Memory controller: memory accounting

Add the accounting hooks.  The accounting is carried out for RSS and Page
Cache (unmapped) pages.  There is now a common limit and accounting for both.
The RSS accounting is accounted at page_add_*_rmap() and page_remove_rmap()
time.  Page cache is accounted at add_to_page_cache(),
__delete_from_page_cache().  Swap cache is also accounted for.

Each page's page_cgroup is protected with the last bit of the
page_cgroup pointer, this makes handling of race conditions involving
simultaneous mappings of a page easier.  A reference count is kept in the
page_cgroup to deal with cases where a page might be unmapped from the RSS
of all tasks, but still lives in the page cache.

Credits go to Vaidyanathan Srinivasan for helping with reference counting work
of the page cgroup.  Almost all of the page cache accounting code has help
from Vaidyanathan Srinivasan.

[[email protected]: fix swapoff breakage]
[[email protected]: fix locking]
Signed-off-by: Vaidyanathan Srinivasan <[email protected]>
Signed-off-by: Balbir Singh <[email protected]>
Cc: Pavel Emelianov <[email protected]>
Cc: Paul Menage <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Cc: Nick Piggin <[email protected]>
Cc: Kirill Korotaev <[email protected]>
Cc: Herbert Poetzl <[email protected]>
Cc: David Rientjes <[email protected]>
Cc: <[email protected]>
Signed-off-by: Hugh Dickins <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>

include/linux/memcontrol.h

@@ -30,6 +30,13 @@ extern void mm_free_cgroup(struct mm_struct *mm);
 extern void page_assign_page_cgroup(struct page *page,
 					struct page_cgroup *pc);
 extern struct page_cgroup *page_get_page_cgroup(struct page *page);
+extern int mem_cgroup_charge(struct page *page, struct mm_struct *mm);
+extern void mem_cgroup_uncharge(struct page_cgroup *pc);
+
+static inline void mem_cgroup_uncharge_page(struct page *page)
+{
+	mem_cgroup_uncharge(page_get_page_cgroup(page));
+}
 
 #else /* CONFIG_CGROUP_MEM_CONT */
 static inline void mm_init_cgroup(struct mm_struct *mm,
@@ -51,6 +58,19 @@ static inline struct page_cgroup *page_get_page_cgroup(struct page *page)
 	return NULL;
 }
 
+static inline int mem_cgroup_charge(struct page *page, struct mm_struct *mm)
+{
+	return 0;
+}
+
+static inline void mem_cgroup_uncharge(struct page_cgroup *pc)
+{
+}
+
+static inline void mem_cgroup_uncharge_page(struct page *page)
+{
+}
+
 #endif /* CONFIG_CGROUP_MEM_CONT */
 
 #endif /* _LINUX_MEMCONTROL_H */

mm/filemap.c

@@ -33,6 +33,7 @@
 #include <linux/syscalls.h>
 #include <linux/cpuset.h>
 #include <linux/hardirq.h> /* for BUG_ON(!in_atomic()) only */
+#include <linux/memcontrol.h>
 #include "internal.h"
 
 /*
@@ -118,6 +119,7 @@ void __remove_from_page_cache(struct page *page)
 {
 	struct address_space *mapping = page->mapping;
 
+	mem_cgroup_uncharge_page(page);
 	radix_tree_delete(&mapping->page_tree, page->index);
 	page->mapping = NULL;
 	mapping->nrpages--;
@@ -461,6 +463,11 @@ int add_to_page_cache(struct page *page, struct address_space *mapping,
 	int error = radix_tree_preload(gfp_mask & ~__GFP_HIGHMEM);
 
 	if (error == 0) {
+
+		error = mem_cgroup_charge(page, current->mm);
+		if (error)
+			goto out;
+
 		write_lock_irq(&mapping->tree_lock);
 		error = radix_tree_insert(&mapping->page_tree, offset, page);
 		if (!error) {
@@ -470,10 +477,13 @@ int add_to_page_cache(struct page *page, struct address_space *mapping,
 			page->index = offset;
 			mapping->nrpages++;
 			__inc_zone_page_state(page, NR_FILE_PAGES);
-		}
+		} else
+			mem_cgroup_uncharge_page(page);
+
 		write_unlock_irq(&mapping->tree_lock);
 		radix_tree_preload_end();
 	}
+out:
 	return error;
 }
 EXPORT_SYMBOL(add_to_page_cache);

mm/memcontrol.c

@@ -21,6 +21,9 @@
 #include <linux/memcontrol.h>
 #include <linux/cgroup.h>
 #include <linux/mm.h>
+#include <linux/page-flags.h>
+#include <linux/bit_spinlock.h>
+#include <linux/rcupdate.h>
 
 struct cgroup_subsys mem_cgroup_subsys;
 
@@ -31,7 +34,9 @@ struct cgroup_subsys mem_cgroup_subsys;
  * to help the administrator determine what knobs to tune.
  *
  * TODO: Add a water mark for the memory controller. Reclaim will begin when
- * we hit the water mark.
+ * we hit the water mark. May be even add a low water mark, such that
+ * no reclaim occurs from a cgroup at it's low water mark, this is
+ * a feature that will be implemented much later in the future.
  */
 struct mem_cgroup {
 	struct cgroup_subsys_state css;
@@ -48,6 +53,14 @@ struct mem_cgroup {
 	struct list_head inactive_list;
 };
 
+/*
+ * We use the lower bit of the page->page_cgroup pointer as a bit spin
+ * lock. We need to ensure that page->page_cgroup is atleast two
+ * byte aligned (based on comments from Nick Piggin)
+ */
+#define PAGE_CGROUP_LOCK_BIT 	0x0
+#define PAGE_CGROUP_LOCK 		(1 << PAGE_CGROUP_LOCK_BIT)
+
 /*
  * A page_cgroup page is associated with every page descriptor. The
  * page_cgroup helps us identify information about the cgroup
@@ -56,6 +69,8 @@ struct page_cgroup {
 	struct list_head lru;		/* per cgroup LRU list */
 	struct page *page;
 	struct mem_cgroup *mem_cgroup;
+	atomic_t ref_cnt;		/* Helpful when pages move b/w  */
+					/* mapped and cached states     */
 };
 
 
@@ -88,14 +103,157 @@ void mm_free_cgroup(struct mm_struct *mm)
 	css_put(&mm->mem_cgroup->css);
 }
 
+static inline int page_cgroup_locked(struct page *page)
+{
+	return bit_spin_is_locked(PAGE_CGROUP_LOCK_BIT,
+					&page->page_cgroup);
+}
+
 void page_assign_page_cgroup(struct page *page, struct page_cgroup *pc)
 {
-	page->page_cgroup = (unsigned long)pc;
+	int locked;
+
+	/*
+	 * While resetting the page_cgroup we might not hold the
+	 * page_cgroup lock. free_hot_cold_page() is an example
+	 * of such a scenario
+	 */
+	if (pc)
+		VM_BUG_ON(!page_cgroup_locked(page));
+	locked = (page->page_cgroup & PAGE_CGROUP_LOCK);
+	page->page_cgroup = ((unsigned long)pc | locked);
 }
 
 struct page_cgroup *page_get_page_cgroup(struct page *page)
 {
-	return page->page_cgroup;
+	return (struct page_cgroup *)
+		(page->page_cgroup & ~PAGE_CGROUP_LOCK);
+}
+
+void __always_inline lock_page_cgroup(struct page *page)
+{
+	bit_spin_lock(PAGE_CGROUP_LOCK_BIT, &page->page_cgroup);
+	VM_BUG_ON(!page_cgroup_locked(page));
+}
+
+void __always_inline unlock_page_cgroup(struct page *page)
+{
+	bit_spin_unlock(PAGE_CGROUP_LOCK_BIT, &page->page_cgroup);
+}
+
+/*
+ * Charge the memory controller for page usage.
+ * Return
+ * 0 if the charge was successful
+ * < 0 if the cgroup is over its limit
+ */
+int mem_cgroup_charge(struct page *page, struct mm_struct *mm)
+{
+	struct mem_cgroup *mem;
+	struct page_cgroup *pc, *race_pc;
+
+	/*
+	 * Should page_cgroup's go to their own slab?
+	 * One could optimize the performance of the charging routine
+	 * by saving a bit in the page_flags and using it as a lock
+	 * to see if the cgroup page already has a page_cgroup associated
+	 * with it
+	 */
+	lock_page_cgroup(page);
+	pc = page_get_page_cgroup(page);
+	/*
+	 * The page_cgroup exists and the page has already been accounted
+	 */
+	if (pc) {
+		atomic_inc(&pc->ref_cnt);
+		goto done;
+	}
+
+	unlock_page_cgroup(page);
+
+	pc = kzalloc(sizeof(struct page_cgroup), GFP_KERNEL);
+	if (pc == NULL)
+		goto err;
+
+	rcu_read_lock();
+	/*
+	 * We always charge the cgroup the mm_struct belongs to
+	 * the mm_struct's mem_cgroup changes on task migration if the
+	 * 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;
+
+	mem = rcu_dereference(mm->mem_cgroup);
+	/*
+	 * For every charge from the cgroup, increment reference
+	 * count
+	 */
+	css_get(&mem->css);
+	rcu_read_unlock();
+
+	/*
+	 * If we created the page_cgroup, we should free it on exceeding
+	 * the cgroup limit.
+	 */
+	if (res_counter_charge(&mem->res, 1)) {
+		css_put(&mem->css);
+		goto free_pc;
+	}
+
+	lock_page_cgroup(page);
+	/*
+	 * Check if somebody else beat us to allocating the page_cgroup
+	 */
+	race_pc = page_get_page_cgroup(page);
+	if (race_pc) {
+		kfree(pc);
+		pc = race_pc;
+		atomic_inc(&pc->ref_cnt);
+		res_counter_uncharge(&mem->res, 1);
+		css_put(&mem->css);
+		goto done;
+	}
+
+	atomic_set(&pc->ref_cnt, 1);
+	pc->mem_cgroup = mem;
+	pc->page = page;
+	page_assign_page_cgroup(page, pc);
+
+done:
+	unlock_page_cgroup(page);
+	return 0;
+free_pc:
+	kfree(pc);
+	return -ENOMEM;
+err:
+	unlock_page_cgroup(page);
+	return -ENOMEM;
+}
+
+/*
+ * Uncharging is always a welcome operation, we never complain, simply
+ * uncharge.
+ */
+void mem_cgroup_uncharge(struct page_cgroup *pc)
+{
+	struct mem_cgroup *mem;
+	struct page *page;
+
+	if (!pc)
+		return;
+
+	if (atomic_dec_and_test(&pc->ref_cnt)) {
+		page = pc->page;
+		lock_page_cgroup(page);
+		mem = pc->mem_cgroup;
+		css_put(&mem->css);
+		page_assign_page_cgroup(page, NULL);
+		unlock_page_cgroup(page);
+		res_counter_uncharge(&mem->res, 1);
+		kfree(pc);
+	}
 }
 
 static ssize_t mem_cgroup_read(struct cgroup *cont, struct cftype *cft,
@@ -150,6 +308,8 @@ mem_cgroup_create(struct cgroup_subsys *ss, struct cgroup *cont)
 		return NULL;
 
 	res_counter_init(&mem->res);
+	INIT_LIST_HEAD(&mem->active_list);
+	INIT_LIST_HEAD(&mem->inactive_list);
 	return &mem->css;
 }