linux-kernel - Forwarded: [PATCH] mm/workingset: fix NULL pointer dereference in lru_gen_test

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: Windows password security audit tool. GUI, reports in PDF.

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

Message-ID: <6936388d.a70a0220.38f243.0075.GAE@google.com>
Date: Sun, 07 Dec 2025 18:31:41 -0800
From: syzbot <syzbot+e008db2ac01e282550ee@...kaller.appspotmail.com>
To: linux-kernel@...r.kernel.org, syzkaller-bugs@...glegroups.com
Subject: Forwarded: [PATCH] mm/workingset: fix NULL pointer dereference in lru_gen_test_recent()

For archival purposes, forwarding an incoming command email to
linux-kernel@...r.kernel.org, syzkaller-bugs@...glegroups.com.

***

Subject: [PATCH] mm/workingset: fix NULL pointer dereference in lru_gen_test_recent()
Author: kartikey406@...il.com

#syz test: git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master

Syzbot reported a general protection fault in lru_gen_test_recent() when
accessing memcg->nodeinfo with a NULL memcg pointer:

  Oops: general protection fault in lru_gen_test_recent+0xfc/0x370
  KASAN: probably user-memory-access in range [0x0000000000004e00-0x0000000000004e07]
  RIP: 0010:lru_gen_test_recent+0xfc/0x370

The crash occurs when unpack_shadow() extracts parameters from a shadow
entry. There are two cases where NULL pointers can be returned:

1. pgdat can be NULL when NODE_DATA(nid) returns NULL for an invalid or
   offlined NUMA node ID stored in the shadow entry.

2. memcg can be NULL when mem_cgroup_from_id() fails to find the memory
   cgroup (e.g., if it was destroyed).

The existing code directly passes these potentially NULL pointers to
mem_cgroup_lruvec(), which dereferences memcg->nodeinfo without checking,
leading to the crash.

Fix this by:
- Checking if pgdat is NULL and returning early if so, as we cannot
  determine page recency without a valid node.
- Checking if memcg is NULL and falling back to pgdat->__lruvec (the
  root memcg's lruvec) instead of calling mem_cgroup_lruvec() which
  would dereference NULL.

Reported-by: syzbot+e008db2ac01e282550ee@...kaller.appspot.com
Link: https://syzkaller.appspot.com/bug?extid=e008db2ac01e282550ee
Signed-off-by: Deepanshu Kartikey <kartikey406@...il.com>
---
 mm/workingset.c | 14 +++++++++++---
 1 file changed, 11 insertions(+), 3 deletions(-)

diff --git a/mm/workingset.c b/mm/workingset.c
index e9f05634747a..335c2d34ac94 100644
--- a/mm/workingset.c
+++ b/mm/workingset.c
@@ -270,10 +270,18 @@ static bool lru_gen_test_recent(void *shadow, struct lruvec **lruvec,
 	struct pglist_data *pgdat;
 
 	unpack_shadow(shadow, &memcg_id, &pgdat, token, workingset);
-
+	if(unlikely(!pgdat))
+		return false;
 	memcg = mem_cgroup_from_id(memcg_id);
-	*lruvec = mem_cgroup_lruvec(memcg, pgdat);
-
+	if (unlikely(!memcg)) {
+		pr_warn("DEBUG: memcg is NULL (memcg_id=%d), pgdat=%p, returning false\n",memcg_id, pgdat);
+		pr_warn("DEBUG: shadow=%p token=%lx workingset=%d\n",shadow, *token, *workingset);
+		 memcg = root_mem_cgroup;
+		*lruvec = &pgdat->__lruvec;
+	} else {
+		*lruvec = mem_cgroup_lruvec(memcg, pgdat);
+	}
+	pr_warn("DEBUG: memcg=%p, lruvec=%p, continuing normally\n", memcg, *lruvec);
 	max_seq = READ_ONCE((*lruvec)->lrugen.max_seq);
 	max_seq &= EVICTION_MASK >> LRU_REFS_WIDTH;
 
-- 
2.43.0