memcg: better migration handling

This patch changes page migration under memory controller to use a
different algorithm.  (thanks to Christoph for new idea.)

Before:
 - page_cgroup is migrated from an old page to a new page.
After:
 - a new page is accounted , no reuse of page_cgroup.

Pros:

 - We can avoid compliated lock depndencies and races in migration.

Cons:

 - new param to mem_cgroup_charge_common().

 - mem_cgroup_getref() is added for handling ref_cnt ping-pong.

This version simplifies complicated lock dependency in page migraiton
under memory resource controller.

  new refcnt sequence is following.

a mapped page:
  prepage_migration() ..... +1 to NEW page
  try_to_unmap()      ..... all refs to OLD page is gone.
  move_pages()        ..... +1 to NEW page if page cache.
  remap...            ..... all refs from *map* is added to NEW one.
  end_migration()     ..... -1 to New page.

  page's mapcount + (page_is_cache) refs are added to NEW one.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Pavel Emelyanov <xemul@openvz.org>
Cc: Li Zefan <lizf@cn.fujitsu.com>
Cc: YAMAMOTO Takashi <yamamoto@valinux.co.jp>
Cc: Hugh Dickins <hugh@veritas.com>
Cc: Christoph Lameter <cl@linux-foundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

1 files changed, 65 insertions, 63 deletions

diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 90ccc1326356..da5912b84551 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -524,7 +524,8 @@ unsigned long mem_cgroup_isolate_pages(unsigned long nr_to_scan,
  * < 0 if the cgroup is over its limit
  */
 static int mem_cgroup_charge_common(struct page *page, struct mm_struct *mm,
-                                gfp_t gfp_mask, enum charge_type ctype)
+                                gfp_t gfp_mask, enum charge_type ctype,
+                                struct mem_cgroup *memcg)
 {
         struct mem_cgroup *mem;
         struct page_cgroup *pc;
@@ -569,16 +570,21 @@ retry:
          * thread group leader migrates. It's possible that mm is not
          * set, if so charge the init_mm (happens for pagecache usage).
          */
-        if (!mm)
-                mm = &init_mm;
+        if (!memcg) {
+                if (!mm)
+                        mm = &init_mm;
 
-        rcu_read_lock();
-        mem = mem_cgroup_from_task(rcu_dereference(mm->owner));
-        /*
-         * For every charge from the cgroup, increment reference count
-         */
-        css_get(&mem->css);
-        rcu_read_unlock();
+                rcu_read_lock();
+                mem = mem_cgroup_from_task(rcu_dereference(mm->owner));
+                /*
+                 * For every charge from the cgroup, increment reference count
+                 */
+                css_get(&mem->css);
+                rcu_read_unlock();
+        } else {
+                mem = memcg;
+                css_get(&memcg->css);
+        }
 
         while (res_counter_charge(&mem->res, PAGE_SIZE)) {
                 if (!(gfp_mask & __GFP_WAIT))
@@ -648,7 +654,7 @@ err:
 int mem_cgroup_charge(struct page *page, struct mm_struct *mm, gfp_t gfp_mask)
 {
         return mem_cgroup_charge_common(page, mm, gfp_mask,
-                                MEM_CGROUP_CHARGE_TYPE_MAPPED);
+                                MEM_CGROUP_CHARGE_TYPE_MAPPED, NULL);
 }
 
 int mem_cgroup_cache_charge(struct page *page, struct mm_struct *mm,
@@ -657,7 +663,22 @@ int mem_cgroup_cache_charge(struct page *page, struct mm_struct *mm,
         if (!mm)
                 mm = &init_mm;
         return mem_cgroup_charge_common(page, mm, gfp_mask,
-                                MEM_CGROUP_CHARGE_TYPE_CACHE);
+                                MEM_CGROUP_CHARGE_TYPE_CACHE, NULL);
+}
+
+int mem_cgroup_getref(struct page *page)
+{
+        struct page_cgroup *pc;
+
+        if (mem_cgroup_subsys.disabled)
+                return 0;
+
+        lock_page_cgroup(page);
+        pc = page_get_page_cgroup(page);
+        VM_BUG_ON(!pc);
+        pc->ref_cnt++;
+        unlock_page_cgroup(page);
+        return 0;
 }
 
 /*
@@ -707,65 +728,39 @@ unlock:
 }
 
 /*
- * Returns non-zero if a page (under migration) has valid page_cgroup member.
- * Refcnt of page_cgroup is incremented.
+ * Before starting migration, account against new page.
  */
-int mem_cgroup_prepare_migration(struct page *page)
+int mem_cgroup_prepare_migration(struct page *page, struct page *newpage)
 {
         struct page_cgroup *pc;
+        struct mem_cgroup *mem = NULL;
+        enum charge_type ctype = MEM_CGROUP_CHARGE_TYPE_MAPPED;
+        int ret = 0;
 
         if (mem_cgroup_subsys.disabled)
                 return 0;
 
         lock_page_cgroup(page);
         pc = page_get_page_cgroup(page);
-        if (pc)
-                pc->ref_cnt++;
+        if (pc) {
+                mem = pc->mem_cgroup;
+                css_get(&mem->css);
+                if (pc->flags & PAGE_CGROUP_FLAG_CACHE)
+                        ctype = MEM_CGROUP_CHARGE_TYPE_CACHE;
+        }
         unlock_page_cgroup(page);
-        return pc != NULL;
-}
-
-void mem_cgroup_end_migration(struct page *page)
-{
-        mem_cgroup_uncharge_page(page);
+        if (mem) {
+                ret = mem_cgroup_charge_common(newpage, NULL, GFP_KERNEL,
+                        ctype, mem);
+                css_put(&mem->css);
+        }
+        return ret;
 }
 
-/*
- * We know both *page* and *newpage* are now not-on-LRU and PG_locked.
- * And no race with uncharge() routines because page_cgroup for *page*
- * has extra one reference by mem_cgroup_prepare_migration.
- */
-void mem_cgroup_page_migration(struct page *page, struct page *newpage)
+/* remove redundant charge */
+void mem_cgroup_end_migration(struct page *newpage)
 {
-        struct page_cgroup *pc;
-        struct mem_cgroup_per_zone *mz;
-        unsigned long flags;
-
-        lock_page_cgroup(page);
-        pc = page_get_page_cgroup(page);
-        if (!pc) {
-                unlock_page_cgroup(page);
-                return;
-        }
-
-        mz = page_cgroup_zoneinfo(pc);
-        spin_lock_irqsave(&mz->lru_lock, flags);
-        __mem_cgroup_remove_list(mz, pc);
-        spin_unlock_irqrestore(&mz->lru_lock, flags);
-
-        page_assign_page_cgroup(page, NULL);
-        unlock_page_cgroup(page);
-
-        pc->page = newpage;
-        lock_page_cgroup(newpage);
-        page_assign_page_cgroup(newpage, pc);
-
-        mz = page_cgroup_zoneinfo(pc);
-        spin_lock_irqsave(&mz->lru_lock, flags);
-        __mem_cgroup_add_list(mz, pc);
-        spin_unlock_irqrestore(&mz->lru_lock, flags);
-
-        unlock_page_cgroup(newpage);
+        mem_cgroup_uncharge_page(newpage);
 }
 
 /*
@@ -795,12 +790,19 @@ static void mem_cgroup_force_empty_list(struct mem_cgroup *mem,
                 page = pc->page;
                 get_page(page);
                 spin_unlock_irqrestore(&mz->lru_lock, flags);
-                mem_cgroup_uncharge_page(page);
-                put_page(page);
-                if (--count <= 0) {
-                        count = FORCE_UNCHARGE_BATCH;
+                /*
+                 * Check if this page is on LRU. !LRU page can be found
+                 * if it's under page migration.
+                 */
+                if (PageLRU(page)) {
+                        mem_cgroup_uncharge_page(page);
+                        put_page(page);
+                        if (--count <= 0) {
+                                count = FORCE_UNCHARGE_BATCH;
+                                cond_resched();
+                        }
+                } else
                         cond_resched();
-                }
                 spin_lock_irqsave(&mz->lru_lock, flags);
         }
         spin_unlock_irqrestore(&mz->lru_lock, flags);