]> git.ipfire.org Git - thirdparty/linux.git/commit
ksmbd: fix use-after-free in __ksmbd_close_fd() via durable scavenger
authorNamjae Jeon <linkinjeon@kernel.org>
Sat, 4 Apr 2026 12:09:02 +0000 (21:09 +0900)
committerSteve French <stfrench@microsoft.com>
Sun, 12 Apr 2026 23:07:54 +0000 (18:07 -0500)
commit235e32320a470fcd3998fb3774f2290a0eb302a1
tree233707af6173c3ba2c58f1ef0de7afdb08b86ad0
parent3df614ebc976bb23d2f99734695c1b7ff126d7fc
ksmbd: fix use-after-free in __ksmbd_close_fd() via durable scavenger

When a durable file handle survives session disconnect (TCP close without
SMB2_LOGOFF), session_fd_check() sets fp->conn = NULL to preserve the
handle for later reconnection. However, it did not clean up the byte-range
locks on fp->lock_list.

Later, when the durable scavenger thread times out and calls
__ksmbd_close_fd(NULL, fp), the lock cleanup loop did:

    spin_lock(&fp->conn->llist_lock);

This caused a slab use-after-free because fp->conn was NULL and the
original connection object had already been freed by
ksmbd_tcp_disconnect().

The root cause is asymmetric cleanup: lock entries (smb_lock->clist) were
left dangling on the freed conn->lock_list while fp->conn was nulled out.

To fix this issue properly, we need to handle the lifetime of
smb_lock->clist across three paths:
 - Safely skip clist deletion when list is empty and fp->conn is NULL.
 - Remove the lock from the old connection's lock_list in
   session_fd_check()
 - Re-add the lock to the new connection's lock_list in
   ksmbd_reopen_durable_fd().

Fixes: c8efcc786146 ("ksmbd: add support for durable handles v1/v2")
Co-developed-by: munan Huang <munanevil@gmail.com>
Signed-off-by: munan Huang <munanevil@gmail.com>
Reviewed-by: ChenXiaoSong <chenxiaosong@kylinos.cn>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
fs/smb/server/vfs_cache.c