The patch titled
memcg: fix shmem's swap accounting
has been added to the -mm tree. Its filename is
memcg-fix-shmems-swap-accounting.patch
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/SubmitChecklist when testing your code ***
See http://userweb.kernel.org/~akpm/stuff/added-to-mm.txt to find
out what to do about this
The current -mm tree may be found at http://userweb.kernel.org/~akpm/mmotm/
------------------------------------------------------
Subject: memcg: fix shmem's swap accounting
From: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>
Now, you can see following even when swap accounting is enabled.
1. Create Group 01, and 02.
2. allocate a "file" on tmpfs by a task under 01.
3. swap out the "file" (by memory pressure)
4. Read "file" from a task in group 02.
5. the charge of "file" is moved to group 02.
This is not ideal behavior. This is because SwapCache which was loaded
by read-ahead is not taken into account..
This is a patch to fix shmem's swapcache behavior.
- remove mem_cgroup_cache_charge_swapin().
- Add SwapCache handler routine to mem_cgroup_cache_charge().
By this, shmem's file cache is charged at add_to_page_cache()
with GFP_NOWAIT.
- pass the page of swapcache to shrink_mem_cgroup.
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>
Cc: Daisuke Nishimura <nishimura@xxxxxxxxxxxxxxxxx>
Cc: Balbir Singh <balbir@xxxxxxxxxx>
Cc: Paul Menage <menage@xxxxxxxxxx>
Cc: Li Zefan <lizf@xxxxxxxxxxxxxx>
Cc: Hugh Dickins <hugh@xxxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
include/linux/memcontrol.h | 6 +
include/linux/swap.h | 8 --
mm/memcontrol.c | 134 +++++++++++++++--------------------
mm/shmem.c | 30 +++----
4 files changed, 76 insertions(+), 102 deletions(-)
diff -puN include/linux/memcontrol.h~memcg-fix-shmems-swap-accounting include/linux/memcontrol.h
--- a/include/linux/memcontrol.h~memcg-fix-shmems-swap-accounting
+++ a/include/linux/memcontrol.h
@@ -56,7 +56,8 @@ extern void mem_cgroup_move_lists(struct
enum lru_list from, enum lru_list to);
extern void mem_cgroup_uncharge_page(struct page *page);
extern void mem_cgroup_uncharge_cache_page(struct page *page);
-extern int mem_cgroup_shrink_usage(struct mm_struct *mm, gfp_t gfp_mask);
+extern int mem_cgroup_shrink_usage(struct page *page,
+ struct mm_struct *mm, gfp_t gfp_mask);
extern unsigned long mem_cgroup_isolate_pages(unsigned long nr_to_scan,
struct list_head *dst,
@@ -155,7 +156,8 @@ static inline void mem_cgroup_uncharge_c
{
}
-static inline int mem_cgroup_shrink_usage(struct mm_struct *mm, gfp_t gfp_mask)
+static inline int mem_cgroup_shrink_usage(struct page *page,
+ struct mm_struct *mm, gfp_t gfp_mask)
{
return 0;
}
diff -puN include/linux/swap.h~memcg-fix-shmems-swap-accounting include/linux/swap.h
--- a/include/linux/swap.h~memcg-fix-shmems-swap-accounting
+++ a/include/linux/swap.h
@@ -338,16 +338,8 @@ static inline void disable_swap_token(vo
}
#ifdef CONFIG_CGROUP_MEM_RES_CTLR
-extern int mem_cgroup_cache_charge_swapin(struct page *page,
- struct mm_struct *mm, gfp_t mask, bool locked);
extern void mem_cgroup_uncharge_swapcache(struct page *page, swp_entry_t ent);
#else
-static inline
-int mem_cgroup_cache_charge_swapin(struct page *page,
- struct mm_struct *mm, gfp_t mask, bool locked)
-{
- return 0;
-}
static inline void
mem_cgroup_uncharge_swapcache(struct page *page, swp_entry_t ent)
{
diff -puN mm/memcontrol.c~memcg-fix-shmems-swap-accounting mm/memcontrol.c
--- a/mm/memcontrol.c~memcg-fix-shmems-swap-accounting
+++ a/mm/memcontrol.c
@@ -888,6 +888,23 @@ nomem:
return -ENOMEM;
}
+static struct mem_cgroup *try_get_mem_cgroup_from_swapcache(struct page *page)
+{
+ struct mem_cgroup *mem;
+ swp_entry_t ent;
+
+ if (!PageSwapCache(page))
+ return NULL;
+
+ ent.val = page_private(page);
+ mem = lookup_swap_cgroup(ent);
+ if (!mem)
+ return NULL;
+ if (!css_tryget(&mem->css))
+ return NULL;
+ return mem;
+}
+
/*
* commit a charge got by __mem_cgroup_try_charge() and makes page_cgroup to be
* USED state. If already USED, uncharge and return.
@@ -1079,6 +1096,9 @@ int mem_cgroup_newpage_charge(struct pag
int mem_cgroup_cache_charge(struct page *page, struct mm_struct *mm,
gfp_t gfp_mask)
{
+ struct mem_cgroup *mem = NULL;
+ int ret;
+
if (mem_cgroup_disabled())
return 0;
if (PageCompound(page))
@@ -1091,6 +1111,8 @@ int mem_cgroup_cache_charge(struct page
* For GFP_NOWAIT case, the page may be pre-charged before calling
* add_to_page_cache(). (See shmem.c) check it here and avoid to call
* charge twice. (It works but has to pay a bit larger cost.)
+ * And when the page is SwapCache, it should take swap information
+ * into account. This is under lock_page() now.
*/
if (!(gfp_mask & __GFP_WAIT)) {
struct page_cgroup *pc;
@@ -1107,15 +1129,40 @@ int mem_cgroup_cache_charge(struct page
unlock_page_cgroup(pc);
}
- if (unlikely(!mm))
+ if (do_swap_account && PageSwapCache(page)) {
+ mem = try_get_mem_cgroup_from_swapcache(page);
+ if (mem)
+ mm = NULL;
+ else
+ mem = NULL;
+ /* SwapCache may be still linked to LRU now. */
+ mem_cgroup_lru_del_before_commit_swapcache(page);
+ }
+
+ if (unlikely(!mm && !mem))
mm = &init_mm;
if (page_is_file_cache(page))
return mem_cgroup_charge_common(page, mm, gfp_mask,
MEM_CGROUP_CHARGE_TYPE_CACHE, NULL);
- else
- return mem_cgroup_charge_common(page, mm, gfp_mask,
- MEM_CGROUP_CHARGE_TYPE_SHMEM, NULL);
+
+ ret = mem_cgroup_charge_common(page, mm, gfp_mask,
+ MEM_CGROUP_CHARGE_TYPE_SHMEM, mem);
+ if (mem)
+ css_put(&mem->css);
+ if (PageSwapCache(page))
+ mem_cgroup_lru_add_after_commit_swapcache(page);
+
+ if (do_swap_account && !ret && PageSwapCache(page)) {
+ swp_entry_t ent = {.val = page_private(page)};
+ /* avoid double counting */
+ mem = swap_cgroup_record(ent, NULL);
+ if (mem) {
+ res_counter_uncharge(&mem->memsw, PAGE_SIZE);
+ mem_cgroup_put(mem);
+ }
+ }
+ return ret;
}
/*
@@ -1129,7 +1176,6 @@ int mem_cgroup_try_charge_swapin(struct
gfp_t mask, struct mem_cgroup **ptr)
{
struct mem_cgroup *mem;
- swp_entry_t ent;
int ret;
if (mem_cgroup_disabled())
@@ -1137,7 +1183,6 @@ int mem_cgroup_try_charge_swapin(struct
if (!do_swap_account)
goto charge_cur_mm;
-
/*
* A racing thread's fault, or swapoff, may have already updated
* the pte, and even removed page from swap cache: return success
@@ -1145,14 +1190,9 @@ int mem_cgroup_try_charge_swapin(struct
*/
if (!PageSwapCache(page))
return 0;
-
- ent.val = page_private(page);
-
- mem = lookup_swap_cgroup(ent);
+ mem = try_get_mem_cgroup_from_swapcache(page);
if (!mem)
goto charge_cur_mm;
- if (!css_tryget(&mem->css))
- goto charge_cur_mm;
*ptr = mem;
ret = __mem_cgroup_try_charge(NULL, mask, ptr, true);
/* drop extra refcnt from tryget */
@@ -1164,62 +1204,6 @@ charge_cur_mm:
return __mem_cgroup_try_charge(mm, mask, ptr, true);
}
-#ifdef CONFIG_SWAP
-
-int mem_cgroup_cache_charge_swapin(struct page *page,
- struct mm_struct *mm, gfp_t mask, bool locked)
-{
- int ret = 0;
-
- if (mem_cgroup_disabled())
- return 0;
- if (unlikely(!mm))
- mm = &init_mm;
- if (!locked)
- lock_page(page);
- /*
- * If not locked, the page can be dropped from SwapCache until
- * we reach here.
- */
- if (PageSwapCache(page)) {
- struct mem_cgroup *mem = NULL;
- swp_entry_t ent;
-
- ent.val = page_private(page);
- if (do_swap_account) {
- mem = lookup_swap_cgroup(ent);
- if (mem) {
- if (css_tryget(&mem->css))
- mm = NULL; /* charge to recorded */
- else
- mem = NULL; /* charge to current */
- }
- }
- /* SwapCache may be still linked to LRU now. */
- mem_cgroup_lru_del_before_commit_swapcache(page);
- ret = mem_cgroup_charge_common(page, mm, mask,
- MEM_CGROUP_CHARGE_TYPE_SHMEM, mem);
- mem_cgroup_lru_add_after_commit_swapcache(page);
- /* drop extra refcnt from tryget */
- if (mem)
- css_put(&mem->css);
-
- if (!ret && do_swap_account) {
- /* avoid double counting */
- mem = swap_cgroup_record(ent, NULL);
- if (mem) {
- res_counter_uncharge(&mem->memsw, PAGE_SIZE);
- mem_cgroup_put(mem);
- }
- }
- }
- if (!locked)
- unlock_page(page);
-
- return ret;
-}
-#endif
-
void mem_cgroup_commit_charge_swapin(struct page *page, struct mem_cgroup *ptr)
{
struct page_cgroup *pc;
@@ -1481,18 +1465,20 @@ void mem_cgroup_end_migration(struct mem
* This is typically used for page reclaiming for shmem for reducing side
* effect of page allocation from shmem, which is used by some mem_cgroup.
*/
-int mem_cgroup_shrink_usage(struct mm_struct *mm, gfp_t gfp_mask)
+int mem_cgroup_shrink_usage(struct page *page,
+ struct mm_struct *mm,
+ gfp_t gfp_mask)
{
- struct mem_cgroup *mem;
+ struct mem_cgroup *mem = NULL;
int progress = 0;
int retry = MEM_CGROUP_RECLAIM_RETRIES;
if (mem_cgroup_disabled())
return 0;
- if (!mm)
- return 0;
-
- mem = try_get_mem_cgroup_from_mm(mm);
+ if (page)
+ mem = try_get_mem_cgroup_from_swapcache(page);
+ if (!mem && mm)
+ mem = try_get_mem_cgroup_from_mm(mm);
if (unlikely(!mem))
return 0;
diff -puN mm/shmem.c~memcg-fix-shmems-swap-accounting mm/shmem.c
--- a/mm/shmem.c~memcg-fix-shmems-swap-accounting
+++ a/mm/shmem.c
@@ -929,11 +929,11 @@ found:
if (!inode)
goto out;
/*
- * Charge page using GFP_HIGHUSER_MOVABLE while we can wait.
- * charged back to the user(not to caller) when swap account is used.
+ * Charge page using GFP_KERNEL while we can wait.
+ * Charged back to the user(not to caller) when swap account is used.
+ * add_to_page_cache() will be called with GFP_NOWAIT.
*/
- error = mem_cgroup_cache_charge_swapin(page, current->mm, GFP_KERNEL,
- true);
+ error = mem_cgroup_cache_charge(page, current->mm, GFP_KERNEL);
if (error)
goto out;
error = radix_tree_preload(GFP_KERNEL);
@@ -1270,16 +1270,6 @@ repeat:
goto repeat;
}
wait_on_page_locked(swappage);
- /*
- * We want to avoid charge at add_to_page_cache().
- * charge against this swap cache here.
- */
- if (mem_cgroup_cache_charge_swapin(swappage,
- current->mm, gfp & GFP_RECLAIM_MASK, false)) {
- page_cache_release(swappage);
- error = -ENOMEM;
- goto failed;
- }
page_cache_release(swappage);
goto repeat;
}
@@ -1334,15 +1324,19 @@ repeat:
} else {
shmem_swp_unmap(entry);
spin_unlock(&info->lock);
- unlock_page(swappage);
- page_cache_release(swappage);
if (error == -ENOMEM) {
/* allow reclaim from this memory cgroup */
- error = mem_cgroup_shrink_usage(current->mm,
+ error = mem_cgroup_shrink_usage(swappage,
+ current->mm,
gfp);
- if (error)
+ if (error) {
+ unlock_page(swappage);
+ page_cache_release(swappage);
goto failed;
+ }
}
+ unlock_page(swappage);
+ page_cache_release(swappage);
goto repeat;
}
} else if (sgp == SGP_READ && !filepage) {
_
Patches currently in -mm which might be from kamezawa.hiroyu@xxxxxxxxxxxxxx are
mm-gup-persist-for-write-permission.patch
mm-wp-lock-page-before-deciding-cow.patch
mm-reuse_swap_page-replaces-can_share_swap_page.patch
mm-try_to_free_swap-replaces-remove_exclusive_swap_page.patch
mm-try_to_unuse-check-removing-right-swap.patch
mm-remove-try_to_munlock-from-vmscan.patch
mm-remove-gfp_mask-from-add_to_swap.patch
mm-add-add_to_swap-stub.patch
mm-optimize-get_scan_ratio-for-no-swap.patch
memcg-reclaim-shouldnt-change-zone-recent_rotated-statistics.patch
swapfile-swapon-needs-larger-size-type.patch
swapfile-remove-swp_active-mask.patch
swapfile-remove-surplus-whitespace.patch
swapfile-remove-v0-swap-space-message.patch
swapfile-rearrange-scan-and-swap_info.patch
swapfile-swapon-use-discard-trim.patch
swapfile-swap-allocation-use-discard.patch
swapfile-swapon-randomize-if-nonrot.patch
swapfile-swap-allocation-cycle-if-nonrot.patch
swapfile-change-discard-pgoff_t-to-sector_t.patch
swapfile-change-discard-pgoff_t-to-sector_t-fix.patch
swapfile-let-others-seed-random.patch
mm-remove-config_out_of_line_pfn_to_page.patch
cgroups-make-cgroup-config-a-submenu.patch
cgroups-documentation-updates.patch
cgroups-remove-some-redundant-null-checks.patch
ns_cgroup-remove-unused-spinlock.patch
memcg-fix-a-typo-in-kconfig.patch
cgroups-add-lock-for-child-cgroups-in-cgroup_post_fork.patch
cgroups-fix-cgroup_iter_next-bug.patch
cgroups-dont-put-struct-cgroupfs_root-protected-by-rcu.patch
cgroups-use-task_lock-for-access-tsk-cgroups-safe-in-cgroup_clone.patch
cgroups-call-find_css_set-safely-in-cgroup_attach_task.patch
cgroups-remove-rcu_read_lock-in-cgroupstats_build.patch
cgroups-make-root_list-contains-active-hierarchies-only.patch
cgroups-add-inactive-subsystems-to-rootnodesubsys_list.patch
cgroups-add-inactive-subsystems-to-rootnodesubsys_list-fix.patch
cgroups-introduce-link_css_set-to-remove-duplicate-code.patch
cgroups-introduce-link_css_set-to-remove-duplicate-code-fix.patch
devcgroup-use-list_for_each_entry_rcu.patch
memcg-introduce-charge-commit-cancel-style-of-functions.patch
memcg-introduce-charge-commit-cancel-style-of-functions-fix.patch
memcg-fix-gfp_mask-of-callers-of-charge.patch
memcg-simple-migration-handling.patch
memcg-do-not-recalculate-section-unnecessarily-in-init_section_page_cgroup.patch
memcg-move-all-acccounts-to-parent-at-rmdir.patch
memcg-reduce-size-of-mem_cgroup-by-using-nr_cpu_ids.patch
memcg-new-force_empty-to-free-pages-under-group.patch
memcg-new-force_empty-to-free-pages-under-group-fix.patch
memcg-new-force_empty-to-free-pages-under-group-fix-fix.patch
memcg-handle-swap-caches.patch
memcg-handle-swap-caches-build-fix.patch
memcg-memswap-controller-kconfig.patch
memcg-swap-cgroup-for-remembering-usage.patch
memcg-memswap-controller-core.patch
memcg-memswap-controller-core-make-resize-limit-hold-mutex.patch
memcg-memswap-controller-core-swapcache-fixes.patch
memcg-synchronized-lru.patch
memcg-add-mem_cgroup_disabled.patch
memcg-add-mem_cgroup_disabled-fix.patch
memory-cgroup-hierarchy-documentation-v4.patch
memory-cgroup-resource-counters-for-hierarchy-v4.patch
memory-cgroup-resource-counters-for-hierarchy-v4-checkpatch-fixes.patch
memory-cgroup-hierarchical-reclaim-v4.patch
memory-cgroup-hierarchical-reclaim-v4-checkpatch-fixes.patch
memory-cgroup-hierarchical-reclaim-v4-fix-for-hierarchical-reclaim.patch
memory-cgroup-hierarchy-feature-selector-v4.patch
memory-cgroup-hierarchy-feature-selector-v4-fix.patch
memcontrol-rcu_read_lock-to-protect-mm_match_cgroup.patch
memcg-avoid-unnecessary-system-wide-oom-killer.patch
memcg-avoid-unnecessary-system-wide-oom-killer-fix.patch
memcg-fix-reclaim-result-checks.patch
memcg-revert-gfp-mask-fix.patch
memcg-check-group-leader-fix.patch
memcg-memoryswap-controller-fix-limit-check.patch
memcg-swapout-refcnt-fix.patch
memcg-hierarchy-avoid-unnecessary-reclaim.patch
inactive_anon_is_low-move-to-vmscan.patch
mm-introduce-zone_reclaim-struct.patch
mm-add-zone-nr_pages-helper-function.patch
memcg-add-null-check-to-page_cgroup_zoneinfo.patch
memcg-add-inactive_anon_is_low.patch
memcg-add-mem_cgroup_zone_nr_pages.patch
memcg-add-zone_reclaim_stat.patch
memcg-show-reclaim-stat.patch
memcg-rename-scan-global-lru.patch
memcg-protect-prev_priority.patch
memcg-swappiness.patch
memcg-fix-calclation-of-active_ratio.patch
memcg-fix-calclation-of-active_ratio-build-error-fix.patch
memcg-show-real-limit-under-hierarchy-mode.patch
memcg-explain-details-and-test-document.patch
memcg-explain-details-and-test-document-fix.patch
memcg-dont-trigger-oom-at-page-migration.patch
memcg-remove-mem_cgroup_try_charge.patch
memcg-avoid-dead-lock-caused-by-race-between-oom-and-cpuset_attach.patch
memcg-change-try_to_free_pages-to-hierarchical_reclaim.patch
memcg-fix-swap-accounting-leak-v3.patch
memcg-fix-swap-accounting-leak-doc-fix.patch
memcg-fix-double-free-and-make-refcnt-sane.patch
memcg-use-css_tryget-in-memcg.patch
memcg-use-css_tryget-in-memcg-fix.patch
memcg-fix-lru-accounting-for-swapcache-v2.patch
memcg-fix-shmems-swap-accounting.patch
cgroups-add-a-per-subsystem-hierarchy_mutex.patch
cgroups-use-hierarchy_mutex-in-memory-controller.patch
cgroups-add-css_tryget.patch
cpuset-rcu_read_lock-to-protect-task_cs.patch
--
To unsubscribe from this list: send the line "unsubscribe mm-commits" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
