[PATCH 5.10 61/95] mm, slub: fix incorrect memcg slab count for bulk free

From: Greg Kroah-Hartman
Date: Mon Oct 25 2021 - 15:42:23 EST


From: Miaohe Lin <linmiaohe@xxxxxxxxxx>

commit 3ddd60268c24bcac9d744404cc277e9dc52fe6b6 upstream.

kmem_cache_free_bulk() will call memcg_slab_free_hook() for all objects
when doing bulk free. So we shouldn't call memcg_slab_free_hook() again
for bulk free to avoid incorrect memcg slab count.

Link: https://lkml.kernel.org/r/20210916123920.48704-6-linmiaohe@xxxxxxxxxx
Fixes: d1b2cf6cb84a ("mm: memcg/slab: uncharge during kmem_cache_free_bulk()")
Signed-off-by: Miaohe Lin <linmiaohe@xxxxxxxxxx>
Reviewed-by: Vlastimil Babka <vbabka@xxxxxxx>
Cc: Andrey Konovalov <andreyknvl@xxxxxxxxx>
Cc: Andrey Ryabinin <ryabinin.a.a@xxxxxxxxx>
Cc: Bharata B Rao <bharata@xxxxxxxxxxxxx>
Cc: Christoph Lameter <cl@xxxxxxxxx>
Cc: David Rientjes <rientjes@xxxxxxxxxx>
Cc: Faiyaz Mohammed <faiyazm@xxxxxxxxxxxxxx>
Cc: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
Cc: Joonsoo Kim <iamjoonsoo.kim@xxxxxxx>
Cc: Kees Cook <keescook@xxxxxxxxxxxx>
Cc: Pekka Enberg <penberg@xxxxxxxxxx>
Cc: Roman Gushchin <guro@xxxxxx>
Cc: <stable@xxxxxxxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
Signed-off-by: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
mm/slub.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)

--- a/mm/slub.c
+++ b/mm/slub.c
@@ -3100,7 +3100,9 @@ static __always_inline void do_slab_free
struct kmem_cache_cpu *c;
unsigned long tid;

- memcg_slab_free_hook(s, &head, 1);
+ /* memcg_slab_free_hook() is already called for bulk free. */
+ if (!tail)
+ memcg_slab_free_hook(s, &head, 1);
redo:
/*
* Determine the currently cpus per cpu slab.