NFSv4: Fix a NFSv4 state manager deadlock

Fix a deadlock whereby the NFSv4 state manager can get stuck in the
delegation return code, waiting for a layout return to complete in
another thread. If the server reboots before that other thread
completes, then we need to be able to start a second state
manager thread in order to perform recovery.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

2 files changed, 13 insertions, 5 deletions

diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h
index 8d59c9655ec4..1b994b527518 100644
--- a/fs/nfs/nfs4_fs.h
+++ b/fs/nfs/nfs4_fs.h
@@ -41,6 +41,8 @@ enum nfs4_client_state {
 	NFS4CLNT_MOVED,
 	NFS4CLNT_LEASE_MOVED,
 	NFS4CLNT_DELEGATION_EXPIRED,
+	NFS4CLNT_RUN_MANAGER,
+	NFS4CLNT_DELEGRETURN_RUNNING,
 };
 
 #define NFS4_RENEW_TIMEOUT		0x01
diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c
index ffea57885394..d8decf2ec48f 100644
--- a/fs/nfs/nfs4state.c
+++ b/fs/nfs/nfs4state.c
@@ -1210,6 +1210,7 @@ void nfs4_schedule_state_manager(struct nfs_client *clp)
 	struct task_struct *task;
 	char buf[INET6_ADDRSTRLEN + sizeof("-manager") + 1];
 
+	set_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state);
 	if (test_and_set_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state) != 0)
 		return;
 	__module_get(THIS_MODULE);
@@ -2503,6 +2504,7 @@ static void nfs4_state_manager(struct nfs_client *clp)
 
 	/* Ensure exclusive access to NFSv4 state */
 	do {
+		clear_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state);
 		if (test_bit(NFS4CLNT_PURGE_STATE, &clp->cl_state)) {
 			section = "purge state";
 			status = nfs4_purge_lease(clp);
@@ -2593,14 +2595,18 @@ static void nfs4_state_manager(struct nfs_client *clp)
 		}
 
 		nfs4_end_drain_session(clp);
-		if (test_and_clear_bit(NFS4CLNT_DELEGRETURN, &clp->cl_state)) {
-			nfs_client_return_marked_delegations(clp);
-			continue;
+		nfs4_clear_state_manager_bit(clp);
+
+		if (!test_and_set_bit(NFS4CLNT_DELEGRETURN_RUNNING, &clp->cl_state)) {
+			if (test_and_clear_bit(NFS4CLNT_DELEGRETURN, &clp->cl_state)) {
+				nfs_client_return_marked_delegations(clp);
+				set_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state);
+			}
+			clear_bit(NFS4CLNT_DELEGRETURN_RUNNING, &clp->cl_state);
 		}
 
-		nfs4_clear_state_manager_bit(clp);
 		/* Did we race with an attempt to give us more work? */
-		if (clp->cl_state == 0)
+		if (!test_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state))
 			return;
 		if (test_and_set_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state) != 0)
 			return;