[PATCH v2 1/6] NFSD: Prevent lock owner use-after-free during client teardown

[Chuck Lever <cel@kernel.org>]

After __destroy_client() releases a client's open owners, a lock owner
whose only remaining reference is a blocked lock (nbl) is left on
cl_ownerstr_hashtbl.  client_has_state() does not account for a bare
lock owner, so DESTROY_CLIENTID can reach __destroy_client() with these
lock owners still present.

__destroy_client() then walks cl_ownerstr_hashtbl and calls
remove_blocked_locks() on each lock owner without holding a reference.
Freeing a blocked lock drops the lock owner reference held through the
file_lock's flc_owner, so the per-net laundromat, which reaps timed-out
blocked locks from nn->blocked_locks_lru independently of client state,
can free the same lock owner concurrently.  The two paths serialize on
blocked_locks_lock for the list splice only, not for the lock owner's
lifetime.  The laundromat can therefore free the lock owner while
__destroy_client() is about to dereference it, and the freed, zeroed
slab object produces a NULL dereference in remove_blocked_locks().

nfsd4_release_lockowner() holds a reference across the same call;
__destroy_client() does not.  Hold cl_lock across the walk, and
take a reference and unhash each lock owner before dropping the
lock, so the laundromat cannot reap a blocked lock and free the
lock owner underneath this loop.  cl_lock is released before
remove_blocked_locks() and nfs4_put_stateowner(), which take
blocked_locks_lock and cl_lock respectively.

Reported-by: Wolfgang Walter <linux@stwm.de>
Closes: https://lore.kernel.org/linux-nfs/6eccafaaaa60651ef091257c3439c46b@stwm.de/
Fixes: 68ef3bc31664 ("nfsd: remove blocked locks on client teardown")
Signed-off-by: Chuck Lever <cel@kernel.org>
---
 fs/nfsd/nfs4state.c | 16 +++++++++++++---
 1 file changed, 13 insertions(+), 3 deletions(-)

diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
index a4398dc861a5..e000ed3e96e9 100644
--- a/fs/nfsd/nfs4state.c
+++ b/fs/nfsd/nfs4state.c
@@ -2758,14 +2758,24 @@ __destroy_client(struct nfs4_client *clp)
 		release_openowner(oo);
 	}
 	for (i = 0; i < OWNER_HASH_SIZE; i++) {
-		struct nfs4_stateowner *so, *tmp;
+		struct nfs4_stateowner *so;
 
-		list_for_each_entry_safe(so, tmp, &clp->cl_ownerstr_hashtbl[i],
-					 so_strhash) {
+		spin_lock(&clp->cl_lock);
+		while (!list_empty(&clp->cl_ownerstr_hashtbl[i])) {
+			so = list_first_entry(&clp->cl_ownerstr_hashtbl[i],
+					      struct nfs4_stateowner, so_strhash);
 			/* Should be no openowners at this point */
 			WARN_ON_ONCE(so->so_is_open_owner);
+			nfs4_get_stateowner(so);
+			unhash_lockowner_locked(lockowner(so));
+			spin_unlock(&clp->cl_lock);
+
 			remove_blocked_locks(lockowner(so));
+			nfs4_put_stateowner(so);
+
+			spin_lock(&clp->cl_lock);
 		}
+		spin_unlock(&clp->cl_lock);
 	}
 	nfsd4_return_all_client_layouts(clp);
 	nfsd4_shutdown_copy(clp);

-- 
2.54.0