memcg: add memcg sanity checks at allocating and freeing pages

Add checks at allocating or freeing a page whether the page is used (iow,
charged) from the view point of memcg.

This check may be useful in debugging a problem and we did similar checks
before the commit 52d4b9ac(memcg: allocate all page_cgroup at boot).

This patch adds some overheads at allocating or freeing memory, so it's
enabled only when CONFIG_DEBUG_VM is enabled.

Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

include/linux/memcontrol.h

@@ -151,6 +151,10 @@ u64 mem_cgroup_get_limit(struct mem_cgroup *mem);
 void mem_cgroup_split_huge_fixup(struct page *head, struct page *tail);
 #endif
 
+#ifdef CONFIG_DEBUG_VM
+bool mem_cgroup_bad_page_check(struct page *page);
+void mem_cgroup_print_bad_page(struct page *page);
+#endif
 #else /* CONFIG_CGROUP_MEM_RES_CTLR */
 struct mem_cgroup;
 
@@ -352,5 +356,18 @@ static inline void mem_cgroup_split_huge_fixup(struct page *head,
 
 #endif /* CONFIG_CGROUP_MEM_CONT */
 
+#if !defined(CONFIG_CGROUP_MEM_RES_CTLR) || !defined(CONFIG_DEBUG_VM)
+static inline bool
+mem_cgroup_bad_page_check(struct page *page)
+{
+	return false;
+}
+
+static inline void
+mem_cgroup_print_bad_page(struct page *page)
+{
+}
+#endif
+
 #endif /* _LINUX_MEMCONTROL_H */
 

mm/memcontrol.c

@@ -3046,6 +3046,52 @@ int mem_cgroup_shmem_charge_fallback(struct page *page,
 	return ret;
 }
 
+#ifdef CONFIG_DEBUG_VM
+static struct page_cgroup *lookup_page_cgroup_used(struct page *page)
+{
+	struct page_cgroup *pc;
+
+	pc = lookup_page_cgroup(page);
+	if (likely(pc) && PageCgroupUsed(pc))
+		return pc;
+	return NULL;
+}
+
+bool mem_cgroup_bad_page_check(struct page *page)
+{
+	if (mem_cgroup_disabled())
+		return false;
+
+	return lookup_page_cgroup_used(page) != NULL;
+}
+
+void mem_cgroup_print_bad_page(struct page *page)
+{
+	struct page_cgroup *pc;
+
+	pc = lookup_page_cgroup_used(page);
+	if (pc) {
+		int ret = -1;
+		char *path;
+
+		printk(KERN_ALERT "pc:%p pc->flags:%lx pc->mem_cgroup:%p",
+		       pc, pc->flags, pc->mem_cgroup);
+
+		path = kmalloc(PATH_MAX, GFP_KERNEL);
+		if (path) {
+			rcu_read_lock();
+			ret = cgroup_path(pc->mem_cgroup->css.cgroup,
+							path, PATH_MAX);
+			rcu_read_unlock();
+		}
+
+		printk(KERN_CONT "(%s)\n",
+				(ret < 0) ? "cannot get the path" : path);
+		kfree(path);
+	}
+}
+#endif
+
 static DEFINE_MUTEX(set_limit_mutex);
 
 static int mem_cgroup_resize_limit(struct mem_cgroup *memcg,

mm/page_alloc.c

@@ -53,6 +53,7 @@
 #include <linux/compaction.h>
 #include <trace/events/kmem.h>
 #include <linux/ftrace_event.h>
+#include <linux/memcontrol.h>
 
 #include <asm/tlbflush.h>
 #include <asm/div64.h>
@@ -565,7 +566,8 @@ static inline int free_pages_check(struct page *page)
 	if (unlikely(page_mapcount(page) |
 		(page->mapping != NULL)  |
 		(atomic_read(&page->_count) != 0) |
-		(page->flags & PAGE_FLAGS_CHECK_AT_FREE))) {
+		(page->flags & PAGE_FLAGS_CHECK_AT_FREE) |
+		(mem_cgroup_bad_page_check(page)))) {
 		bad_page(page);
 		return 1;
 	}
@@ -754,7 +756,8 @@ static inline int check_new_page(struct page *page)
 	if (unlikely(page_mapcount(page) |
 		(page->mapping != NULL)  |
 		(atomic_read(&page->_count) != 0)  |
-		(page->flags & PAGE_FLAGS_CHECK_AT_PREP))) {
+		(page->flags & PAGE_FLAGS_CHECK_AT_PREP) |
+		(mem_cgroup_bad_page_check(page)))) {
 		bad_page(page);
 		return 1;
 	}
@@ -5684,4 +5687,5 @@ void dump_page(struct page *page)
 		page, atomic_read(&page->_count), page_mapcount(page),
 		page->mapping, page->index);
 	dump_page_flags(page->flags);
+	mem_cgroup_print_bad_page(page);
 }