memcg: add comments explaining memory barriers
authorKAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Wed, 23 Sep 2009 22:56:33 +0000 (15:56 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Thu, 24 Sep 2009 14:20:58 +0000 (07:20 -0700)
Add comments for the reason of smp_wmb() in mem_cgroup_commit_charge().

[akpm@linux-foundation.org: coding-style fixes]
Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Cc: Balbir Singh <balbir@linux.vnet.ibm.com>
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/memcontrol.c

index b075766..eb95718 100644 (file)
@@ -1134,6 +1134,13 @@ static void __mem_cgroup_commit_charge(struct mem_cgroup *mem,
        }
 
        pc->mem_cgroup = mem;
+       /*
+        * We access a page_cgroup asynchronously without lock_page_cgroup().
+        * Especially when a page_cgroup is taken from a page, pc->mem_cgroup
+        * is accessed after testing USED bit. To make pc->mem_cgroup visible
+        * before USED bit, we need memory barrier here.
+        * See mem_cgroup_add_lru_list(), etc.
+        */
        smp_wmb();
        switch (ctype) {
        case MEM_CGROUP_CHARGE_TYPE_CACHE: