@@ -490,7 +490,8 @@ bool mem_cgroup_oom_synchronize(bool wait);
extern int do_swap_account;
#endif
-void lock_page_memcg(struct page *page);
+struct mem_cgroup *lock_page_memcg(struct page *page);
+void __unlock_page_memcg(struct mem_cgroup *memcg);
void unlock_page_memcg(struct page *page);
/**
@@ -529,6 +530,27 @@ static inline void mem_cgroup_dec_page_stat(struct page *page,
mem_cgroup_update_page_stat(page, idx, -1);
}
+static inline void mem_cgroup_update_stat(struct mem_cgroup *memcg,
+ enum mem_cgroup_stat_index idx, int val)
+{
+ VM_BUG_ON(!(rcu_read_lock_held()));
+
+ if (memcg)
+ this_cpu_add(memcg->stat->count[idx], val);
+}
+
+static inline void mem_cgroup_inc_stat(struct mem_cgroup *memcg,
+ enum mem_cgroup_stat_index idx)
+{
+ mem_cgroup_update_stat(memcg, idx, 1);
+}
+
+static inline void mem_cgroup_dec_stat(struct mem_cgroup *memcg,
+ enum mem_cgroup_stat_index idx)
+{
+ mem_cgroup_update_stat(memcg, idx, -1);
+}
+
unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order,
gfp_t gfp_mask,
unsigned long *total_scanned);
@@ -709,7 +731,12 @@ mem_cgroup_print_oom_info(struct mem_cgroup *memcg, struct task_struct *p)
{
}
-static inline void lock_page_memcg(struct page *page)
+static inline struct mem_cgroup *lock_page_memcg(struct page *page)
+{
+ return NULL;
+}
+
+static inline void __unlock_page_memcg(struct mem_cgroup *memcg)
{
}
@@ -1619,9 +1619,13 @@ bool mem_cgroup_oom_synchronize(bool handle)
* @page: the page
*
* This function protects unlocked LRU pages from being moved to
- * another cgroup and stabilizes their page->mem_cgroup binding.
+ * another cgroup.
+ *
+ * It ensures lifetime of the returned memcg. Caller is responsible
+ * for the lifetime of the page; __unlock_page_memcg() is available
+ * when @page might get freed inside the locked section.
*/
-void lock_page_memcg(struct page *page)
+struct mem_cgroup *lock_page_memcg(struct page *page)
{
struct mem_cgroup *memcg;
unsigned long flags;
@@ -1630,18 +1634,24 @@ void lock_page_memcg(struct page *page)
* The RCU lock is held throughout the transaction. The fast
* path can get away without acquiring the memcg->move_lock
* because page moving starts with an RCU grace period.
- */
+ *
+ * The RCU lock also protects the memcg from being freed when
+ * the page state that is going to change is the only thing
+ * preventing the page itself from being freed. E.g. writeback
+ * doesn't hold a page reference and relies on PG_writeback to
+ * keep off truncation, migration and so forth.
+ */
rcu_read_lock();
if (mem_cgroup_disabled())
- return;
+ return NULL;
again:
memcg = page->mem_cgroup;
if (unlikely(!memcg))
- return;
+ return NULL;
if (atomic_read(&memcg->moving_account) <= 0)
- return;
+ return memcg;
spin_lock_irqsave(&memcg->move_lock, flags);
if (memcg != page->mem_cgroup) {
@@ -1657,18 +1667,18 @@ void lock_page_memcg(struct page *page)
memcg->move_lock_task = current;
memcg->move_lock_flags = flags;
- return;
+ return memcg;
}
EXPORT_SYMBOL(lock_page_memcg);
/**
- * unlock_page_memcg - unlock a page->mem_cgroup binding
- * @page: the page
+ * __unlock_page_memcg - unlock and unpin a memcg
+ * @memcg: the memcg
+ *
+ * Unlock and unpin a memcg returned by lock_page_memcg().
*/
-void unlock_page_memcg(struct page *page)
+void __unlock_page_memcg(struct mem_cgroup *memcg)
{
- struct mem_cgroup *memcg = page->mem_cgroup;
-
if (memcg && memcg->move_lock_task == current) {
unsigned long flags = memcg->move_lock_flags;
@@ -1680,6 +1690,15 @@ void unlock_page_memcg(struct page *page)
rcu_read_unlock();
}
+
+/**
+ * unlock_page_memcg - unlock a page->mem_cgroup binding
+ * @page: the page
+ */
+void unlock_page_memcg(struct page *page)
+{
+ __unlock_page_memcg(page->mem_cgroup);
+}
EXPORT_SYMBOL(unlock_page_memcg);
/*
@@ -2704,9 +2704,10 @@ EXPORT_SYMBOL(clear_page_dirty_for_io);
int test_clear_page_writeback(struct page *page)
{
struct address_space *mapping = page_mapping(page);
+ struct mem_cgroup *memcg;
int ret;
- lock_page_memcg(page);
+ memcg = lock_page_memcg(page);
if (mapping && mapping_use_writeback_tags(mapping)) {
struct inode *inode = mapping->host;
struct backing_dev_info *bdi = inode_to_bdi(inode);
@@ -2734,13 +2735,20 @@ int test_clear_page_writeback(struct page *page)
} else {
ret = TestClearPageWriteback(page);
}
+
+ /*
+ * NOTE: Page might be free now! Writeback doesn't hold a page
+ * reference on its own, it relies on truncation to wait for
+ * the clearing of PG_writeback. The below can only access
+ * page state that is static across allocation cycles.
+ */
if (ret) {
- mem_cgroup_dec_page_stat(page, MEM_CGROUP_STAT_WRITEBACK);
+ mem_cgroup_dec_stat(memcg, MEM_CGROUP_STAT_WRITEBACK);
dec_node_page_state(page, NR_WRITEBACK);
dec_zone_page_state(page, NR_ZONE_WRITE_PENDING);
inc_node_page_state(page, NR_WRITTEN);
}
- unlock_page_memcg(page);
+ __unlock_page_memcg(memcg);
return ret;
}