split_page_memcg() does not care about the returned memcg for kmem
pages, so folio_memcg_charged() should be used, otherwise obj_cgroup_memcg
will complain about this.

Reported-by: syzbot+ef4ecf7b6bdc4157bfa4@syzkaller.appspotmail.com
Signed-off-by: Muchun Song <songmuchun@bytedance.com>
---
 mm/memcontrol.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index e0b3b7ee6de6e..6c84af0a9ede6 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -3018,12 +3018,11 @@ void __memcg_slab_free_hook(struct kmem_cache *s, struct slab *slab,
 void split_page_memcg(struct page *head, int old_order, int new_order)
 {
 	struct folio *folio = page_folio(head);
-	struct mem_cgroup *memcg = folio_memcg(folio);
 	int i;
 	unsigned int old_nr = 1 << old_order;
 	unsigned int new_nr = 1 << new_order;
 
-	if (mem_cgroup_disabled() || !memcg)
+	if (mem_cgroup_disabled() || !folio_memcg_charged(folio))
 		return;
 
 	for (i = new_nr; i < old_nr; i += new_nr)
@@ -3032,7 +3031,7 @@ void split_page_memcg(struct page *head, int old_order, int new_order)
 	if (folio_memcg_kmem(folio))
 		obj_cgroup_get_many(__folio_objcg(folio), old_nr / new_nr - 1);
 	else
-		css_get_many(&memcg->css, old_nr / new_nr - 1);
+		css_get_many(&folio_memcg(folio)->css, old_nr / new_nr - 1);
 }
 
 unsigned long mem_cgroup_usage(struct mem_cgroup *memcg, bool swap)
-- 
2.20.1

On Mon, Aug 19, 2024 at 04:04:15PM GMT, Muchun Song wrote:
> split_page_memcg() does not care about the returned memcg for kmem
> pages, so folio_memcg_charged() should be used, otherwise obj_cgroup_memcg
> will complain about this.

Basically avoid calling folio_memcg() for folio_memcg_kmem(folio),
correct?

> 
> Reported-by: syzbot+ef4ecf7b6bdc4157bfa4@syzkaller.appspotmail.com
> Signed-off-by: Muchun Song <songmuchun@bytedance.com>

Acked-by: Shakeel Butt <shakeel.butt@linux.dev>

> On Aug 19, 2024, at 23:37, Shakeel Butt <shakeel.butt@linux.dev> wrote:
> 
> On Mon, Aug 19, 2024 at 04:04:15PM GMT, Muchun Song wrote:
>> split_page_memcg() does not care about the returned memcg for kmem
>> pages, so folio_memcg_charged() should be used, otherwise obj_cgroup_memcg
>> will complain about this.
> 
> Basically avoid calling folio_memcg() for folio_memcg_kmem(folio),
> correct?

Yes. Unless you hold rcu lock.

> 
>> 
>> Reported-by: syzbot+ef4ecf7b6bdc4157bfa4@syzkaller.appspotmail.com
>> Signed-off-by: Muchun Song <songmuchun@bytedance.com>
> 
> Acked-by: Shakeel Butt <shakeel.butt@linux.dev>
> 

Thanks.