lkml.org 
[lkml]   [2014]   [Feb]   [4]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
From
Subject[PATCH -v2 5/6] memcg, kmem: clean up memcg parameter handling
Date
memcg_kmem_newpage_charge doesn't always set the given memcg parameter.
Some early escape paths skip setting *memcg while
__memcg_kmem_newpage_charge down the call chain sets *memcg even if no
memcg is charged due to other escape paths.

The current code is correct because the memcg is initialized to NULL
at the highest level in __alloc_pages_nodemask but this all is very
confusing and error prone. Let's make the semantic clear and move the
memcg parameter initialization to the highest level of kmem accounting
(memcg_kmem_newpage_charge).

Signed-off-by: Michal Hocko <mhocko@suse.cz>
---
include/linux/memcontrol.h | 4 +++-
mm/memcontrol.c | 2 --
mm/page_alloc.c | 2 +-
3 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h
index abd0113b6620..7bcb39668917 100644
--- a/include/linux/memcontrol.h
+++ b/include/linux/memcontrol.h
@@ -522,11 +522,13 @@ void kmem_cache_destroy_memcg_children(struct kmem_cache *s);
* allocation.
*
* We return true automatically if this allocation is not to be accounted to
- * any memcg.
+ * any memcg when *memcg is set to NULL.
*/
static inline bool
memcg_kmem_newpage_charge(gfp_t gfp, struct mem_cgroup **memcg, int order)
{
+ *memcg = NULL;
+
if (!memcg_kmem_enabled())
return true;

diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index d06743a9a765..46b9f461cedf 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -3637,8 +3637,6 @@ __memcg_kmem_newpage_charge(gfp_t gfp, struct mem_cgroup **_memcg, int order)
struct mem_cgroup *memcg;
int ret;

- *_memcg = NULL;
-
/*
* Disabling accounting is only relevant for some specific memcg
* internal allocations. Therefore we would initially not have such
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index e3758a09a009..6f6099d38772 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -2692,7 +2692,7 @@ __alloc_pages_nodemask(gfp_t gfp_mask, unsigned int order,
int migratetype = allocflags_to_migratetype(gfp_mask);
unsigned int cpuset_mems_cookie;
int alloc_flags = ALLOC_WMARK_LOW|ALLOC_CPUSET;
- struct mem_cgroup *memcg = NULL;
+ struct mem_cgroup *memcg;

gfp_mask &= gfp_allowed_mask;

--
1.9.rc1


\
 
 \ /
  Last update: 2014-02-04 16:01    [W:0.522 / U:0.096 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site