[PATCH 1/2] NFSv4.x: Handle bad/dead sessions correctly in nfs41_sequence_process()

[PATCH 1/2] NFSv4.x: Handle bad/dead sessions correctly in nfs41_sequence_process()

[Trond Myklebust]

If the server returns a bad or dead session error, the we don't want
to update the session slot number, but just immediately schedule
recovery and allow it to proceed.

We can/should then remove handling in other places

Fixes: 3453d5708b33 ("NFSv4.1: Avoid false retries when RPC calls are interrupted")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
---
 fs/nfs/nfs4proc.c | 34 +++++++++++++++++++++++++---------
 1 file changed, 25 insertions(+), 9 deletions(-)

diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index 970172dcdba1..d1a7facfa926 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -521,9 +521,7 @@ static int nfs4_do_handle_exception(struct nfs_server *server,
 		case -NFS4ERR_DEADSESSION:
 		case -NFS4ERR_SEQ_FALSE_RETRY:
 		case -NFS4ERR_SEQ_MISORDERED:
-			dprintk("%s ERROR: %d Reset session\n", __func__,
-				errorcode);
-			nfs4_schedule_session_recovery(clp->cl_session, errorcode);
+			/* Handled in nfs41_sequence_process() */
 			goto wait_on_recovery;
 #endif /* defined(CONFIG_NFS_V4_1) */
 		case -NFS4ERR_FILE_OPEN:
@@ -782,6 +780,7 @@ static int nfs41_sequence_process(struct rpc_task *task,
 	struct nfs4_session *session;
 	struct nfs4_slot *slot = res->sr_slot;
 	struct nfs_client *clp;
+	int status;
 	int ret = 1;
 
 	if (slot == NULL)
@@ -793,8 +792,13 @@ static int nfs41_sequence_process(struct rpc_task *task,
 	session = slot->table->session;
 
 	trace_nfs4_sequence_done(session, res);
+
+	status = res->sr_status;
+	if (task->tk_status == -NFS4ERR_DEADSESSION)
+		status = -NFS4ERR_DEADSESSION;
+
 	/* Check the SEQUENCE operation status */
-	switch (res->sr_status) {
+	switch (status) {
 	case 0:
 		/* Mark this sequence number as having been acked */
 		nfs4_slot_sequence_acked(slot, slot->seq_nr);
@@ -866,6 +870,10 @@ static int nfs41_sequence_process(struct rpc_task *task,
 		 */
 		slot->seq_nr = slot->seq_nr_highest_sent;
 		goto out_retry;
+	case -NFS4ERR_BADSESSION:
+	case -NFS4ERR_DEADSESSION:
+	case -NFS4ERR_CONN_NOT_BOUND_TO_SESSION:
+		goto session_recover;
 	default:
 		/* Just update the slot sequence no. */
 		slot->seq_done = 1;
@@ -876,8 +884,10 @@ static int nfs41_sequence_process(struct rpc_task *task,
 out_noaction:
 	return ret;
 session_recover:
-	nfs4_schedule_session_recovery(session, res->sr_status);
-	goto retry_nowait;
+	nfs4_schedule_session_recovery(session, status);
+	dprintk("%s ERROR: %d Reset session\n", __func__, status);
+	nfs41_sequence_free_slot(res);
+	goto out;
 retry_new_seq:
 	++slot->seq_nr;
 retry_nowait:
@@ -2188,7 +2198,6 @@ static int nfs4_handle_delegation_recall_error(struct nfs_server *server, struct
 		case -NFS4ERR_BAD_HIGH_SLOT:
 		case -NFS4ERR_CONN_NOT_BOUND_TO_SESSION:
 		case -NFS4ERR_DEADSESSION:
-			nfs4_schedule_session_recovery(server->nfs_client->cl_session, err);
 			return -EAGAIN;
 		case -NFS4ERR_STALE_CLIENTID:
 		case -NFS4ERR_STALE_STATEID:
@@ -7824,6 +7833,15 @@ nfs41_same_server_scope(struct nfs41_server_scope *a,
 static void
 nfs4_bind_one_conn_to_session_done(struct rpc_task *task, void *calldata)
 {
+	struct nfs41_bind_conn_to_session_args *args = task->tk_msg.rpc_argp;
+	struct nfs_client *clp = args->client;
+
+	switch (task->tk_status) {
+	case -NFS4ERR_BADSESSION:
+	case -NFS4ERR_DEADSESSION:
+		nfs4_schedule_session_recovery(clp->cl_session,
+				task->tk_status);
+	}
 }
 
 static const struct rpc_call_ops nfs4_bind_one_conn_to_session_ops = {
@@ -8871,8 +8889,6 @@ static int nfs41_reclaim_complete_handle_errors(struct rpc_task *task, struct nf
 	case -NFS4ERR_BADSESSION:
 	case -NFS4ERR_DEADSESSION:
 	case -NFS4ERR_CONN_NOT_BOUND_TO_SESSION:
-		nfs4_schedule_session_recovery(clp->cl_session,
-				task->tk_status);
 		break;
 	default:
 		nfs4_schedule_lease_recovery(clp);
-- 
2.23.0

[PATCH 2/2] NFSv4.x: Drop the slot if nfs4_delegreturn_prepare waits for layoutreturn

[Trond Myklebust]

If nfs4_delegreturn_prepare needs to wait for a layoutreturn to complete
then make sure we drop the sequence slot if we hold it.

Fixes: 1c5bd76d17cc ("pNFS: Enable layoutreturn operation for return-on-close")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
---
 fs/nfs/nfs4proc.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index d1a7facfa926..2d3c12f68204 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -6256,8 +6256,10 @@ static void nfs4_delegreturn_prepare(struct rpc_task *task, void *data)
 
 	d_data = (struct nfs4_delegreturndata *)data;
 
-	if (!d_data->lr.roc && nfs4_wait_on_layoutreturn(d_data->inode, task))
+	if (!d_data->lr.roc && nfs4_wait_on_layoutreturn(d_data->inode, task)) {
+		nfs4_sequence_done(task, &d_data->res.seq_res);
 		return;
+	}
 
 	lo = d_data->args.lr_args ? d_data->args.lr_args->layout : NULL;
 	if (lo && !pnfs_layout_is_valid(lo)) {
-- 
2.23.0