mm: memcontrol: remove unnecessary PCG_USED pc->mem_cgroup valid flag
authorJohannes Weiner <hannes@cmpxchg.org>
Wed, 10 Dec 2014 23:44:02 +0000 (15:44 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Thu, 11 Dec 2014 01:41:07 +0000 (17:41 -0800)
pc->mem_cgroup had to be left intact after uncharge for the final LRU
removal, and !PCG_USED indicated whether the page was uncharged.  But
since commit 0a31bc97c80c ("mm: memcontrol: rewrite uncharge API") pages
are uncharged after the final LRU removal.  Uncharge can simply clear
the pointer and the PCG_USED/PageCgroupUsed sites can test that instead.

Because this is the last page_cgroup flag, this patch reduces the memcg
per-page overhead to a single pointer.

[akpm@linux-foundation.org: remove unneeded initialization of `memcg', per Michal]
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: Hugh Dickins <hughd@google.com>
Acked-by: Michal Hocko <mhocko@suse.cz>
Reviewed-by: Vladimir Davydov <vdavydov@parallels.com>
Acked-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>
include/linux/page_cgroup.h
mm/memcontrol.c

index 97536e6858430064f85dc71d509463582fb460bd..1289be6b436c54619a705b3dcbe51eed62ad6b23 100644 (file)
@@ -1,11 +1,6 @@
 #ifndef __LINUX_PAGE_CGROUP_H
 #define __LINUX_PAGE_CGROUP_H
 
-enum {
-       /* flags for mem_cgroup */
-       PCG_USED = 0x01,        /* This page is charged to a memcg */
-};
-
 struct pglist_data;
 
 #ifdef CONFIG_MEMCG
@@ -19,7 +14,6 @@ struct mem_cgroup;
  * then the page cgroup for pfn always exists.
  */
 struct page_cgroup {
-       unsigned long flags;
        struct mem_cgroup *mem_cgroup;
 };
 
@@ -39,10 +33,6 @@ static inline void page_cgroup_init(void)
 
 struct page_cgroup *lookup_page_cgroup(struct page *page);
 
-static inline int PageCgroupUsed(struct page_cgroup *pc)
-{
-       return !!(pc->flags & PCG_USED);
-}
 #else /* !CONFIG_MEMCG */
 struct page_cgroup;
 
index 3dfb56a9311730f789b59afeb0dfea8f0b25a32b..09fece0eb9f156404fd06832fb3e924dcb13d20a 100644 (file)
@@ -1284,14 +1284,12 @@ struct lruvec *mem_cgroup_page_lruvec(struct page *page, struct zone *zone)
 
        pc = lookup_page_cgroup(page);
        memcg = pc->mem_cgroup;
-
        /*
         * Swapcache readahead pages are added to the LRU - and
-        * possibly migrated - before they are charged.  Ensure
-        * pc->mem_cgroup is sane.
+        * possibly migrated - before they are charged.
         */
-       if (!PageLRU(page) && !PageCgroupUsed(pc) && memcg != root_mem_cgroup)
-               pc->mem_cgroup = memcg = root_mem_cgroup;
+       if (!memcg)
+               memcg = root_mem_cgroup;
 
        mz = mem_cgroup_page_zoneinfo(memcg, page);
        lruvec = &mz->lruvec;
@@ -2151,7 +2149,7 @@ struct mem_cgroup *mem_cgroup_begin_page_stat(struct page *page,
        pc = lookup_page_cgroup(page);
 again:
        memcg = pc->mem_cgroup;
-       if (unlikely(!memcg || !PageCgroupUsed(pc)))
+       if (unlikely(!memcg))
                return NULL;
 
        *locked = false;
@@ -2159,7 +2157,7 @@ again:
                return memcg;
 
        move_lock_mem_cgroup(memcg, flags);
-       if (memcg != pc->mem_cgroup || !PageCgroupUsed(pc)) {
+       if (memcg != pc->mem_cgroup) {
                move_unlock_mem_cgroup(memcg, flags);
                goto again;
        }
@@ -2525,7 +2523,7 @@ static struct mem_cgroup *mem_cgroup_lookup(unsigned short id)
  */
 struct mem_cgroup *try_get_mem_cgroup_from_page(struct page *page)
 {
-       struct mem_cgroup *memcg = NULL;
+       struct mem_cgroup *memcg;
        struct page_cgroup *pc;
        unsigned short id;
        swp_entry_t ent;
@@ -2533,9 +2531,10 @@ struct mem_cgroup *try_get_mem_cgroup_from_page(struct page *page)
        VM_BUG_ON_PAGE(!PageLocked(page), page);
 
        pc = lookup_page_cgroup(page);
-       if (PageCgroupUsed(pc)) {
-               memcg = pc->mem_cgroup;
-               if (memcg && !css_tryget_online(&memcg->css))
+       memcg = pc->mem_cgroup;
+
+       if (memcg) {
+               if (!css_tryget_online(&memcg->css))
                        memcg = NULL;
        } else if (PageSwapCache(page)) {
                ent.val = page_private(page);
@@ -2586,7 +2585,7 @@ static void commit_charge(struct page *page, struct mem_cgroup *memcg,
        struct page_cgroup *pc = lookup_page_cgroup(page);
        int isolated;
 
-       VM_BUG_ON_PAGE(PageCgroupUsed(pc), page);
+       VM_BUG_ON_PAGE(pc->mem_cgroup, page);
        /*
         * we don't need page_cgroup_lock about tail pages, becase they are not
         * accessed by any other context at this point.
@@ -2601,7 +2600,7 @@ static void commit_charge(struct page *page, struct mem_cgroup *memcg,
 
        /*
         * Nobody should be changing or seriously looking at
-        * pc->mem_cgroup and pc->flags at this point:
+        * pc->mem_cgroup at this point:
         *
         * - the page is uncharged
         *
@@ -2614,7 +2613,6 @@ static void commit_charge(struct page *page, struct mem_cgroup *memcg,
         *   have the page locked
         */
        pc->mem_cgroup = memcg;
-       pc->flags = PCG_USED;
 
        if (lrucare)
                unlock_page_lru(page, isolated);
@@ -3126,37 +3124,22 @@ void __memcg_kmem_commit_charge(struct page *page, struct mem_cgroup *memcg,
                memcg_uncharge_kmem(memcg, 1 << order);
                return;
        }
-       /*
-        * The page is freshly allocated and not visible to any
-        * outside callers yet.  Set up pc non-atomically.
-        */
        pc = lookup_page_cgroup(page);
        pc->mem_cgroup = memcg;
-       pc->flags = PCG_USED;
 }
 
 void __memcg_kmem_uncharge_pages(struct page *page, int order)
 {
-       struct mem_cgroup *memcg = NULL;
-       struct page_cgroup *pc;
-
-
-       pc = lookup_page_cgroup(page);
-       if (!PageCgroupUsed(pc))
-               return;
-
-       memcg = pc->mem_cgroup;
-       pc->flags = 0;
+       struct page_cgroup *pc = lookup_page_cgroup(page);
+       struct mem_cgroup *memcg = pc->mem_cgroup;
 
-       /*
-        * We trust that only if there is a memcg associated with the page, it
-        * is a valid allocation
-        */
        if (!memcg)
                return;
 
        VM_BUG_ON_PAGE(mem_cgroup_is_root(memcg), page);
+
        memcg_uncharge_kmem(memcg, 1 << order);
+       pc->mem_cgroup = NULL;
 }
 #else
 static inline void memcg_unregister_all_caches(struct mem_cgroup *memcg)
@@ -3174,23 +3157,16 @@ static inline void memcg_unregister_all_caches(struct mem_cgroup *memcg)
  */
 void mem_cgroup_split_huge_fixup(struct page *head)
 {
-       struct page_cgroup *head_pc;
-       struct page_cgroup *pc;
-       struct mem_cgroup *memcg;
+       struct page_cgroup *pc = lookup_page_cgroup(head);
        int i;
 
        if (mem_cgroup_disabled())
                return;
 
-       head_pc = lookup_page_cgroup(head);
+       for (i = 1; i < HPAGE_PMD_NR; i++)
+               pc[i].mem_cgroup = pc[0].mem_cgroup;
 
-       memcg = head_pc->mem_cgroup;
-       for (i = 1; i < HPAGE_PMD_NR; i++) {
-               pc = head_pc + i;
-               pc->mem_cgroup = memcg;
-               pc->flags = head_pc->flags;
-       }
-       __this_cpu_sub(memcg->stat->count[MEM_CGROUP_STAT_RSS_HUGE],
+       __this_cpu_sub(pc[0].mem_cgroup->stat->count[MEM_CGROUP_STAT_RSS_HUGE],
                       HPAGE_PMD_NR);
 }
 #endif /* CONFIG_TRANSPARENT_HUGEPAGE */
@@ -3240,7 +3216,7 @@ static int mem_cgroup_move_account(struct page *page,
                goto out;
 
        ret = -EINVAL;
-       if (!PageCgroupUsed(pc) || pc->mem_cgroup != from)
+       if (pc->mem_cgroup != from)
                goto out_unlock;
 
        move_lock_mem_cgroup(from, &flags);
@@ -3350,7 +3326,7 @@ static struct page_cgroup *lookup_page_cgroup_used(struct page *page)
         * the first time, i.e. during boot or memory hotplug;
         * or when mem_cgroup_disabled().
         */
-       if (likely(pc) && PageCgroupUsed(pc))
+       if (likely(pc) && pc->mem_cgroup)
                return pc;
        return NULL;
 }
@@ -3368,10 +3344,8 @@ void mem_cgroup_print_bad_page(struct page *page)
        struct page_cgroup *pc;
 
        pc = lookup_page_cgroup_used(page);
-       if (pc) {
-               pr_alert("pc:%p pc->flags:%lx pc->mem_cgroup:%p\n",
-                        pc, pc->flags, pc->mem_cgroup);
-       }
+       if (pc)
+               pr_alert("pc:%p pc->mem_cgroup:%p\n", pc, pc->mem_cgroup);
 }
 #endif
 
@@ -5308,7 +5282,7 @@ static enum mc_target_type get_mctgt_type(struct vm_area_struct *vma,
                 * mem_cgroup_move_account() checks the pc is valid or
                 * not under LRU exclusion.
                 */
-               if (PageCgroupUsed(pc) && pc->mem_cgroup == mc.from) {
+               if (pc->mem_cgroup == mc.from) {
                        ret = MC_TARGET_PAGE;
                        if (target)
                                target->page = page;
@@ -5344,7 +5318,7 @@ static enum mc_target_type get_mctgt_type_thp(struct vm_area_struct *vma,
        if (!move_anon())
                return ret;
        pc = lookup_page_cgroup(page);
-       if (PageCgroupUsed(pc) && pc->mem_cgroup == mc.from) {
+       if (pc->mem_cgroup == mc.from) {
                ret = MC_TARGET_PAGE;
                if (target) {
                        get_page(page);
@@ -5788,18 +5762,17 @@ void mem_cgroup_swapout(struct page *page, swp_entry_t entry)
                return;
 
        pc = lookup_page_cgroup(page);
+       memcg = pc->mem_cgroup;
 
        /* Readahead page, never charged */
-       if (!PageCgroupUsed(pc))
+       if (!memcg)
                return;
 
-       memcg = pc->mem_cgroup;
-
        oldid = swap_cgroup_record(entry, mem_cgroup_id(memcg));
        VM_BUG_ON_PAGE(oldid, page);
        mem_cgroup_swap_statistics(memcg, true);
 
-       pc->flags = 0;
+       pc->mem_cgroup = NULL;
 
        if (!mem_cgroup_is_root(memcg))
                page_counter_uncharge(&memcg->memory, 1);
@@ -5874,7 +5847,7 @@ int mem_cgroup_try_charge(struct page *page, struct mm_struct *mm,
                 * the page lock, which serializes swap cache removal, which
                 * in turn serializes uncharging.
                 */
-               if (PageCgroupUsed(pc))
+               if (pc->mem_cgroup)
                        goto out;
        }
 
@@ -6036,13 +6009,13 @@ static void uncharge_list(struct list_head *page_list)
                VM_BUG_ON_PAGE(page_count(page), page);
 
                pc = lookup_page_cgroup(page);
-               if (!PageCgroupUsed(pc))
+               if (!pc->mem_cgroup)
                        continue;
 
                /*
                 * Nobody should be changing or seriously looking at
-                * pc->mem_cgroup and pc->flags at this point, we have
-                * fully exclusive access to the page.
+                * pc->mem_cgroup at this point, we have fully
+                * exclusive access to the page.
                 */
 
                if (memcg != pc->mem_cgroup) {
@@ -6065,7 +6038,7 @@ static void uncharge_list(struct list_head *page_list)
                else
                        nr_file += nr_pages;
 
-               pc->flags = 0;
+               pc->mem_cgroup = NULL;
 
                pgpgout++;
        } while (next != page_list);
@@ -6091,7 +6064,7 @@ void mem_cgroup_uncharge(struct page *page)
 
        /* Don't touch page->lru of any random page, pre-check: */
        pc = lookup_page_cgroup(page);
-       if (!PageCgroupUsed(pc))
+       if (!pc->mem_cgroup)
                return;
 
        INIT_LIST_HEAD(&page->lru);
@@ -6127,6 +6100,7 @@ void mem_cgroup_uncharge_list(struct list_head *page_list)
 void mem_cgroup_migrate(struct page *oldpage, struct page *newpage,
                        bool lrucare)
 {
+       struct mem_cgroup *memcg;
        struct page_cgroup *pc;
        int isolated;
 
@@ -6143,7 +6117,7 @@ void mem_cgroup_migrate(struct page *oldpage, struct page *newpage,
 
        /* Page cache replacement: new page already charged? */
        pc = lookup_page_cgroup(newpage);
-       if (PageCgroupUsed(pc))
+       if (pc->mem_cgroup)
                return;
 
        /*
@@ -6153,18 +6127,19 @@ void mem_cgroup_migrate(struct page *oldpage, struct page *newpage,
         * reclaim just put back on the LRU but has not released yet.
         */
        pc = lookup_page_cgroup(oldpage);
-       if (!PageCgroupUsed(pc))
+       memcg = pc->mem_cgroup;
+       if (!memcg)
                return;
 
        if (lrucare)
                lock_page_lru(oldpage, &isolated);
 
-       pc->flags = 0;
+       pc->mem_cgroup = NULL;
 
        if (lrucare)
                unlock_page_lru(oldpage, isolated);
 
-       commit_charge(newpage, pc->mem_cgroup, lrucare);
+       commit_charge(newpage, memcg, lrucare);
 }
 
 /*