Re: [PATCH v2 0/3] percpu: memcg memory accounting rework

[Dennis Zhou]

Hello,

On Wed, Jun 02, 2021 at 06:09:28PM -0700, Roman Gushchin wrote:
> This patchset reworks memcg-based accounting of the percpu memory:
> instead of using two types of chunks (memcg-aware and non-memcg-aware)
> it treats all chunks as memcg-aware. An extra memory used for objcg
> arrays will is compensated by a better chunks utilization, so the
> total memory footprint should be the same or smaller.
> 
> First two patches are preparational changes and cleanups on the memcg side.
> The third one is percpu accounting rework.
> 
> v2:
>   - make all chunks memcg-aware instead of on-demand objcg allocation, by Dennis
> 
> v1:
>   https://lkml.org/lkml/2021/5/11/1343
> 
> 
> Roman Gushchin (3):
>   mm, memcg: mark cgroup_memory_nosocket, nokmem and noswap as
>     __ro_after_init
>   mm, memcg: introduce mem_cgroup_kmem_disabled()
>   percpu: rework memcg accounting
> 
>  include/linux/memcontrol.h |   5 ++
>  mm/memcontrol.c            |  11 ++-
>  mm/percpu-internal.h       |  52 +--------------
>  mm/percpu-km.c             |   5 +-
>  mm/percpu-stats.c          |  46 +++++--------
>  mm/percpu-vm.c             |  11 ++-
>  mm/percpu.c                | 133 +++++++++++++++----------------------
>  7 files changed, 88 insertions(+), 175 deletions(-)
> 
> -- 
> 2.31.1
> 

Thanks for this! It greatly simplifies the chunk maintenance.

I've applied this to for-5.14.

There is a trivial merge conflict though due to my for-5.14 being based
on torvalds#v5.12-rc7, but that's fine.

Thanks,
Dennis