mm: memcontrol: directly access page->memcg_data in mm/page_alloc.c
authorMuchun Song <songmuchun@bytedance.com>
Fri, 30 Apr 2021 05:56:45 +0000 (22:56 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Fri, 30 Apr 2021 18:20:38 +0000 (11:20 -0700)
page_memcg() is not suitable for use by page_expected_state() and
page_bad_reason().  Because it can BUG_ON() for the slab pages when
CONFIG_DEBUG_VM is enabled.  As neither lru, nor kmem, nor slab page
should have anything left in there by the time the page is freed, what
we care about is whether the value of page->memcg_data is 0.  So just
directly access page->memcg_data here.

Link: https://lkml.kernel.org/r/20210319163821.20704-4-songmuchun@bytedance.com
Signed-off-by: Muchun Song <songmuchun@bytedance.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Reviewed-by: Shakeel Butt <shakeelb@google.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Roman Gushchin <guro@fb.com>
Cc: Vladimir Davydov <vdavydov.dev@gmail.com>
Cc: Xiongchun Duan <duanxiongchun@bytedance.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/page_alloc.c

index e2f19bf..56a8103 100644 (file)
@@ -1103,7 +1103,7 @@ static inline bool page_expected_state(struct page *page,
        if (unlikely((unsigned long)page->mapping |
                        page_ref_count(page) |
 #ifdef CONFIG_MEMCG
-                       (unsigned long)page_memcg(page) |
+                       page->memcg_data |
 #endif
                        (page->flags & check_flags)))
                return false;
@@ -1128,7 +1128,7 @@ static const char *page_bad_reason(struct page *page, unsigned long flags)
                        bad_reason = "PAGE_FLAGS_CHECK_AT_FREE flag(s) set";
        }
 #ifdef CONFIG_MEMCG
-       if (unlikely(page_memcg(page)))
+       if (unlikely(page->memcg_data))
                bad_reason = "page still charged to cgroup";
 #endif
        return bad_reason;