path: root/mm
diff options
authorHugh Dickins <hughd@google.com>2012-03-05 14:59:18 -0800
committerLinus Torvalds <torvalds@linux-foundation.org>2012-03-05 15:49:43 -0800
commit7512102cf64d36e3c7444480273623c7aab3563f (patch)
tree4c5b589455ed0d343384d7eeaedfab5057648a0d /mm
parent9f78ff005a6b6313728247113948450b2adddde8 (diff)
memcg: fix GPF when cgroup removal races with last exit
When moving tasks from old memcg (with move_charge_at_immigrate on new memcg), followed by removal of old memcg, hit General Protection Fault in mem_cgroup_lru_del_list() (called from release_pages called from free_pages_and_swap_cache from tlb_flush_mmu from tlb_finish_mmu from exit_mmap from mmput from exit_mm from do_exit). Somewhat reproducible, takes a few hours: the old struct mem_cgroup has been freed and poisoned by SLAB_DEBUG, but mem_cgroup_lru_del_list() is still trying to update its stats, and take page off lru before freeing. A task, or a charge, or a page on lru: each secures a memcg against removal. In this case, the last task has been moved out of the old memcg, and it is exiting: anonymous pages are uncharged one by one from the memcg, as they are zapped from its pagetables, so the charge gets down to 0; but the pages themselves are queued in an mmu_gather for freeing. Most of those pages will be on lru (and force_empty is careful to lru_add_drain_all, to add pages from pagevec to lru first), but not necessarily all: perhaps some have been isolated for page reclaim, perhaps some isolated for other reasons. So, force_empty may find no task, no charge and no page on lru, and let the removal proceed. There would still be no problem if these pages were immediately freed; but typically (and the put_page_testzero protocol demands it) they have to be added back to lru before they are found freeable, then removed from lru and freed. We don't see the issue when adding, because the mem_cgroup_iter() loops keep their own reference to the memcg being scanned; but when it comes to mem_cgroup_lru_del_list(). I believe this was not an issue in v3.2: there, PageCgroupAcctLRU and PageCgroupUsed flags were used (like a trick with mirrors) to deflect view of pc->mem_cgroup to the stable root_mem_cgroup when neither set. 38c5d72f3ebe ("memcg: simplify LRU handling by new rule") mercifully removed those convolutions, but left this General Protection Fault. But it's surprisingly easy to restore the old behaviour: just check PageCgroupUsed in mem_cgroup_lru_add_list() (which decides on which lruvec to add), and reset pc to root_mem_cgroup if page is uncharged. A risky change? just going back to how it worked before; testing, and an audit of uses of pc->mem_cgroup, show no problem. And there's a nice bonus: with mem_cgroup_lru_add_list() itself making sure that an uncharged page goes to root lru, mem_cgroup_reset_owner() no longer has any purpose, and we can safely revert 4e5f01c2b9b9 ("memcg: clear pc->mem_cgroup if necessary"). Calling update_page_reclaim_stat() after add_page_to_lru_list() in swap.c is not strictly necessary: the lru_lock there, with RCU before memcg structures are freed, makes mem_cgroup_get_reclaim_stat_from_page safe without that; but it seems cleaner to rely on one dependency less. Signed-off-by: Hugh Dickins <hughd@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Konstantin Khlebnikov <khlebnikov@openvz.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Diffstat (limited to 'mm')
5 files changed, 18 insertions, 43 deletions
diff --git a/mm/ksm.c b/mm/ksm.c
index 1925ffbfb27..310544a379a 100644
--- a/mm/ksm.c
+++ b/mm/ksm.c
@@ -28,7 +28,6 @@
#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>
@@ -1572,16 +1571,6 @@ 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);
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 1097d8098f8..d0e57a3cda1 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -1042,6 +1042,19 @@ struct lruvec *mem_cgroup_lru_add_list(struct zone *zone, struct page *page,
pc = lookup_page_cgroup(page);
memcg = pc->mem_cgroup;
+ /*
+ * Surreptitiously switch any uncharged page to root:
+ * an uncharged page off lru does nothing to secure
+ * its former mem_cgroup from sudden removal.
+ *
+ * Our caller holds lru_lock, and PageCgroupUsed is updated
+ * under page_cgroup lock: between them, they make all uses
+ * of pc->mem_cgroup safe.
+ */
+ if (!PageCgroupUsed(pc) && memcg != root_mem_cgroup)
+ pc->mem_cgroup = memcg = root_mem_cgroup;
mz = page_cgroup_zoneinfo(memcg, page);
/* compound_order() is stabilized through lru_lock */
MEM_CGROUP_ZSTAT(mz, lru) += 1 << compound_order(page);
@@ -3029,23 +3042,6 @@ 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;
* called after __delete_from_swap_cache() and drop "page" account.
diff --git a/mm/migrate.c b/mm/migrate.c
index df141f60289..1503b6b54ec 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -839,8 +839,6 @@ 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.c b/mm/swap.c
index fff1ff7fb9a..14380e9fbe3 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -652,7 +652,7 @@ EXPORT_SYMBOL(__pagevec_release);
void lru_add_page_tail(struct zone* zone,
struct page *page, struct page *page_tail)
- int active;
+ int uninitialized_var(active);
enum lru_list lru;
const int file = 0;
@@ -672,7 +672,6 @@ void lru_add_page_tail(struct zone* zone,
active = 0;
- update_page_reclaim_stat(zone, page_tail, file, active);
} else {
@@ -693,6 +692,9 @@ void lru_add_page_tail(struct zone* zone,
list_head = page_tail->lru.prev;
list_move_tail(&page_tail->lru, list_head);
+ if (!PageUnevictable(page))
+ update_page_reclaim_stat(zone, page_tail, file, active);
@@ -710,8 +712,8 @@ static void __pagevec_lru_add_fn(struct page *page, void *arg)
if (active)
- update_page_reclaim_stat(zone, page, file, active);
add_page_to_lru_list(zone, page, lru);
+ update_page_reclaim_stat(zone, page, file, active);
diff --git a/mm/swap_state.c b/mm/swap_state.c
index 470038a9187..ea6b32d6187 100644
--- a/mm/swap_state.c
+++ b/mm/swap_state.c
@@ -300,16 +300,6 @@ 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);