NFSD: Fix possible sleep during nfsd4_release_lockowner()
authorChuck Lever <chuck.lever@oracle.com>
Sat, 21 May 2022 23:06:13 +0000 (19:06 -0400)
committerChuck Lever <chuck.lever@oracle.com>
Thu, 26 May 2022 14:50:49 +0000 (10:50 -0400)
commitce3c4ad7f4ce5db7b4f08a1e237d8dd94b39180b
treede2d585897a4956f207e6361d23da1280c39d149
parentfd5e363eac77ef81542db77ddad0559fa0f9204e
NFSD: Fix possible sleep during nfsd4_release_lockowner()

nfsd4_release_lockowner() holds clp->cl_lock when it calls
check_for_locks(). However, check_for_locks() calls nfsd_file_get()
/ nfsd_file_put() to access the backing inode's flc_posix list, and
nfsd_file_put() can sleep if the inode was recently removed.

Let's instead rely on the stateowner's reference count to gate
whether the release is permitted. This should be a reliable
indication of locks-in-use since file lock operations and
->lm_get_owner take appropriate references, which are released
appropriately when file locks are removed.

Reported-by: Dai Ngo <dai.ngo@oracle.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: stable@vger.kernel.org
fs/nfsd/nfs4state.c