linux-kernel - [PATCH v2] slab: Add check for memcg_data's upper bits in folio_memcg

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [day] [month] [year] [list]

Message-Id: <20251014140815.383823-1-hao.ge@linux.dev>
Date: Tue, 14 Oct 2025 22:08:15 +0800
From: Hao Ge <hao.ge@...ux.dev>
To: Vlastimil Babka <vbabka@...e.cz>,
	Alexei Starovoitov <ast@...nel.org>,
	Andrew Morton <akpm@...ux-foundation.org>,
	Johannes Weiner <hannes@...xchg.org>,
	Shakeel Butt <shakeel.butt@...ux.dev>,
	Michal Hocko <mhocko@...nel.org>,
	Roman Gushchin <roman.gushchin@...ux.dev>,
	Muchun Song <muchun.song@...ux.dev>,
	Suren Baghdasaryan <surenb@...gle.com>
Cc: Harry Yoo <harry.yoo@...cle.com>,
	cgroups@...r.kernel.org,
	linux-mm@...ck.org,
	linux-kernel@...r.kernel.org,
	Hao Ge <gehao@...inos.cn>
Subject: [PATCH v2] slab: Add check for memcg_data's upper bits in folio_memcg_kmem

From: Hao Ge <gehao@...inos.cn>

This is because OBJEXTS_ALLOC_FAIL and OBJEXTS_ALLOC_FAIL currently share
the same bit position. Therefore, we cannot simply determine whether
memcg_data still points to the slabobj_ext vector by checking
folio->memcg_data & MEMCG_DATA_OBJEXTS.

We can distinguish between these two cases by checking whether the upper
bits set:

1) MEMCG_DATA_OBJEXTS is set, but upper bits are not set,
   so it should mean obj_exts allocation failed (OBJEXTS_ALLOC_FAIL),
   thus do not report error, or

2) MEMCG_DATA_OBJEXTS is set, and upper bits are also set, so someone
   did not clear a valid folio->memcg_data before freeing the folio
   (report error).

So let's add check for memcg_data's upper bits in folio_memcg_kmem.

Fixes: 7612833192d5 ("slab: Reuse first bit for OBJEXTS_ALLOC_FAIL")
Signed-off-by: Hao Ge <gehao@...inos.cn>
---
v2: Per Vlastimil and Harry's suggestion, instead of introducing a new bit,
    implement this by checking if the highest bit is set.
    Many thanks to Vlastimil and Harry.
---
 include/linux/memcontrol.h | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h
index 873e510d6f8d..f9f7ba14be04 100644
--- a/include/linux/memcontrol.h
+++ b/include/linux/memcontrol.h
@@ -534,7 +534,9 @@ static inline struct mem_cgroup *get_mem_cgroup_from_objcg(struct obj_cgroup *ob
 static inline bool folio_memcg_kmem(struct folio *folio)
 {
 	VM_BUG_ON_PGFLAGS(PageTail(&folio->page), &folio->page);
-	VM_BUG_ON_FOLIO(folio->memcg_data & MEMCG_DATA_OBJEXTS, folio);
+	VM_BUG_ON_FOLIO((folio->memcg_data & MEMCG_DATA_OBJEXTS) &&
+			(folio->memcg_data & ~(ULONG_MAX >> 1)),
+			folio);
 	return folio->memcg_data & MEMCG_DATA_KMEM;
 }
 
-- 
2.25.1