memcg: clear pc->mem_cgroup if necessary.

This is a preparation before removing a flag PCG_ACCT_LRU in page_cgroup
and reducing atomic ops/complexity in memcg LRU handling.

In some cases, pages are added to lru before charge to memcg and pages
are not classfied to memory cgroup at lru addtion.  Now, the lru where
the page should be added is determined a bit in page_cgroup->flags and
pc->mem_cgroup.  I'd like to remove the check of flag.

To handle the case pc->mem_cgroup may contain stale pointers if pages
are added to LRU before classification.  This patch resets
pc->mem_cgroup to root_mem_cgroup before lru additions.

[akpm@linux-foundation.org: fix CONFIG_CGROUP_MEM_CONT=n build]
[hughd@google.com: fix CONFIG_CGROUP_MEM_RES_CTLR=y CONFIG_CGROUP_MEM_RES_CTLR_SWAP=n build]
[akpm@linux-foundation.org: ksm.c needs memcontrol.h, per Michal]
[hughd@google.com: stop oops in mem_cgroup_reset_owner()]
[hughd@google.com: fix page migration to reset_owner]
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Miklos Szeredi <mszeredi@suse.cz>
Acked-by: Michal Hocko <mhocko@suse.cz>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Ying Han <yinghan@google.com>
Signed-off-by: Hugh Dickins <hughd@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

4 files changed, 40 insertions, 0 deletions

diff --git a/mm/ksm.c b/mm/ksm.c
index 310544a379a..1925ffbfb27 100644
--- a/mm/ksm.c
+++ b/mm/ksm.c
@@ -28,6 +28,7 @@
 #include <linux/kthread.h>
 #include <linux/wait.h>
 #include <linux/slab.h>
+#include <linux/memcontrol.h>
 #include <linux/rbtree.h>
 #include <linux/memory.h>
 #include <linux/mmu_notifier.h>
@@ -1571,6 +1572,16 @@ struct page *ksm_does_need_to_copy(struct page *page,
 
 	new_page = alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma, address);
 	if (new_page) {
+		/*
+		 * The memcg-specific accounting when moving
+		 * pages around the LRU lists relies on the
+		 * page's owner (memcg) to be valid.  Usually,
+		 * pages are assigned to a new owner before
+		 * being put on the LRU list, but since this
+		 * is not the case here, the stale owner from
+		 * a previous allocation cycle must be reset.
+		 */
+		mem_cgroup_reset_owner(new_page);
 		copy_user_highpage(new_page, page, address, vma);
 
 		SetPageDirty(new_page);
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index d58bb5fa440..c74102d6eb5 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -3050,6 +3050,23 @@ void mem_cgroup_uncharge_end(void)
 	batch->memcg = NULL;
 }
 
+/*
+ * A function for resetting pc->mem_cgroup for newly allocated pages.
+ * This function should be called if the newpage will be added to LRU
+ * before start accounting.
+ */
+void mem_cgroup_reset_owner(struct page *newpage)
+{
+	struct page_cgroup *pc;
+
+	if (mem_cgroup_disabled())
+		return;
+
+	pc = lookup_page_cgroup(newpage);
+	VM_BUG_ON(PageCgroupUsed(pc));
+	pc->mem_cgroup = root_mem_cgroup;
+}
+
 #ifdef CONFIG_SWAP
 /*
  * called after __delete_from_swap_cache() and drop "page" account.
diff --git a/mm/migrate.c b/mm/migrate.c
index 89ea0854332..fc391985899 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -777,6 +777,8 @@ static int unmap_and_move(new_page_t get_new_page, unsigned long private,
 	if (!newpage)
 		return -ENOMEM;
 
+	mem_cgroup_reset_owner(newpage);
+
 	if (page_count(page) == 1) {
 		/* page was freed from under us. So we are done. */
 		goto out;
diff --git a/mm/swap_state.c b/mm/swap_state.c
index ea6b32d6187..470038a9187 100644
--- a/mm/swap_state.c
+++ b/mm/swap_state.c
@@ -300,6 +300,16 @@ struct page *read_swap_cache_async(swp_entry_t entry, gfp_t gfp_mask,
 			new_page = alloc_page_vma(gfp_mask, vma, addr);
 			if (!new_page)
 				break;		/* Out of memory */
+			/*
+			 * The memcg-specific accounting when moving
+			 * pages around the LRU lists relies on the
+			 * page's owner (memcg) to be valid.  Usually,
+			 * pages are assigned to a new owner before
+			 * being put on the LRU list, but since this
+			 * is not the case here, the stale owner from
+			 * a previous allocation cycle must be reset.
+			 */
+			mem_cgroup_reset_owner(new_page);
 		}
 
 		/*