From aede517207b2bf4e5176492ce3be0e45feb581bd Mon Sep 17 00:00:00 2001
From: Chuck Lever <chuck.lever@oracle.com>
Date: Mon, 19 Jul 2021 10:48:04 -0400
Subject: SUNRPC: Refactor rpc_ping()

Make it use the rpc_null_call_helper() so that it can share the
new rpc_call_ops structure to be introduced in the next patch.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 net/sunrpc/clnt.c | 24 +++++++++++++-----------
 1 file changed, 13 insertions(+), 11 deletions(-)

(limited to 'net')

diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index 8b4de70e8ead..ca2000d8cf64 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -2694,17 +2694,6 @@ static const struct rpc_procinfo rpcproc_null = {
 	.p_decode = rpcproc_decode_null,
 };
 
-static int rpc_ping(struct rpc_clnt *clnt)
-{
-	struct rpc_message msg = {
-		.rpc_proc = &rpcproc_null,
-	};
-	int err;
-	err = rpc_call_sync(clnt, &msg, RPC_TASK_SOFT | RPC_TASK_SOFTCONN |
-			    RPC_TASK_NULLCREDS);
-	return err;
-}
-
 static
 struct rpc_task *rpc_call_null_helper(struct rpc_clnt *clnt,
 		struct rpc_xprt *xprt, struct rpc_cred *cred, int flags,
@@ -2733,6 +2722,19 @@ struct rpc_task *rpc_call_null(struct rpc_clnt *clnt, struct rpc_cred *cred, int
 }
 EXPORT_SYMBOL_GPL(rpc_call_null);
 
+static int rpc_ping(struct rpc_clnt *clnt)
+{
+	struct rpc_task	*task;
+	int status;
+
+	task = rpc_call_null_helper(clnt, NULL, NULL, 0, NULL, NULL);
+	if (IS_ERR(task))
+		return PTR_ERR(task);
+	status = task->tk_status;
+	rpc_put_task(task);
+	return status;
+}
+
 struct rpc_cb_add_xprt_calldata {
 	struct rpc_xprt_switch *xps;
 	struct rpc_xprt *xprt;
-- 
cgit v1.2.3


From 823c73d0c539a0c1740c7929a7aad3f844e69d70 Mon Sep 17 00:00:00 2001
From: Chuck Lever <chuck.lever@oracle.com>
Date: Mon, 19 Jul 2021 10:48:10 -0400
Subject: SUNRPC: Unset RPC_TASK_NO_RETRANS_TIMEOUT for NULL RPCs

In some rare failure modes, the server is actually reading the
transport, but then just dropping the requests on the floor.
TCP_USER_TIMEOUT cannot detect that case.

Prevent such a stuck server from pinning client resources
indefinitely by ensuring that certain idempotent requests
(such as NULL) can time out even if the connection is still
operational.

Otherwise rpc_bind_new_program(), gss_destroy_cred(), or
rpc_clnt_test_and_add_xprt() can wait forever.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 net/sunrpc/clnt.c | 15 ++++++++++++++-
 1 file changed, 14 insertions(+), 1 deletion(-)

(limited to 'net')

diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index ca2000d8cf64..d34737a8a68a 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -2694,6 +2694,18 @@ static const struct rpc_procinfo rpcproc_null = {
 	.p_decode = rpcproc_decode_null,
 };
 
+static void
+rpc_null_call_prepare(struct rpc_task *task, void *data)
+{
+	task->tk_flags &= ~RPC_TASK_NO_RETRANS_TIMEOUT;
+	rpc_call_start(task);
+}
+
+static const struct rpc_call_ops rpc_null_ops = {
+	.rpc_call_prepare = rpc_null_call_prepare,
+	.rpc_call_done = rpc_default_callback,
+};
+
 static
 struct rpc_task *rpc_call_null_helper(struct rpc_clnt *clnt,
 		struct rpc_xprt *xprt, struct rpc_cred *cred, int flags,
@@ -2707,7 +2719,7 @@ struct rpc_task *rpc_call_null_helper(struct rpc_clnt *clnt,
 		.rpc_xprt = xprt,
 		.rpc_message = &msg,
 		.rpc_op_cred = cred,
-		.callback_ops = (ops != NULL) ? ops : &rpc_default_ops,
+		.callback_ops = ops ?: &rpc_null_ops,
 		.callback_data = data,
 		.flags = flags | RPC_TASK_SOFT | RPC_TASK_SOFTCONN |
 			 RPC_TASK_NULLCREDS,
@@ -2758,6 +2770,7 @@ static void rpc_cb_add_xprt_release(void *calldata)
 }
 
 static const struct rpc_call_ops rpc_cb_add_xprt_call_ops = {
+	.rpc_call_prepare = rpc_null_call_prepare,
 	.rpc_call_done = rpc_cb_add_xprt_done,
 	.rpc_release = rpc_cb_add_xprt_release,
 };
-- 
cgit v1.2.3


From 1143129e4d0d27740ce680d2fb0161ad4f27aa7e Mon Sep 17 00:00:00 2001
From: Chuck Lever <chuck.lever@oracle.com>
Date: Mon, 2 Aug 2021 14:44:17 -0400
Subject: xprtrdma: Disconnect after an ib_post_send() immediate error

ib_post_send() does not disconnect the QP when it returns an
immediate error. Thus, the code that posts LocalInv has to
explicitly disconnect after an immediate error. This is just
like the frwr_send() callers handle it.

If a disconnect isn't done here, the transport deadlocks.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 net/sunrpc/xprtrdma/frwr_ops.c  | 8 ++++++++
 net/sunrpc/xprtrdma/verbs.c     | 2 +-
 net/sunrpc/xprtrdma/xprt_rdma.h | 1 +
 3 files changed, 10 insertions(+), 1 deletion(-)

(limited to 'net')

diff --git a/net/sunrpc/xprtrdma/frwr_ops.c b/net/sunrpc/xprtrdma/frwr_ops.c
index 229fcc9a9064..754c5dffe127 100644
--- a/net/sunrpc/xprtrdma/frwr_ops.c
+++ b/net/sunrpc/xprtrdma/frwr_ops.c
@@ -557,6 +557,10 @@ void frwr_unmap_sync(struct rpcrdma_xprt *r_xprt, struct rpcrdma_req *req)
 
 	/* On error, the MRs get destroyed once the QP has drained. */
 	trace_xprtrdma_post_linv_err(req, rc);
+
+	/* Force a connection loss to ensure complete recovery.
+	 */
+	rpcrdma_force_disconnect(ep);
 }
 
 /**
@@ -653,4 +657,8 @@ void frwr_unmap_async(struct rpcrdma_xprt *r_xprt, struct rpcrdma_req *req)
 	 * retransmission.
 	 */
 	rpcrdma_unpin_rqst(req->rl_reply);
+
+	/* Force a connection loss to ensure complete recovery.
+	 */
+	rpcrdma_force_disconnect(ep);
 }
diff --git a/net/sunrpc/xprtrdma/verbs.c b/net/sunrpc/xprtrdma/verbs.c
index 649c23518ec0..c1797ea19418 100644
--- a/net/sunrpc/xprtrdma/verbs.c
+++ b/net/sunrpc/xprtrdma/verbs.c
@@ -124,7 +124,7 @@ static void rpcrdma_xprt_drain(struct rpcrdma_xprt *r_xprt)
  * connection is closed or lost. (The important thing is it needs
  * to be invoked "at least" once).
  */
-static void rpcrdma_force_disconnect(struct rpcrdma_ep *ep)
+void rpcrdma_force_disconnect(struct rpcrdma_ep *ep)
 {
 	if (atomic_add_unless(&ep->re_force_disconnect, 1, 1))
 		xprt_force_disconnect(ep->re_xprt);
diff --git a/net/sunrpc/xprtrdma/xprt_rdma.h b/net/sunrpc/xprtrdma/xprt_rdma.h
index 5d231d94e944..927e20a2c04e 100644
--- a/net/sunrpc/xprtrdma/xprt_rdma.h
+++ b/net/sunrpc/xprtrdma/xprt_rdma.h
@@ -454,6 +454,7 @@ extern unsigned int xprt_rdma_memreg_strategy;
 /*
  * Endpoint calls - xprtrdma/verbs.c
  */
+void rpcrdma_force_disconnect(struct rpcrdma_ep *ep);
 void rpcrdma_flush_disconnect(struct rpcrdma_xprt *r_xprt, struct ib_wc *wc);
 int rpcrdma_xprt_connect(struct rpcrdma_xprt *r_xprt);
 void rpcrdma_xprt_disconnect(struct rpcrdma_xprt *r_xprt);
-- 
cgit v1.2.3


From 97480cae13ca3a9c1de3eb6fd66cf9650a60db42 Mon Sep 17 00:00:00 2001
From: Chuck Lever <chuck.lever@oracle.com>
Date: Mon, 2 Aug 2021 14:44:24 -0400
Subject: xprtrdma: Put rpcrdma_reps before waking the tear-down completion

Ensure the tear-down completion is awoken only /after/ we've stopped
fiddling with rpcrdma_rep objects in rpcrdma_post_recvs().

Fixes: 15788d1d1077 ("xprtrdma: Do not refresh Receive Queue while it is draining")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 net/sunrpc/xprtrdma/verbs.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

(limited to 'net')

diff --git a/net/sunrpc/xprtrdma/verbs.c b/net/sunrpc/xprtrdma/verbs.c
index c1797ea19418..016f10a781b4 100644
--- a/net/sunrpc/xprtrdma/verbs.c
+++ b/net/sunrpc/xprtrdma/verbs.c
@@ -1416,11 +1416,6 @@ void rpcrdma_post_recvs(struct rpcrdma_xprt *r_xprt, int needed, bool temp)
 
 	rc = ib_post_recv(ep->re_id->qp, wr,
 			  (const struct ib_recv_wr **)&bad_wr);
-	if (atomic_dec_return(&ep->re_receiving) > 0)
-		complete(&ep->re_done);
-
-out:
-	trace_xprtrdma_post_recvs(r_xprt, count, rc);
 	if (rc) {
 		for (wr = bad_wr; wr;) {
 			struct rpcrdma_rep *rep;
@@ -1431,6 +1426,11 @@ out:
 			--count;
 		}
 	}
+	if (atomic_dec_return(&ep->re_receiving) > 0)
+		complete(&ep->re_done);
+
+out:
+	trace_xprtrdma_post_recvs(r_xprt, count, rc);
 	ep->re_receive_count += count;
 	return;
 }
-- 
cgit v1.2.3


From 683f31c3ab2e60b323b9b88c0ac389ff9cacec1a Mon Sep 17 00:00:00 2001
From: Chuck Lever <chuck.lever@oracle.com>
Date: Mon, 2 Aug 2021 14:44:30 -0400
Subject: xprtrdma: Add xprtrdma_post_recvs_err() tracepoint

In the vast majority of cases, rc=0. Don't record that in the
post_recvs tracepoint. Instead, add a separate tracepoint that can
be left enabled all the time to capture the very rare immediate
errors returned by ib_post_recv().

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 include/trace/events/rpcrdma.h | 41 ++++++++++++++++++++++++++++++++++-------
 net/sunrpc/xprtrdma/verbs.c    |  3 ++-
 2 files changed, 36 insertions(+), 8 deletions(-)

(limited to 'net')

diff --git a/include/trace/events/rpcrdma.h b/include/trace/events/rpcrdma.h
index bd55908c1bef..d65a84bd040c 100644
--- a/include/trace/events/rpcrdma.h
+++ b/include/trace/events/rpcrdma.h
@@ -818,16 +818,14 @@ TRACE_EVENT(xprtrdma_post_recv,
 TRACE_EVENT(xprtrdma_post_recvs,
 	TP_PROTO(
 		const struct rpcrdma_xprt *r_xprt,
-		unsigned int count,
-		int status
+		unsigned int count
 	),
 
-	TP_ARGS(r_xprt, count, status),
+	TP_ARGS(r_xprt, count),
 
 	TP_STRUCT__entry(
 		__field(u32, cq_id)
 		__field(unsigned int, count)
-		__field(int, status)
 		__field(int, posted)
 		__string(addr, rpcrdma_addrstr(r_xprt))
 		__string(port, rpcrdma_portstr(r_xprt))
@@ -838,15 +836,44 @@ TRACE_EVENT(xprtrdma_post_recvs,
 
 		__entry->cq_id = ep->re_attr.recv_cq->res.id;
 		__entry->count = count;
-		__entry->status = status;
 		__entry->posted = ep->re_receive_count;
 		__assign_str(addr, rpcrdma_addrstr(r_xprt));
 		__assign_str(port, rpcrdma_portstr(r_xprt));
 	),
 
-	TP_printk("peer=[%s]:%s cq.id=%d %u new recvs, %d active (rc %d)",
+	TP_printk("peer=[%s]:%s cq.id=%d %u new recvs, %d active",
+		__get_str(addr), __get_str(port), __entry->cq_id,
+		__entry->count, __entry->posted
+	)
+);
+
+TRACE_EVENT(xprtrdma_post_recvs_err,
+	TP_PROTO(
+		const struct rpcrdma_xprt *r_xprt,
+		int status
+	),
+
+	TP_ARGS(r_xprt, status),
+
+	TP_STRUCT__entry(
+		__field(u32, cq_id)
+		__field(int, status)
+		__string(addr, rpcrdma_addrstr(r_xprt))
+		__string(port, rpcrdma_portstr(r_xprt))
+	),
+
+	TP_fast_assign(
+		const struct rpcrdma_ep *ep = r_xprt->rx_ep;
+
+		__entry->cq_id = ep->re_attr.recv_cq->res.id;
+		__entry->status = status;
+		__assign_str(addr, rpcrdma_addrstr(r_xprt));
+		__assign_str(port, rpcrdma_portstr(r_xprt));
+	),
+
+	TP_printk("peer=[%s]:%s cq.id=%d rc=%d",
 		__get_str(addr), __get_str(port), __entry->cq_id,
-		__entry->count, __entry->posted, __entry->status
+		__entry->status
 	)
 );
 
diff --git a/net/sunrpc/xprtrdma/verbs.c b/net/sunrpc/xprtrdma/verbs.c
index 016f10a781b4..1e9041c022b6 100644
--- a/net/sunrpc/xprtrdma/verbs.c
+++ b/net/sunrpc/xprtrdma/verbs.c
@@ -1417,6 +1417,7 @@ void rpcrdma_post_recvs(struct rpcrdma_xprt *r_xprt, int needed, bool temp)
 	rc = ib_post_recv(ep->re_id->qp, wr,
 			  (const struct ib_recv_wr **)&bad_wr);
 	if (rc) {
+		trace_xprtrdma_post_recvs_err(r_xprt, rc);
 		for (wr = bad_wr; wr;) {
 			struct rpcrdma_rep *rep;
 
@@ -1430,7 +1431,7 @@ void rpcrdma_post_recvs(struct rpcrdma_xprt *r_xprt, int needed, bool temp)
 		complete(&ep->re_done);
 
 out:
-	trace_xprtrdma_post_recvs(r_xprt, count, rc);
+	trace_xprtrdma_post_recvs(r_xprt, count);
 	ep->re_receive_count += count;
 	return;
 }
-- 
cgit v1.2.3


From d9ae8134f253e8d0e15b1f0127af3b8b5552b90c Mon Sep 17 00:00:00 2001
From: Chuck Lever <chuck.lever@oracle.com>
Date: Mon, 2 Aug 2021 14:44:36 -0400
Subject: xprtrdma: Add an xprtrdma_post_send_err tracepoint

Unlike xprtrdma_post_send(), this one can be left enabled all the
time, and should almost never fire. But we do want to know about
immediate errors when they happen.

Note that there is already a similar post_linv_err tracepoint.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 include/trace/events/rpcrdma.h | 33 +++++++++++++++++++++++++++++++++
 net/sunrpc/xprtrdma/frwr_ops.c |  6 +++++-
 2 files changed, 38 insertions(+), 1 deletion(-)

(limited to 'net')

diff --git a/include/trace/events/rpcrdma.h b/include/trace/events/rpcrdma.h
index d65a84bd040c..de4195499592 100644
--- a/include/trace/events/rpcrdma.h
+++ b/include/trace/events/rpcrdma.h
@@ -793,6 +793,39 @@ TRACE_EVENT(xprtrdma_post_send,
 	)
 );
 
+TRACE_EVENT(xprtrdma_post_send_err,
+	TP_PROTO(
+		const struct rpcrdma_xprt *r_xprt,
+		const struct rpcrdma_req *req,
+		int rc
+	),
+
+	TP_ARGS(r_xprt, req, rc),
+
+	TP_STRUCT__entry(
+		__field(u32, cq_id)
+		__field(unsigned int, task_id)
+		__field(unsigned int, client_id)
+		__field(int, rc)
+	),
+
+	TP_fast_assign(
+		const struct rpc_rqst *rqst = &req->rl_slot;
+		const struct rpcrdma_ep *ep = r_xprt->rx_ep;
+
+		__entry->cq_id = ep ? ep->re_attr.recv_cq->res.id : 0;
+		__entry->task_id = rqst->rq_task->tk_pid;
+		__entry->client_id = rqst->rq_task->tk_client ?
+				     rqst->rq_task->tk_client->cl_clid : -1;
+		__entry->rc = rc;
+	),
+
+	TP_printk("task:%u@%u cq.id=%u rc=%d",
+		__entry->task_id, __entry->client_id,
+		__entry->cq_id, __entry->rc
+	)
+);
+
 TRACE_EVENT(xprtrdma_post_recv,
 	TP_PROTO(
 		const struct rpcrdma_rep *rep
diff --git a/net/sunrpc/xprtrdma/frwr_ops.c b/net/sunrpc/xprtrdma/frwr_ops.c
index 754c5dffe127..f700b34a5bfd 100644
--- a/net/sunrpc/xprtrdma/frwr_ops.c
+++ b/net/sunrpc/xprtrdma/frwr_ops.c
@@ -394,6 +394,7 @@ int frwr_send(struct rpcrdma_xprt *r_xprt, struct rpcrdma_req *req)
 	struct rpcrdma_ep *ep = r_xprt->rx_ep;
 	struct rpcrdma_mr *mr;
 	unsigned int num_wrs;
+	int ret;
 
 	num_wrs = 1;
 	post_wr = send_wr;
@@ -420,7 +421,10 @@ int frwr_send(struct rpcrdma_xprt *r_xprt, struct rpcrdma_req *req)
 	}
 
 	trace_xprtrdma_post_send(req);
-	return ib_post_send(ep->re_id->qp, post_wr, NULL);
+	ret = ib_post_send(ep->re_id->qp, post_wr, NULL);
+	if (ret)
+		trace_xprtrdma_post_send_err(r_xprt, req, ret);
+	return ret;
 }
 
 /**
-- 
cgit v1.2.3


From 8d863b1f0541ea540cb47d1454eb060963a79d5f Mon Sep 17 00:00:00 2001
From: Chuck Lever <chuck.lever@oracle.com>
Date: Mon, 2 Aug 2021 14:44:42 -0400
Subject: xprtrdma: Eliminate rpcrdma_post_sends()

Clean up.

Now that there is only one registration mode, there is only one
target "post_send" method: frwr_send(). rpcrdma_post_sends() no
longer adds much value, especially since all of its call sites
ignore the return code value except to check if it's non-zero.

Just have them call frwr_send() directly instead.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 net/sunrpc/xprtrdma/backchannel.c |  2 +-
 net/sunrpc/xprtrdma/transport.c   |  2 +-
 net/sunrpc/xprtrdma/verbs.c       | 15 ---------------
 net/sunrpc/xprtrdma/xprt_rdma.h   |  1 -
 4 files changed, 2 insertions(+), 18 deletions(-)

(limited to 'net')

diff --git a/net/sunrpc/xprtrdma/backchannel.c b/net/sunrpc/xprtrdma/backchannel.c
index 1151efd09b27..17f174d6ea3b 100644
--- a/net/sunrpc/xprtrdma/backchannel.c
+++ b/net/sunrpc/xprtrdma/backchannel.c
@@ -115,7 +115,7 @@ int xprt_rdma_bc_send_reply(struct rpc_rqst *rqst)
 	if (rc < 0)
 		goto failed_marshal;
 
-	if (rpcrdma_post_sends(r_xprt, req))
+	if (frwr_send(r_xprt, req))
 		goto drop_connection;
 	return 0;
 
diff --git a/net/sunrpc/xprtrdma/transport.c b/net/sunrpc/xprtrdma/transport.c
index 9c2ffc67c0fd..a463400ed5a3 100644
--- a/net/sunrpc/xprtrdma/transport.c
+++ b/net/sunrpc/xprtrdma/transport.c
@@ -661,7 +661,7 @@ xprt_rdma_send_request(struct rpc_rqst *rqst)
 		goto drop_connection;
 	rqst->rq_xtime = ktime_get();
 
-	if (rpcrdma_post_sends(r_xprt, req))
+	if (frwr_send(r_xprt, req))
 		goto drop_connection;
 
 	rqst->rq_xmit_bytes_sent += rqst->rq_snd_buf.len;
diff --git a/net/sunrpc/xprtrdma/verbs.c b/net/sunrpc/xprtrdma/verbs.c
index 1e9041c022b6..aaec3c9be8db 100644
--- a/net/sunrpc/xprtrdma/verbs.c
+++ b/net/sunrpc/xprtrdma/verbs.c
@@ -1349,21 +1349,6 @@ static void rpcrdma_regbuf_free(struct rpcrdma_regbuf *rb)
 	kfree(rb);
 }
 
-/**
- * rpcrdma_post_sends - Post WRs to a transport's Send Queue
- * @r_xprt: controlling transport instance
- * @req: rpcrdma_req containing the Send WR to post
- *
- * Returns 0 if the post was successful, otherwise -ENOTCONN
- * is returned.
- */
-int rpcrdma_post_sends(struct rpcrdma_xprt *r_xprt, struct rpcrdma_req *req)
-{
-	if (frwr_send(r_xprt, req))
-		return -ENOTCONN;
-	return 0;
-}
-
 /**
  * rpcrdma_post_recvs - Refill the Receive Queue
  * @r_xprt: controlling transport instance
diff --git a/net/sunrpc/xprtrdma/xprt_rdma.h b/net/sunrpc/xprtrdma/xprt_rdma.h
index 927e20a2c04e..d91f54eae00b 100644
--- a/net/sunrpc/xprtrdma/xprt_rdma.h
+++ b/net/sunrpc/xprtrdma/xprt_rdma.h
@@ -459,7 +459,6 @@ void rpcrdma_flush_disconnect(struct rpcrdma_xprt *r_xprt, struct ib_wc *wc);
 int rpcrdma_xprt_connect(struct rpcrdma_xprt *r_xprt);
 void rpcrdma_xprt_disconnect(struct rpcrdma_xprt *r_xprt);
 
-int rpcrdma_post_sends(struct rpcrdma_xprt *r_xprt, struct rpcrdma_req *req);
 void rpcrdma_post_recvs(struct rpcrdma_xprt *r_xprt, int needed, bool temp);
 
 /*
-- 
cgit v1.2.3


From 71d3d0ebc894294ef9454e45a3ac2e9ba60b3351 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <trond.myklebust@hammerspace.com>
Date: Mon, 26 Jul 2021 08:01:27 -0400
Subject: SUNRPC: Convert rpc_client refcount to use refcount_t

There are now tools in the refcount library that allow us to convert the
client shutdown code.

Reported-by: Xiyu Yang <xiyuyang19@fudan.edu.cn>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 include/linux/sunrpc/clnt.h          |  3 ++-
 net/sunrpc/auth_gss/gss_rpc_upcall.c |  2 +-
 net/sunrpc/clnt.c                    | 22 ++++++++++------------
 net/sunrpc/debugfs.c                 |  2 +-
 net/sunrpc/rpc_pipe.c                |  2 +-
 5 files changed, 15 insertions(+), 16 deletions(-)

(limited to 'net')

diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h
index 8b5d5c97553e..b2edd5fc2f0c 100644
--- a/include/linux/sunrpc/clnt.h
+++ b/include/linux/sunrpc/clnt.h
@@ -14,6 +14,7 @@
 #include <linux/socket.h>
 #include <linux/in.h>
 #include <linux/in6.h>
+#include <linux/refcount.h>
 
 #include <linux/sunrpc/msg_prot.h>
 #include <linux/sunrpc/sched.h>
@@ -35,7 +36,7 @@ struct rpc_sysfs_client;
  * The high-level client handle
  */
 struct rpc_clnt {
-	atomic_t		cl_count;	/* Number of references */
+	refcount_t		cl_count;	/* Number of references */
 	unsigned int		cl_clid;	/* client id */
 	struct list_head	cl_clients;	/* Global list of clients */
 	struct list_head	cl_tasks;	/* List of tasks */
diff --git a/net/sunrpc/auth_gss/gss_rpc_upcall.c b/net/sunrpc/auth_gss/gss_rpc_upcall.c
index d1c003a25b0f..61c276bddaf2 100644
--- a/net/sunrpc/auth_gss/gss_rpc_upcall.c
+++ b/net/sunrpc/auth_gss/gss_rpc_upcall.c
@@ -160,7 +160,7 @@ static struct rpc_clnt *get_gssp_clnt(struct sunrpc_net *sn)
 	mutex_lock(&sn->gssp_lock);
 	clnt = sn->gssp_clnt;
 	if (clnt)
-		atomic_inc(&clnt->cl_count);
+		refcount_inc(&clnt->cl_count);
 	mutex_unlock(&sn->gssp_lock);
 	return clnt;
 }
diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index d34737a8a68a..a5b7f6e34d15 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -167,7 +167,7 @@ static int rpc_clnt_skip_event(struct rpc_clnt *clnt, unsigned long event)
 	case RPC_PIPEFS_MOUNT:
 		if (clnt->cl_pipedir_objects.pdh_dentry != NULL)
 			return 1;
-		if (atomic_read(&clnt->cl_count) == 0)
+		if (refcount_read(&clnt->cl_count) == 0)
 			return 1;
 		break;
 	case RPC_PIPEFS_UMOUNT:
@@ -419,7 +419,7 @@ static struct rpc_clnt * rpc_new_client(const struct rpc_create_args *args,
 	clnt->cl_rtt = &clnt->cl_rtt_default;
 	rpc_init_rtt(&clnt->cl_rtt_default, clnt->cl_timeout->to_initval);
 
-	atomic_set(&clnt->cl_count, 1);
+	refcount_set(&clnt->cl_count, 1);
 
 	if (nodename == NULL)
 		nodename = utsname()->nodename;
@@ -431,7 +431,7 @@ static struct rpc_clnt * rpc_new_client(const struct rpc_create_args *args,
 	if (err)
 		goto out_no_path;
 	if (parent)
-		atomic_inc(&parent->cl_count);
+		refcount_inc(&parent->cl_count);
 
 	trace_rpc_clnt_new(clnt, xprt, program->name, args->servername);
 	return clnt;
@@ -918,18 +918,16 @@ rpc_free_client(struct rpc_clnt *clnt)
 static struct rpc_clnt *
 rpc_free_auth(struct rpc_clnt *clnt)
 {
-	if (clnt->cl_auth == NULL)
-		return rpc_free_client(clnt);
-
 	/*
 	 * Note: RPCSEC_GSS may need to send NULL RPC calls in order to
 	 *       release remaining GSS contexts. This mechanism ensures
 	 *       that it can do so safely.
 	 */
-	atomic_inc(&clnt->cl_count);
-	rpcauth_release(clnt->cl_auth);
-	clnt->cl_auth = NULL;
-	if (atomic_dec_and_test(&clnt->cl_count))
+	if (clnt->cl_auth != NULL) {
+		rpcauth_release(clnt->cl_auth);
+		clnt->cl_auth = NULL;
+	}
+	if (refcount_dec_and_test(&clnt->cl_count))
 		return rpc_free_client(clnt);
 	return NULL;
 }
@@ -943,7 +941,7 @@ rpc_release_client(struct rpc_clnt *clnt)
 	do {
 		if (list_empty(&clnt->cl_tasks))
 			wake_up(&destroy_wait);
-		if (!atomic_dec_and_test(&clnt->cl_count))
+		if (refcount_dec_not_one(&clnt->cl_count))
 			break;
 		clnt = rpc_free_auth(clnt);
 	} while (clnt != NULL);
@@ -1082,7 +1080,7 @@ void rpc_task_set_client(struct rpc_task *task, struct rpc_clnt *clnt)
 	if (clnt != NULL) {
 		rpc_task_set_transport(task, clnt);
 		task->tk_client = clnt;
-		atomic_inc(&clnt->cl_count);
+		refcount_inc(&clnt->cl_count);
 		if (clnt->cl_softrtry)
 			task->tk_flags |= RPC_TASK_SOFT;
 		if (clnt->cl_softerr)
diff --git a/net/sunrpc/debugfs.c b/net/sunrpc/debugfs.c
index 56029e3af6ff..79995eb95927 100644
--- a/net/sunrpc/debugfs.c
+++ b/net/sunrpc/debugfs.c
@@ -90,7 +90,7 @@ static int tasks_open(struct inode *inode, struct file *filp)
 		struct seq_file *seq = filp->private_data;
 		struct rpc_clnt *clnt = seq->private = inode->i_private;
 
-		if (!atomic_inc_not_zero(&clnt->cl_count)) {
+		if (!refcount_inc_not_zero(&clnt->cl_count)) {
 			seq_release(inode, filp);
 			ret = -EINVAL;
 		}
diff --git a/net/sunrpc/rpc_pipe.c b/net/sunrpc/rpc_pipe.c
index 09c000d490a1..ee5336d73fdd 100644
--- a/net/sunrpc/rpc_pipe.c
+++ b/net/sunrpc/rpc_pipe.c
@@ -423,7 +423,7 @@ rpc_info_open(struct inode *inode, struct file *file)
 		spin_lock(&file->f_path.dentry->d_lock);
 		if (!d_unhashed(file->f_path.dentry))
 			clnt = RPC_I(inode)->private;
-		if (clnt != NULL && atomic_inc_not_zero(&clnt->cl_count)) {
+		if (clnt != NULL && refcount_inc_not_zero(&clnt->cl_count)) {
 			spin_unlock(&file->f_path.dentry->d_lock);
 			m->private = clnt;
 		} else {
-- 
cgit v1.2.3


From c2dc3e5fad13aca5d7bdf4bcb52b1a1d707c8555 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <trond.myklebust@hammerspace.com>
Date: Mon, 26 Jul 2021 07:59:23 -0400
Subject: SUNRPC: Fix potential memory corruption

We really should not call rpc_wake_up_queued_task_set_status() with
xprt->snd_task as an argument unless we are certain that is actually an
rpc_task.

Fixes: 0445f92c5d53 ("SUNRPC: Fix disconnection races")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 include/linux/sunrpc/xprt.h | 1 +
 net/sunrpc/xprt.c           | 6 ++++--
 2 files changed, 5 insertions(+), 2 deletions(-)

(limited to 'net')

diff --git a/include/linux/sunrpc/xprt.h b/include/linux/sunrpc/xprt.h
index c8c39f22d3b1..59cd97da895b 100644
--- a/include/linux/sunrpc/xprt.h
+++ b/include/linux/sunrpc/xprt.h
@@ -432,6 +432,7 @@ void			xprt_release_write(struct rpc_xprt *, struct rpc_task *);
 #define XPRT_CONGESTED		(9)
 #define XPRT_CWND_WAIT		(10)
 #define XPRT_WRITE_SPACE	(11)
+#define XPRT_SND_IS_COOKIE	(12)
 
 static inline void xprt_set_connected(struct rpc_xprt *xprt)
 {
diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c
index fb6db09725c7..bddd354a0076 100644
--- a/net/sunrpc/xprt.c
+++ b/net/sunrpc/xprt.c
@@ -775,9 +775,9 @@ void xprt_force_disconnect(struct rpc_xprt *xprt)
 	/* Try to schedule an autoclose RPC call */
 	if (test_and_set_bit(XPRT_LOCKED, &xprt->state) == 0)
 		queue_work(xprtiod_workqueue, &xprt->task_cleanup);
-	else if (xprt->snd_task)
+	else if (xprt->snd_task && !test_bit(XPRT_SND_IS_COOKIE, &xprt->state))
 		rpc_wake_up_queued_task_set_status(&xprt->pending,
-				xprt->snd_task, -ENOTCONN);
+						   xprt->snd_task, -ENOTCONN);
 	spin_unlock(&xprt->transport_lock);
 }
 EXPORT_SYMBOL_GPL(xprt_force_disconnect);
@@ -866,6 +866,7 @@ bool xprt_lock_connect(struct rpc_xprt *xprt,
 		goto out;
 	if (xprt->snd_task != task)
 		goto out;
+	set_bit(XPRT_SND_IS_COOKIE, &xprt->state);
 	xprt->snd_task = cookie;
 	ret = true;
 out:
@@ -881,6 +882,7 @@ void xprt_unlock_connect(struct rpc_xprt *xprt, void *cookie)
 	if (!test_bit(XPRT_LOCKED, &xprt->state))
 		goto out;
 	xprt->snd_task =NULL;
+	clear_bit(XPRT_SND_IS_COOKIE, &xprt->state);
 	xprt->ops->release_xprt(xprt, NULL);
 	xprt_schedule_autodisconnect(xprt);
 out:
-- 
cgit v1.2.3


From e26d9972720e2484f44cdd94ca4e31cc372ed2ed Mon Sep 17 00:00:00 2001
From: Trond Myklebust <trond.myklebust@hammerspace.com>
Date: Mon, 26 Jul 2021 07:59:24 -0400
Subject: SUNRPC: Clean up scheduling of autoclose

Consolidate duplicated code in xprt_force_disconnect() and
xprt_conditional_disconnect().

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 net/sunrpc/xprt.c | 28 ++++++++++++++++------------
 1 file changed, 16 insertions(+), 12 deletions(-)

(limited to 'net')

diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c
index bddd354a0076..aae5a328b15b 100644
--- a/net/sunrpc/xprt.c
+++ b/net/sunrpc/xprt.c
@@ -760,6 +760,20 @@ void xprt_disconnect_done(struct rpc_xprt *xprt)
 }
 EXPORT_SYMBOL_GPL(xprt_disconnect_done);
 
+/**
+ * xprt_schedule_autoclose_locked - Try to schedule an autoclose RPC call
+ * @xprt: transport to disconnect
+ */
+static void xprt_schedule_autoclose_locked(struct rpc_xprt *xprt)
+{
+	set_bit(XPRT_CLOSE_WAIT, &xprt->state);
+	if (test_and_set_bit(XPRT_LOCKED, &xprt->state) == 0)
+		queue_work(xprtiod_workqueue, &xprt->task_cleanup);
+	else if (xprt->snd_task && !test_bit(XPRT_SND_IS_COOKIE, &xprt->state))
+		rpc_wake_up_queued_task_set_status(&xprt->pending,
+						   xprt->snd_task, -ENOTCONN);
+}
+
 /**
  * xprt_force_disconnect - force a transport to disconnect
  * @xprt: transport to disconnect
@@ -771,13 +785,7 @@ void xprt_force_disconnect(struct rpc_xprt *xprt)
 
 	/* Don't race with the test_bit() in xprt_clear_locked() */
 	spin_lock(&xprt->transport_lock);
-	set_bit(XPRT_CLOSE_WAIT, &xprt->state);
-	/* Try to schedule an autoclose RPC call */
-	if (test_and_set_bit(XPRT_LOCKED, &xprt->state) == 0)
-		queue_work(xprtiod_workqueue, &xprt->task_cleanup);
-	else if (xprt->snd_task && !test_bit(XPRT_SND_IS_COOKIE, &xprt->state))
-		rpc_wake_up_queued_task_set_status(&xprt->pending,
-						   xprt->snd_task, -ENOTCONN);
+	xprt_schedule_autoclose_locked(xprt);
 	spin_unlock(&xprt->transport_lock);
 }
 EXPORT_SYMBOL_GPL(xprt_force_disconnect);
@@ -817,11 +825,7 @@ void xprt_conditional_disconnect(struct rpc_xprt *xprt, unsigned int cookie)
 		goto out;
 	if (test_bit(XPRT_CLOSING, &xprt->state))
 		goto out;
-	set_bit(XPRT_CLOSE_WAIT, &xprt->state);
-	/* Try to schedule an autoclose RPC call */
-	if (test_and_set_bit(XPRT_LOCKED, &xprt->state) == 0)
-		queue_work(xprtiod_workqueue, &xprt->task_cleanup);
-	xprt_wake_pending_tasks(xprt, -EAGAIN);
+	xprt_schedule_autoclose_locked(xprt);
 out:
 	spin_unlock(&xprt->transport_lock);
 }
-- 
cgit v1.2.3


From f99fa50880f5300fbbb3c0754ddc7f8738d24fe7 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <trond.myklebust@hammerspace.com>
Date: Mon, 26 Jul 2021 08:03:12 -0400
Subject: SUNRPC/xprtrdma: Fix reconnection locking

The xprtrdma client code currently relies on the task that initiated the
connect to hold the XPRT_LOCK for the duration of the connection
attempt. If the task is woken early, due to some other event, then that
lock could get released early.
Avoid races by using the same mechanism that the socket code uses of
transferring lock ownership to the RDMA connect worker itself. That
frees us to call rpcrdma_xprt_disconnect() directly since we're now
guaranteed exclusion w.r.t. other callers.

Fixes: 4cf44be6f1e8 ("xprtrdma: Fix recursion into rpcrdma_xprt_disconnect()")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 net/sunrpc/xprt.c               |  2 ++
 net/sunrpc/xprtrdma/transport.c | 11 +++++------
 2 files changed, 7 insertions(+), 6 deletions(-)

(limited to 'net')

diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c
index aae5a328b15b..b88ac8132054 100644
--- a/net/sunrpc/xprt.c
+++ b/net/sunrpc/xprt.c
@@ -877,6 +877,7 @@ out:
 	spin_unlock(&xprt->transport_lock);
 	return ret;
 }
+EXPORT_SYMBOL_GPL(xprt_lock_connect);
 
 void xprt_unlock_connect(struct rpc_xprt *xprt, void *cookie)
 {
@@ -893,6 +894,7 @@ out:
 	spin_unlock(&xprt->transport_lock);
 	wake_up_bit(&xprt->state, XPRT_LOCKED);
 }
+EXPORT_SYMBOL_GPL(xprt_unlock_connect);
 
 /**
  * xprt_connect - schedule a transport connect operation
diff --git a/net/sunrpc/xprtrdma/transport.c b/net/sunrpc/xprtrdma/transport.c
index a463400ed5a3..16e5696314a4 100644
--- a/net/sunrpc/xprtrdma/transport.c
+++ b/net/sunrpc/xprtrdma/transport.c
@@ -250,12 +250,9 @@ xprt_rdma_connect_worker(struct work_struct *work)
 					   xprt->stat.connect_start;
 		xprt_set_connected(xprt);
 		rc = -EAGAIN;
-	} else {
-		/* Force a call to xprt_rdma_close to clean up */
-		spin_lock(&xprt->transport_lock);
-		set_bit(XPRT_CLOSE_WAIT, &xprt->state);
-		spin_unlock(&xprt->transport_lock);
-	}
+	} else
+		rpcrdma_xprt_disconnect(r_xprt);
+	xprt_unlock_connect(xprt, r_xprt);
 	xprt_wake_pending_tasks(xprt, rc);
 }
 
@@ -489,6 +486,8 @@ xprt_rdma_connect(struct rpc_xprt *xprt, struct rpc_task *task)
 	struct rpcrdma_ep *ep = r_xprt->rx_ep;
 	unsigned long delay;
 
+	WARN_ON_ONCE(!xprt_lock_connect(xprt, task, r_xprt));
+
 	delay = 0;
 	if (ep && ep->re_connect_status != 0) {
 		delay = xprt_reconnect_delay(xprt);
-- 
cgit v1.2.3


From 5d46dd04cb68771f77ba66dbf6fd323a4a2ce00d Mon Sep 17 00:00:00 2001
From: Anna Schumaker <Anna.Schumaker@Netapp.com>
Date: Tue, 20 Jul 2021 16:04:42 -0400
Subject: sunrpc: Fix return value of get_srcport()

Since bc1c56e9bbe9 transport->srcport may by unset, causing
get_srcport() to return 0 when called. Fix this by querying the port
from the underlying socket instead of the transport.

Fixes: bc1c56e9bbe9 (SUNRPC: prevent port reuse on transports which don't request it)
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 net/sunrpc/xprtsock.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'net')

diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c
index e573dcecdd66..02b071dbdd22 100644
--- a/net/sunrpc/xprtsock.c
+++ b/net/sunrpc/xprtsock.c
@@ -1656,7 +1656,7 @@ static int xs_get_srcport(struct sock_xprt *transport)
 unsigned short get_srcport(struct rpc_xprt *xprt)
 {
 	struct sock_xprt *sock = container_of(xprt, struct sock_xprt, xprt);
-	return sock->srcport;
+	return xs_sock_getport(sock->sock);
 }
 EXPORT_SYMBOL(get_srcport);
 
-- 
cgit v1.2.3


From e44773daf851dc2755144355723c1c305e7246a1 Mon Sep 17 00:00:00 2001
From: Anna Schumaker <Anna.Schumaker@Netapp.com>
Date: Thu, 29 Jul 2021 16:45:23 -0400
Subject: SUNRPC: Add srcaddr as a file in sysfs

I don't support changing it right now, but it could be useful
information for clients with multiple network cards.

Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 net/sunrpc/sysfs.c | 26 ++++++++++++++++++++++++++
 1 file changed, 26 insertions(+)

(limited to 'net')

diff --git a/net/sunrpc/sysfs.c b/net/sunrpc/sysfs.c
index 64da3bfd28e6..2e7a53504974 100644
--- a/net/sunrpc/sysfs.c
+++ b/net/sunrpc/sysfs.c
@@ -100,6 +100,28 @@ static ssize_t rpc_sysfs_xprt_dstaddr_show(struct kobject *kobj,
 	return ret + 1;
 }
 
+static ssize_t rpc_sysfs_xprt_srcaddr_show(struct kobject *kobj,
+					   struct kobj_attribute *attr,
+					   char *buf)
+{
+	struct rpc_xprt *xprt = rpc_sysfs_xprt_kobj_get_xprt(kobj);
+	struct sockaddr_storage saddr;
+	struct sock_xprt *sock;
+	ssize_t ret = -1;
+
+	if (!xprt)
+		return 0;
+
+	sock = container_of(xprt, struct sock_xprt, xprt);
+	if (kernel_getsockname(sock->sock, (struct sockaddr *)&saddr) < 0)
+		goto out;
+
+	ret = sprintf(buf, "%pISc\n", &saddr);
+out:
+	xprt_put(xprt);
+	return ret + 1;
+}
+
 static ssize_t rpc_sysfs_xprt_info_show(struct kobject *kobj,
 					struct kobj_attribute *attr,
 					char *buf)
@@ -376,6 +398,9 @@ static const void *rpc_sysfs_xprt_namespace(struct kobject *kobj)
 static struct kobj_attribute rpc_sysfs_xprt_dstaddr = __ATTR(dstaddr,
 	0644, rpc_sysfs_xprt_dstaddr_show, rpc_sysfs_xprt_dstaddr_store);
 
+static struct kobj_attribute rpc_sysfs_xprt_srcaddr = __ATTR(srcaddr,
+	0644, rpc_sysfs_xprt_srcaddr_show, NULL);
+
 static struct kobj_attribute rpc_sysfs_xprt_info = __ATTR(xprt_info,
 	0444, rpc_sysfs_xprt_info_show, NULL);
 
@@ -384,6 +409,7 @@ static struct kobj_attribute rpc_sysfs_xprt_change_state = __ATTR(xprt_state,
 
 static struct attribute *rpc_sysfs_xprt_attrs[] = {
 	&rpc_sysfs_xprt_dstaddr.attr,
+	&rpc_sysfs_xprt_srcaddr.attr,
 	&rpc_sysfs_xprt_info.attr,
 	&rpc_sysfs_xprt_change_state.attr,
 	NULL,
-- 
cgit v1.2.3


From 69f2cd6df3ee07ae88befafc038d4dd9154e2799 Mon Sep 17 00:00:00 2001
From: Anna Schumaker <Anna.Schumaker@Netapp.com>
Date: Thu, 29 Jul 2021 16:46:05 -0400
Subject: SUNRPC: Add dst_port to the sysfs xprt info file

This is most likely going to be 2049 for NFS, but some servers might be
configured to export on a non-standard port. Let's show this information
just in case somebody needs it.

Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 net/sunrpc/sysfs.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

(limited to 'net')

diff --git a/net/sunrpc/sysfs.c b/net/sunrpc/sysfs.c
index 2e7a53504974..414c664a3199 100644
--- a/net/sunrpc/sysfs.c
+++ b/net/sunrpc/sysfs.c
@@ -136,14 +136,16 @@ static ssize_t rpc_sysfs_xprt_info_show(struct kobject *kobj,
 		       "max_num_slots=%u\nmin_num_slots=%u\nnum_reqs=%u\n"
 		       "binding_q_len=%u\nsending_q_len=%u\npending_q_len=%u\n"
 		       "backlog_q_len=%u\nmain_xprt=%d\nsrc_port=%u\n"
-		       "tasks_queuelen=%ld\n",
+		       "tasks_queuelen=%ld\ndst_port=%s\n",
 		       xprt->last_used, xprt->cong, xprt->cwnd, xprt->max_reqs,
 		       xprt->min_reqs, xprt->num_reqs, xprt->binding.qlen,
 		       xprt->sending.qlen, xprt->pending.qlen,
 		       xprt->backlog.qlen, xprt->main,
 		       (xprt->xprt_class->ident == XPRT_TRANSPORT_TCP) ?
 		       get_srcport(xprt) : 0,
-		       atomic_long_read(&xprt->queuelen));
+		       atomic_long_read(&xprt->queuelen),
+		       (xprt->xprt_class->ident == XPRT_TRANSPORT_TCP) ?
+				xprt->address_strings[RPC_DISPLAY_PORT] : "0");
 	xprt_put(xprt);
 	return ret + 1;
 }
-- 
cgit v1.2.3


From 438623a06bacd69c40c4af633bb09a3bbb9dfc78 Mon Sep 17 00:00:00 2001
From: Chuck Lever <chuck.lever@oracle.com>
Date: Thu, 15 Jul 2021 15:52:06 -0400
Subject: SUNRPC: Add svc_rqst::rq_auth_stat

I'd like to take commit 4532608d71c8 ("SUNRPC: Clean up generic
dispatcher code") even further by using only private local SVC
dispatchers for all kernel RPC services. This change would enable
the removal of the logic that switches between
svc_generic_dispatch() and a service's private dispatcher, and
simplify the invocation of the service's pc_release method
so that humans can visually verify that it is always invoked
properly.

All that will come later.

First, let's provide a better way to return authentication errors
from SVC dispatcher functions. Instead of overloading the dispatch
method's *statp argument, add a field to struct svc_rqst that can
hold an error value.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 include/linux/sunrpc/svc.h        |  1 +
 include/linux/sunrpc/svcauth.h    |  4 ++--
 include/trace/events/sunrpc.h     |  6 +++---
 net/sunrpc/auth_gss/svcauth_gss.c | 43 +++++++++++++++++++--------------------
 net/sunrpc/svc.c                  | 17 ++++++++--------
 net/sunrpc/svcauth.c              |  8 ++++----
 net/sunrpc/svcauth_unix.c         | 12 +++++------
 7 files changed, 46 insertions(+), 45 deletions(-)

(limited to 'net')

diff --git a/include/linux/sunrpc/svc.h b/include/linux/sunrpc/svc.h
index e91d51ea028b..35f12963e1ff 100644
--- a/include/linux/sunrpc/svc.h
+++ b/include/linux/sunrpc/svc.h
@@ -282,6 +282,7 @@ struct svc_rqst {
 	void *			rq_argp;	/* decoded arguments */
 	void *			rq_resp;	/* xdr'd results */
 	void *			rq_auth_data;	/* flavor-specific data */
+	__be32			rq_auth_stat;	/* authentication status */
 	int			rq_auth_slack;	/* extra space xdr code
 						 * should leave in head
 						 * for krb5i, krb5p.
diff --git a/include/linux/sunrpc/svcauth.h b/include/linux/sunrpc/svcauth.h
index b0003866a249..6d9cc9080aca 100644
--- a/include/linux/sunrpc/svcauth.h
+++ b/include/linux/sunrpc/svcauth.h
@@ -127,7 +127,7 @@ struct auth_ops {
 	char *	name;
 	struct module *owner;
 	int	flavour;
-	int	(*accept)(struct svc_rqst *rq, __be32 *authp);
+	int	(*accept)(struct svc_rqst *rq);
 	int	(*release)(struct svc_rqst *rq);
 	void	(*domain_release)(struct auth_domain *);
 	int	(*set_client)(struct svc_rqst *rq);
@@ -149,7 +149,7 @@ struct auth_ops {
 
 struct svc_xprt;
 
-extern int	svc_authenticate(struct svc_rqst *rqstp, __be32 *authp);
+extern int	svc_authenticate(struct svc_rqst *rqstp);
 extern int	svc_authorise(struct svc_rqst *rqstp);
 extern int	svc_set_client(struct svc_rqst *rqstp);
 extern int	svc_auth_register(rpc_authflavor_t flavor, struct auth_ops *aops);
diff --git a/include/trace/events/sunrpc.h b/include/trace/events/sunrpc.h
index 18d552a17c19..c7d9e6c7a979 100644
--- a/include/trace/events/sunrpc.h
+++ b/include/trace/events/sunrpc.h
@@ -1582,9 +1582,9 @@ TRACE_DEFINE_ENUM(SVC_COMPLETE);
 		{ SVC_COMPLETE,	"SVC_COMPLETE" })
 
 TRACE_EVENT(svc_authenticate,
-	TP_PROTO(const struct svc_rqst *rqst, int auth_res, __be32 auth_stat),
+	TP_PROTO(const struct svc_rqst *rqst, int auth_res),
 
-	TP_ARGS(rqst, auth_res, auth_stat),
+	TP_ARGS(rqst, auth_res),
 
 	TP_STRUCT__entry(
 		__field(u32, xid)
@@ -1595,7 +1595,7 @@ TRACE_EVENT(svc_authenticate,
 	TP_fast_assign(
 		__entry->xid = be32_to_cpu(rqst->rq_xid);
 		__entry->svc_status = auth_res;
-		__entry->auth_stat = be32_to_cpu(auth_stat);
+		__entry->auth_stat = be32_to_cpu(rqst->rq_auth_stat);
 	),
 
 	TP_printk("xid=0x%08x auth_res=%s auth_stat=%s",
diff --git a/net/sunrpc/auth_gss/svcauth_gss.c b/net/sunrpc/auth_gss/svcauth_gss.c
index a81be45f40d9..635449ed7af6 100644
--- a/net/sunrpc/auth_gss/svcauth_gss.c
+++ b/net/sunrpc/auth_gss/svcauth_gss.c
@@ -707,11 +707,11 @@ svc_safe_putnetobj(struct kvec *resv, struct xdr_netobj *o)
 /*
  * Verify the checksum on the header and return SVC_OK on success.
  * Otherwise, return SVC_DROP (in the case of a bad sequence number)
- * or return SVC_DENIED and indicate error in authp.
+ * or return SVC_DENIED and indicate error in rqstp->rq_auth_stat.
  */
 static int
 gss_verify_header(struct svc_rqst *rqstp, struct rsc *rsci,
-		  __be32 *rpcstart, struct rpc_gss_wire_cred *gc, __be32 *authp)
+		  __be32 *rpcstart, struct rpc_gss_wire_cred *gc)
 {
 	struct gss_ctx		*ctx_id = rsci->mechctx;
 	struct xdr_buf		rpchdr;
@@ -725,7 +725,7 @@ gss_verify_header(struct svc_rqst *rqstp, struct rsc *rsci,
 	iov.iov_len = (u8 *)argv->iov_base - (u8 *)rpcstart;
 	xdr_buf_from_iov(&iov, &rpchdr);
 
-	*authp = rpc_autherr_badverf;
+	rqstp->rq_auth_stat = rpc_autherr_badverf;
 	if (argv->iov_len < 4)
 		return SVC_DENIED;
 	flavor = svc_getnl(argv);
@@ -737,13 +737,13 @@ gss_verify_header(struct svc_rqst *rqstp, struct rsc *rsci,
 	if (rqstp->rq_deferred) /* skip verification of revisited request */
 		return SVC_OK;
 	if (gss_verify_mic(ctx_id, &rpchdr, &checksum) != GSS_S_COMPLETE) {
-		*authp = rpcsec_gsserr_credproblem;
+		rqstp->rq_auth_stat = rpcsec_gsserr_credproblem;
 		return SVC_DENIED;
 	}
 
 	if (gc->gc_seq > MAXSEQ) {
 		trace_rpcgss_svc_seqno_large(rqstp, gc->gc_seq);
-		*authp = rpcsec_gsserr_ctxproblem;
+		rqstp->rq_auth_stat = rpcsec_gsserr_ctxproblem;
 		return SVC_DENIED;
 	}
 	if (!gss_check_seq_num(rqstp, rsci, gc->gc_seq))
@@ -1142,7 +1142,7 @@ static void gss_free_in_token_pages(struct gssp_in_token *in_token)
 }
 
 static int gss_read_proxy_verf(struct svc_rqst *rqstp,
-			       struct rpc_gss_wire_cred *gc, __be32 *authp,
+			       struct rpc_gss_wire_cred *gc,
 			       struct xdr_netobj *in_handle,
 			       struct gssp_in_token *in_token)
 {
@@ -1151,7 +1151,7 @@ static int gss_read_proxy_verf(struct svc_rqst *rqstp,
 	int pages, i, res, pgto, pgfrom;
 	size_t inlen, to_offs, from_offs;
 
-	res = gss_read_common_verf(gc, argv, authp, in_handle);
+	res = gss_read_common_verf(gc, argv, &rqstp->rq_auth_stat, in_handle);
 	if (res)
 		return res;
 
@@ -1227,7 +1227,7 @@ gss_write_resv(struct kvec *resv, size_t size_limit,
  * Otherwise, drop the request pending an answer to the upcall.
  */
 static int svcauth_gss_legacy_init(struct svc_rqst *rqstp,
-			struct rpc_gss_wire_cred *gc, __be32 *authp)
+				   struct rpc_gss_wire_cred *gc)
 {
 	struct kvec *argv = &rqstp->rq_arg.head[0];
 	struct kvec *resv = &rqstp->rq_res.head[0];
@@ -1236,7 +1236,7 @@ static int svcauth_gss_legacy_init(struct svc_rqst *rqstp,
 	struct sunrpc_net *sn = net_generic(SVC_NET(rqstp), sunrpc_net_id);
 
 	memset(&rsikey, 0, sizeof(rsikey));
-	ret = gss_read_verf(gc, argv, authp,
+	ret = gss_read_verf(gc, argv, &rqstp->rq_auth_stat,
 			    &rsikey.in_handle, &rsikey.in_token);
 	if (ret)
 		return ret;
@@ -1339,7 +1339,7 @@ out:
 }
 
 static int svcauth_gss_proxy_init(struct svc_rqst *rqstp,
-			struct rpc_gss_wire_cred *gc, __be32 *authp)
+				  struct rpc_gss_wire_cred *gc)
 {
 	struct kvec *resv = &rqstp->rq_res.head[0];
 	struct xdr_netobj cli_handle;
@@ -1351,8 +1351,7 @@ static int svcauth_gss_proxy_init(struct svc_rqst *rqstp,
 	struct sunrpc_net *sn = net_generic(net, sunrpc_net_id);
 
 	memset(&ud, 0, sizeof(ud));
-	ret = gss_read_proxy_verf(rqstp, gc, authp,
-				  &ud.in_handle, &ud.in_token);
+	ret = gss_read_proxy_verf(rqstp, gc, &ud.in_handle, &ud.in_token);
 	if (ret)
 		return ret;
 
@@ -1525,7 +1524,7 @@ static void destroy_use_gss_proxy_proc_entry(struct net *net) {}
  * response here and return SVC_COMPLETE.
  */
 static int
-svcauth_gss_accept(struct svc_rqst *rqstp, __be32 *authp)
+svcauth_gss_accept(struct svc_rqst *rqstp)
 {
 	struct kvec	*argv = &rqstp->rq_arg.head[0];
 	struct kvec	*resv = &rqstp->rq_res.head[0];
@@ -1538,7 +1537,7 @@ svcauth_gss_accept(struct svc_rqst *rqstp, __be32 *authp)
 	int		ret;
 	struct sunrpc_net *sn = net_generic(SVC_NET(rqstp), sunrpc_net_id);
 
-	*authp = rpc_autherr_badcred;
+	rqstp->rq_auth_stat = rpc_autherr_badcred;
 	if (!svcdata)
 		svcdata = kmalloc(sizeof(*svcdata), GFP_KERNEL);
 	if (!svcdata)
@@ -1575,22 +1574,22 @@ svcauth_gss_accept(struct svc_rqst *rqstp, __be32 *authp)
 	if ((gc->gc_proc != RPC_GSS_PROC_DATA) && (rqstp->rq_proc != 0))
 		goto auth_err;
 
-	*authp = rpc_autherr_badverf;
+	rqstp->rq_auth_stat = rpc_autherr_badverf;
 	switch (gc->gc_proc) {
 	case RPC_GSS_PROC_INIT:
 	case RPC_GSS_PROC_CONTINUE_INIT:
 		if (use_gss_proxy(SVC_NET(rqstp)))
-			return svcauth_gss_proxy_init(rqstp, gc, authp);
+			return svcauth_gss_proxy_init(rqstp, gc);
 		else
-			return svcauth_gss_legacy_init(rqstp, gc, authp);
+			return svcauth_gss_legacy_init(rqstp, gc);
 	case RPC_GSS_PROC_DATA:
 	case RPC_GSS_PROC_DESTROY:
 		/* Look up the context, and check the verifier: */
-		*authp = rpcsec_gsserr_credproblem;
+		rqstp->rq_auth_stat = rpcsec_gsserr_credproblem;
 		rsci = gss_svc_searchbyctx(sn->rsc_cache, &gc->gc_ctx);
 		if (!rsci)
 			goto auth_err;
-		switch (gss_verify_header(rqstp, rsci, rpcstart, gc, authp)) {
+		switch (gss_verify_header(rqstp, rsci, rpcstart, gc)) {
 		case SVC_OK:
 			break;
 		case SVC_DENIED:
@@ -1600,7 +1599,7 @@ svcauth_gss_accept(struct svc_rqst *rqstp, __be32 *authp)
 		}
 		break;
 	default:
-		*authp = rpc_autherr_rejectedcred;
+		rqstp->rq_auth_stat = rpc_autherr_rejectedcred;
 		goto auth_err;
 	}
 
@@ -1616,13 +1615,13 @@ svcauth_gss_accept(struct svc_rqst *rqstp, __be32 *authp)
 		svc_putnl(resv, RPC_SUCCESS);
 		goto complete;
 	case RPC_GSS_PROC_DATA:
-		*authp = rpcsec_gsserr_ctxproblem;
+		rqstp->rq_auth_stat = rpcsec_gsserr_ctxproblem;
 		svcdata->verf_start = resv->iov_base + resv->iov_len;
 		if (gss_write_verf(rqstp, rsci->mechctx, gc->gc_seq))
 			goto auth_err;
 		rqstp->rq_cred = rsci->cred;
 		get_group_info(rsci->cred.cr_group_info);
-		*authp = rpc_autherr_badcred;
+		rqstp->rq_auth_stat = rpc_autherr_badcred;
 		switch (gc->gc_svc) {
 		case RPC_GSS_SVC_NONE:
 			break;
diff --git a/net/sunrpc/svc.c b/net/sunrpc/svc.c
index 0de918cb3d90..360dab62b6b4 100644
--- a/net/sunrpc/svc.c
+++ b/net/sunrpc/svc.c
@@ -1283,7 +1283,7 @@ svc_process_common(struct svc_rqst *rqstp, struct kvec *argv, struct kvec *resv)
 	struct svc_process_info process;
 	__be32			*statp;
 	u32			prog, vers;
-	__be32			auth_stat, rpc_stat;
+	__be32			rpc_stat;
 	int			auth_res;
 	__be32			*reply_statp;
 
@@ -1326,14 +1326,14 @@ svc_process_common(struct svc_rqst *rqstp, struct kvec *argv, struct kvec *resv)
 	 * We do this before anything else in order to get a decent
 	 * auth verifier.
 	 */
-	auth_res = svc_authenticate(rqstp, &auth_stat);
+	auth_res = svc_authenticate(rqstp);
 	/* Also give the program a chance to reject this call: */
 	if (auth_res == SVC_OK && progp) {
-		auth_stat = rpc_autherr_badcred;
+		rqstp->rq_auth_stat = rpc_autherr_badcred;
 		auth_res = progp->pg_authenticate(rqstp);
 	}
 	if (auth_res != SVC_OK)
-		trace_svc_authenticate(rqstp, auth_res, auth_stat);
+		trace_svc_authenticate(rqstp, auth_res);
 	switch (auth_res) {
 	case SVC_OK:
 		break;
@@ -1392,8 +1392,8 @@ svc_process_common(struct svc_rqst *rqstp, struct kvec *argv, struct kvec *resv)
 			goto release_dropit;
 		if (*statp == rpc_garbage_args)
 			goto err_garbage;
-		auth_stat = svc_get_autherr(rqstp, statp);
-		if (auth_stat != rpc_auth_ok)
+		rqstp->rq_auth_stat = svc_get_autherr(rqstp, statp);
+		if (rqstp->rq_auth_stat != rpc_auth_ok)
 			goto err_release_bad_auth;
 	} else {
 		dprintk("svc: calling dispatcher\n");
@@ -1450,13 +1450,14 @@ err_release_bad_auth:
 	if (procp->pc_release)
 		procp->pc_release(rqstp);
 err_bad_auth:
-	dprintk("svc: authentication failed (%d)\n", ntohl(auth_stat));
+	dprintk("svc: authentication failed (%d)\n",
+		be32_to_cpu(rqstp->rq_auth_stat));
 	serv->sv_stats->rpcbadauth++;
 	/* Restore write pointer to location of accept status: */
 	xdr_ressize_check(rqstp, reply_statp);
 	svc_putnl(resv, 1);	/* REJECT */
 	svc_putnl(resv, 1);	/* AUTH_ERROR */
-	svc_putnl(resv, ntohl(auth_stat));	/* status */
+	svc_putu32(resv, rqstp->rq_auth_stat);	/* status */
 	goto sendit;
 
 err_bad_prog:
diff --git a/net/sunrpc/svcauth.c b/net/sunrpc/svcauth.c
index 998b196b6176..5a8b8e03fdd4 100644
--- a/net/sunrpc/svcauth.c
+++ b/net/sunrpc/svcauth.c
@@ -59,12 +59,12 @@ svc_put_auth_ops(struct auth_ops *aops)
 }
 
 int
-svc_authenticate(struct svc_rqst *rqstp, __be32 *authp)
+svc_authenticate(struct svc_rqst *rqstp)
 {
 	rpc_authflavor_t	flavor;
 	struct auth_ops		*aops;
 
-	*authp = rpc_auth_ok;
+	rqstp->rq_auth_stat = rpc_auth_ok;
 
 	flavor = svc_getnl(&rqstp->rq_arg.head[0]);
 
@@ -72,7 +72,7 @@ svc_authenticate(struct svc_rqst *rqstp, __be32 *authp)
 
 	aops = svc_get_auth_ops(flavor);
 	if (aops == NULL) {
-		*authp = rpc_autherr_badcred;
+		rqstp->rq_auth_stat = rpc_autherr_badcred;
 		return SVC_DENIED;
 	}
 
@@ -80,7 +80,7 @@ svc_authenticate(struct svc_rqst *rqstp, __be32 *authp)
 	init_svc_cred(&rqstp->rq_cred);
 
 	rqstp->rq_authop = aops;
-	return aops->accept(rqstp, authp);
+	return aops->accept(rqstp);
 }
 EXPORT_SYMBOL_GPL(svc_authenticate);
 
diff --git a/net/sunrpc/svcauth_unix.c b/net/sunrpc/svcauth_unix.c
index 35b7966ac3b3..eacfebf326dd 100644
--- a/net/sunrpc/svcauth_unix.c
+++ b/net/sunrpc/svcauth_unix.c
@@ -725,7 +725,7 @@ svcauth_unix_set_client(struct svc_rqst *rqstp)
 EXPORT_SYMBOL_GPL(svcauth_unix_set_client);
 
 static int
-svcauth_null_accept(struct svc_rqst *rqstp, __be32 *authp)
+svcauth_null_accept(struct svc_rqst *rqstp)
 {
 	struct kvec	*argv = &rqstp->rq_arg.head[0];
 	struct kvec	*resv = &rqstp->rq_res.head[0];
@@ -736,12 +736,12 @@ svcauth_null_accept(struct svc_rqst *rqstp, __be32 *authp)
 
 	if (svc_getu32(argv) != 0) {
 		dprintk("svc: bad null cred\n");
-		*authp = rpc_autherr_badcred;
+		rqstp->rq_auth_stat = rpc_autherr_badcred;
 		return SVC_DENIED;
 	}
 	if (svc_getu32(argv) != htonl(RPC_AUTH_NULL) || svc_getu32(argv) != 0) {
 		dprintk("svc: bad null verf\n");
-		*authp = rpc_autherr_badverf;
+		rqstp->rq_auth_stat = rpc_autherr_badverf;
 		return SVC_DENIED;
 	}
 
@@ -785,7 +785,7 @@ struct auth_ops svcauth_null = {
 
 
 static int
-svcauth_unix_accept(struct svc_rqst *rqstp, __be32 *authp)
+svcauth_unix_accept(struct svc_rqst *rqstp)
 {
 	struct kvec	*argv = &rqstp->rq_arg.head[0];
 	struct kvec	*resv = &rqstp->rq_res.head[0];
@@ -827,7 +827,7 @@ svcauth_unix_accept(struct svc_rqst *rqstp, __be32 *authp)
 	}
 	groups_sort(cred->cr_group_info);
 	if (svc_getu32(argv) != htonl(RPC_AUTH_NULL) || svc_getu32(argv) != 0) {
-		*authp = rpc_autherr_badverf;
+		rqstp->rq_auth_stat = rpc_autherr_badverf;
 		return SVC_DENIED;
 	}
 
@@ -839,7 +839,7 @@ svcauth_unix_accept(struct svc_rqst *rqstp, __be32 *authp)
 	return SVC_OK;
 
 badcred:
-	*authp = rpc_autherr_badcred;
+	rqstp->rq_auth_stat = rpc_autherr_badcred;
 	return SVC_DENIED;
 }
 
-- 
cgit v1.2.3


From 5c2465dfd457f3015eebcc3ace50570e1d896aeb Mon Sep 17 00:00:00 2001
From: Chuck Lever <chuck.lever@oracle.com>
Date: Thu, 15 Jul 2021 15:52:12 -0400
Subject: SUNRPC: Set rq_auth_stat in the pg_authenticate() callout

In a few moments, rq_auth_stat will need to be explicitly set to
rpc_auth_ok before execution gets to the dispatcher.

svc_authenticate() already sets it, but it often gets reset to
rpc_autherr_badcred right after that call, even when authentication
is successful. Let's ensure that the pg_authenticate callout and
svc_set_client() set it properly in every case.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 fs/lockd/svc.c                    | 2 ++
 fs/nfs/callback.c                 | 4 ++++
 net/sunrpc/auth_gss/svcauth_gss.c | 4 ++++
 net/sunrpc/svc.c                  | 4 +---
 net/sunrpc/svcauth_unix.c         | 6 +++++-
 5 files changed, 16 insertions(+), 4 deletions(-)

(limited to 'net')

diff --git a/fs/lockd/svc.c b/fs/lockd/svc.c
index 2de048f80eb8..8e936999216c 100644
--- a/fs/lockd/svc.c
+++ b/fs/lockd/svc.c
@@ -649,6 +649,7 @@ static int lockd_authenticate(struct svc_rqst *rqstp)
 	switch (rqstp->rq_authop->flavour) {
 		case RPC_AUTH_NULL:
 		case RPC_AUTH_UNIX:
+			rqstp->rq_auth_stat = rpc_auth_ok;
 			if (rqstp->rq_proc == 0)
 				return SVC_OK;
 			if (is_callback(rqstp->rq_proc)) {
@@ -659,6 +660,7 @@ static int lockd_authenticate(struct svc_rqst *rqstp)
 			}
 			return svc_set_client(rqstp);
 	}
+	rqstp->rq_auth_stat = rpc_autherr_badcred;
 	return SVC_DENIED;
 }
 
diff --git a/fs/nfs/callback.c b/fs/nfs/callback.c
index 7817ad94a6ba..86d856de1389 100644
--- a/fs/nfs/callback.c
+++ b/fs/nfs/callback.c
@@ -429,6 +429,8 @@ check_gss_callback_principal(struct nfs_client *clp, struct svc_rqst *rqstp)
  */
 static int nfs_callback_authenticate(struct svc_rqst *rqstp)
 {
+	rqstp->rq_auth_stat = rpc_autherr_badcred;
+
 	switch (rqstp->rq_authop->flavour) {
 	case RPC_AUTH_NULL:
 		if (rqstp->rq_proc != CB_NULL)
@@ -439,6 +441,8 @@ static int nfs_callback_authenticate(struct svc_rqst *rqstp)
 		 if (svc_is_backchannel(rqstp))
 			return SVC_DENIED;
 	}
+
+	rqstp->rq_auth_stat = rpc_auth_ok;
 	return SVC_OK;
 }
 
diff --git a/net/sunrpc/auth_gss/svcauth_gss.c b/net/sunrpc/auth_gss/svcauth_gss.c
index 635449ed7af6..f89075070fb0 100644
--- a/net/sunrpc/auth_gss/svcauth_gss.c
+++ b/net/sunrpc/auth_gss/svcauth_gss.c
@@ -1038,6 +1038,8 @@ svcauth_gss_set_client(struct svc_rqst *rqstp)
 	struct rpc_gss_wire_cred *gc = &svcdata->clcred;
 	int stat;
 
+	rqstp->rq_auth_stat = rpc_autherr_badcred;
+
 	/*
 	 * A gss export can be specified either by:
 	 * 	export	*(sec=krb5,rw)
@@ -1053,6 +1055,8 @@ svcauth_gss_set_client(struct svc_rqst *rqstp)
 	stat = svcauth_unix_set_client(rqstp);
 	if (stat == SVC_DROP || stat == SVC_CLOSE)
 		return stat;
+
+	rqstp->rq_auth_stat = rpc_auth_ok;
 	return SVC_OK;
 }
 
diff --git a/net/sunrpc/svc.c b/net/sunrpc/svc.c
index 360dab62b6b4..2019d1203641 100644
--- a/net/sunrpc/svc.c
+++ b/net/sunrpc/svc.c
@@ -1328,10 +1328,8 @@ svc_process_common(struct svc_rqst *rqstp, struct kvec *argv, struct kvec *resv)
 	 */
 	auth_res = svc_authenticate(rqstp);
 	/* Also give the program a chance to reject this call: */
-	if (auth_res == SVC_OK && progp) {
-		rqstp->rq_auth_stat = rpc_autherr_badcred;
+	if (auth_res == SVC_OK && progp)
 		auth_res = progp->pg_authenticate(rqstp);
-	}
 	if (auth_res != SVC_OK)
 		trace_svc_authenticate(rqstp, auth_res);
 	switch (auth_res) {
diff --git a/net/sunrpc/svcauth_unix.c b/net/sunrpc/svcauth_unix.c
index eacfebf326dd..d7ed7d49115a 100644
--- a/net/sunrpc/svcauth_unix.c
+++ b/net/sunrpc/svcauth_unix.c
@@ -681,8 +681,9 @@ svcauth_unix_set_client(struct svc_rqst *rqstp)
 
 	rqstp->rq_client = NULL;
 	if (rqstp->rq_proc == 0)
-		return SVC_OK;
+		goto out;
 
+	rqstp->rq_auth_stat = rpc_autherr_badcred;
 	ipm = ip_map_cached_get(xprt);
 	if (ipm == NULL)
 		ipm = __ip_map_lookup(sn->ip_map_cache, rqstp->rq_server->sv_program->pg_class,
@@ -719,6 +720,9 @@ svcauth_unix_set_client(struct svc_rqst *rqstp)
 		put_group_info(cred->cr_group_info);
 		cred->cr_group_info = gi;
 	}
+
+out:
+	rqstp->rq_auth_stat = rpc_auth_ok;
 	return SVC_OK;
 }
 
-- 
cgit v1.2.3


From 9082e1d914f8b27114352b1940bbcc7522f682e7 Mon Sep 17 00:00:00 2001
From: Chuck Lever <chuck.lever@oracle.com>
Date: Thu, 15 Jul 2021 15:52:19 -0400
Subject: SUNRPC: Eliminate the RQ_AUTHERR flag

Now that there is an alternate method for returning an auth_stat
value, replace the RQ_AUTHERR flag with use of that new method.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 fs/nfs/callback_xdr.c         |  3 ++-
 include/linux/sunrpc/svc.h    |  2 --
 include/trace/events/sunrpc.h |  3 +--
 net/sunrpc/svc.c              | 24 ++++--------------------
 4 files changed, 7 insertions(+), 25 deletions(-)

(limited to 'net')

diff --git a/fs/nfs/callback_xdr.c b/fs/nfs/callback_xdr.c
index c5348ba81129..7ff99155b023 100644
--- a/fs/nfs/callback_xdr.c
+++ b/fs/nfs/callback_xdr.c
@@ -988,7 +988,8 @@ static __be32 nfs4_callback_compound(struct svc_rqst *rqstp)
 
 out_invalidcred:
 	pr_warn_ratelimited("NFS: NFSv4 callback contains invalid cred\n");
-	return svc_return_autherr(rqstp, rpc_autherr_badcred);
+	rqstp->rq_auth_stat = rpc_autherr_badcred;
+	return rpc_success;
 }
 
 /*
diff --git a/include/linux/sunrpc/svc.h b/include/linux/sunrpc/svc.h
index 35f12963e1ff..63c9210cae06 100644
--- a/include/linux/sunrpc/svc.h
+++ b/include/linux/sunrpc/svc.h
@@ -275,7 +275,6 @@ struct svc_rqst {
 #define	RQ_VICTIM	(5)			/* about to be shut down */
 #define	RQ_BUSY		(6)			/* request is busy */
 #define	RQ_DATA		(7)			/* request has data */
-#define RQ_AUTHERR	(8)			/* Request status is auth error */
 	unsigned long		rq_flags;	/* flags field */
 	ktime_t			rq_qtime;	/* enqueue time */
 
@@ -533,7 +532,6 @@ unsigned int	   svc_fill_write_vector(struct svc_rqst *rqstp,
 char		  *svc_fill_symlink_pathname(struct svc_rqst *rqstp,
 					     struct kvec *first, void *p,
 					     size_t total);
-__be32		   svc_return_autherr(struct svc_rqst *rqstp, __be32 auth_err);
 __be32		   svc_generic_init_request(struct svc_rqst *rqstp,
 					    const struct svc_program *progp,
 					    struct svc_process_info *procinfo);
diff --git a/include/trace/events/sunrpc.h b/include/trace/events/sunrpc.h
index c7d9e6c7a979..169b93e4dbc1 100644
--- a/include/trace/events/sunrpc.h
+++ b/include/trace/events/sunrpc.h
@@ -1539,8 +1539,7 @@ DEFINE_SVCXDRBUF_EVENT(sendto);
 	svc_rqst_flag(SPLICE_OK)					\
 	svc_rqst_flag(VICTIM)						\
 	svc_rqst_flag(BUSY)						\
-	svc_rqst_flag(DATA)						\
-	svc_rqst_flag_end(AUTHERR)
+	svc_rqst_flag_end(DATA)
 
 #undef svc_rqst_flag
 #undef svc_rqst_flag_end
diff --git a/net/sunrpc/svc.c b/net/sunrpc/svc.c
index 2019d1203641..95836bf514b5 100644
--- a/net/sunrpc/svc.c
+++ b/net/sunrpc/svc.c
@@ -1163,22 +1163,6 @@ void svc_printk(struct svc_rqst *rqstp, const char *fmt, ...)
 static __printf(2,3) void svc_printk(struct svc_rqst *rqstp, const char *fmt, ...) {}
 #endif
 
-__be32
-svc_return_autherr(struct svc_rqst *rqstp, __be32 auth_err)
-{
-	set_bit(RQ_AUTHERR, &rqstp->rq_flags);
-	return auth_err;
-}
-EXPORT_SYMBOL_GPL(svc_return_autherr);
-
-static __be32
-svc_get_autherr(struct svc_rqst *rqstp, __be32 *statp)
-{
-	if (test_and_clear_bit(RQ_AUTHERR, &rqstp->rq_flags))
-		return *statp;
-	return rpc_auth_ok;
-}
-
 static int
 svc_generic_dispatch(struct svc_rqst *rqstp, __be32 *statp)
 {
@@ -1202,7 +1186,7 @@ svc_generic_dispatch(struct svc_rqst *rqstp, __be32 *statp)
 	    test_bit(RQ_DROPME, &rqstp->rq_flags))
 		return 0;
 
-	if (test_bit(RQ_AUTHERR, &rqstp->rq_flags))
+	if (rqstp->rq_auth_stat != rpc_auth_ok)
 		return 1;
 
 	if (*statp != rpc_success)
@@ -1390,15 +1374,15 @@ svc_process_common(struct svc_rqst *rqstp, struct kvec *argv, struct kvec *resv)
 			goto release_dropit;
 		if (*statp == rpc_garbage_args)
 			goto err_garbage;
-		rqstp->rq_auth_stat = svc_get_autherr(rqstp, statp);
-		if (rqstp->rq_auth_stat != rpc_auth_ok)
-			goto err_release_bad_auth;
 	} else {
 		dprintk("svc: calling dispatcher\n");
 		if (!process.dispatch(rqstp, statp))
 			goto release_dropit; /* Release reply info */
 	}
 
+	if (rqstp->rq_auth_stat != rpc_auth_ok)
+		goto err_release_bad_auth;
+
 	/* Check RPC status result */
 	if (*statp != rpc_success)
 		resv->iov_len = ((void*)statp)  - resv->iov_base + 4;
-- 
cgit v1.2.3


From 0a6ff58edbfb26469a095ab964095506352fc960 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <trond.myklebust@hammerspace.com>
Date: Tue, 24 Aug 2021 11:38:17 -0400
Subject: SUNRPC: Simplify socket shutdown when not reusing TCP ports

If we're not required to reuse the TCP port, then we can just
immediately close the socket, and leave the cleanup details to the TCP
layer.

Fixes: e6237b6feb37 ("NFSv4.1: Don't rebind to the same source port when reconnecting to the server")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 net/sunrpc/xprtsock.c | 4 ++++
 1 file changed, 4 insertions(+)

(limited to 'net')

diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c
index 02b071dbdd22..5fb969f8a5ad 100644
--- a/net/sunrpc/xprtsock.c
+++ b/net/sunrpc/xprtsock.c
@@ -2099,6 +2099,10 @@ static void xs_tcp_shutdown(struct rpc_xprt *xprt)
 
 	if (sock == NULL)
 		return;
+	if (!xprt->reuseport) {
+		xs_close(xprt);
+		return;
+	}
 	switch (skst) {
 	default:
 		kernel_sock_shutdown(sock, SHUT_RDWR);
-- 
cgit v1.2.3


From 7c81e6a9d75bd2c094005b64b442cec729dbdf66 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <trond.myklebust@hammerspace.com>
Date: Tue, 24 Aug 2021 11:38:18 -0400
Subject: SUNRPC: Tweak TCP socket shutdown in the RPC client

We only really need to call shutdown() if we're in the ESTABLISHED TCP
state, since that is the only case where the client is initiating a
close of an established connection.

If the socket is in FIN_WAIT1 or FIN_WAIT2, then we've already initiated
socket shutdown and are waiting for the server's reply, so do nothing.

In all other cases where we've already received a FIN from the server,
we should be able to just close the socket.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 net/sunrpc/xprtsock.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

(limited to 'net')

diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c
index 5fb969f8a5ad..d80793a1b75c 100644
--- a/net/sunrpc/xprtsock.c
+++ b/net/sunrpc/xprtsock.c
@@ -2104,12 +2104,15 @@ static void xs_tcp_shutdown(struct rpc_xprt *xprt)
 		return;
 	}
 	switch (skst) {
-	default:
+	case TCP_FIN_WAIT1:
+	case TCP_FIN_WAIT2:
+		break;
+	case TCP_ESTABLISHED:
+	case TCP_CLOSE_WAIT:
 		kernel_sock_shutdown(sock, SHUT_RDWR);
 		trace_rpc_socket_shutdown(xprt, sock);
 		break;
-	case TCP_CLOSE:
-	case TCP_TIME_WAIT:
+	default:
 		xs_reset_transport(transport);
 	}
 }
-- 
cgit v1.2.3


From 3a3f976639f267823e443fdd8bffa03848fa1c3f Mon Sep 17 00:00:00 2001
From: Olga Kornievskaia <kolga@netapp.com>
Date: Fri, 27 Aug 2021 14:37:14 -0400
Subject: SUNRPC keep track of number of transports to unique addresses

Currently, xprt_switch keeps a number of all xprts (xps_nxprts)
that were added to the switch regardless of whethere it's an
nconnect transport or a transport to a trunkable address.
Introduce a new counter to keep track of transports to unique
destination addresses per xprt_switch.

Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 include/linux/sunrpc/xprtmultipath.h | 1 +
 net/sunrpc/clnt.c                    | 2 +-
 net/sunrpc/xprtmultipath.c           | 1 +
 3 files changed, 3 insertions(+), 1 deletion(-)

(limited to 'net')

diff --git a/include/linux/sunrpc/xprtmultipath.h b/include/linux/sunrpc/xprtmultipath.h
index b19addc8b715..bbb8a5fa0816 100644
--- a/include/linux/sunrpc/xprtmultipath.h
+++ b/include/linux/sunrpc/xprtmultipath.h
@@ -18,6 +18,7 @@ struct rpc_xprt_switch {
 	unsigned int		xps_id;
 	unsigned int		xps_nxprts;
 	unsigned int		xps_nactive;
+	unsigned int		xps_nunique_destaddr_xprts;
 	atomic_long_t		xps_queuelen;
 	struct list_head	xps_xprt_list;
 
diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index a5b7f6e34d15..451ac7d031db 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -2799,7 +2799,7 @@ int rpc_clnt_test_and_add_xprt(struct rpc_clnt *clnt,
 
 	task = rpc_call_null_helper(clnt, xprt, NULL, RPC_TASK_ASYNC,
 			&rpc_cb_add_xprt_call_ops, data);
-
+	data->xps->xps_nunique_destaddr_xprts++;
 	rpc_put_task(task);
 success:
 	return 1;
diff --git a/net/sunrpc/xprtmultipath.c b/net/sunrpc/xprtmultipath.c
index c60820e45082..1693f81aae37 100644
--- a/net/sunrpc/xprtmultipath.c
+++ b/net/sunrpc/xprtmultipath.c
@@ -139,6 +139,7 @@ struct rpc_xprt_switch *xprt_switch_alloc(struct rpc_xprt *xprt,
 		xps->xps_iter_ops = &rpc_xprt_iter_singular;
 		rpc_sysfs_xprt_switch_setup(xps, xprt, gfp_flags);
 		xprt_switch_add_xprt_locked(xps, xprt);
+		xps->xps_nunique_destaddr_xprts = 1;
 		rpc_sysfs_xprt_setup(xps, xprt, gfp_flags);
 	}
 
-- 
cgit v1.2.3


From df205d0a8ea1b873a29f1333f232f884e9728acc Mon Sep 17 00:00:00 2001
From: Olga Kornievskaia <kolga@netapp.com>
Date: Fri, 27 Aug 2021 14:37:16 -0400
Subject: SUNRPC add xps_nunique_destaddr_xprts to xprt_switch_info in sysfs

In sysfs's xprt_switch_info attribute also display the value of
number of transports with unique destination addresses for this
xprt_switch.

Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 net/sunrpc/sysfs.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

(limited to 'net')

diff --git a/net/sunrpc/sysfs.c b/net/sunrpc/sysfs.c
index 414c664a3199..9a6f17e18f73 100644
--- a/net/sunrpc/sysfs.c
+++ b/net/sunrpc/sysfs.c
@@ -207,8 +207,10 @@ static ssize_t rpc_sysfs_xprt_switch_info_show(struct kobject *kobj,
 
 	if (!xprt_switch)
 		return 0;
-	ret = sprintf(buf, "num_xprts=%u\nnum_active=%u\nqueue_len=%ld\n",
+	ret = sprintf(buf, "num_xprts=%u\nnum_active=%u\n"
+		      "num_unique_destaddr=%u\nqueue_len=%ld\n",
 		      xprt_switch->xps_nxprts, xprt_switch->xps_nactive,
+		      xprt_switch->xps_nunique_destaddr_xprts,
 		      atomic_long_read(&xprt_switch->xps_queuelen));
 	xprt_switch_put(xprt_switch);
 	return ret + 1;
-- 
cgit v1.2.3


From dc48e0abee245e2f0361bd8d4e3b00f70450fab2 Mon Sep 17 00:00:00 2001
From: Olga Kornievskaia <kolga@netapp.com>
Date: Fri, 27 Aug 2021 14:37:18 -0400
Subject: SUNRPC enforce creation of no more than max_connect xprts

If we are adding new transports via rpc_clnt_test_and_add_xprt()
then check if we've reached the limit. Currently only pnfs path
adds transports via that function but this is done in
preparation when the client would add new transports when
session trunking is detected. A warning is logged if the
limit is reached.

Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 fs/nfs/client.c             | 1 +
 include/linux/sunrpc/clnt.h | 2 ++
 net/sunrpc/clnt.c           | 9 +++++++++
 3 files changed, 12 insertions(+)

(limited to 'net')

diff --git a/fs/nfs/client.c b/fs/nfs/client.c
index 486dec59972b..23e165d5ec9c 100644
--- a/fs/nfs/client.c
+++ b/fs/nfs/client.c
@@ -541,6 +541,7 @@ int nfs_create_rpc_client(struct nfs_client *clp,
 
 	clnt->cl_principal = clp->cl_principal;
 	clp->cl_rpcclient = clnt;
+	clnt->cl_max_connect = clp->cl_max_connect;
 	return 0;
 }
 EXPORT_SYMBOL_GPL(nfs_create_rpc_client);
diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h
index b2edd5fc2f0c..a4661646adc9 100644
--- a/include/linux/sunrpc/clnt.h
+++ b/include/linux/sunrpc/clnt.h
@@ -82,6 +82,7 @@ struct rpc_clnt {
 		struct work_struct	cl_work;
 	};
 	const struct cred	*cl_cred;
+	unsigned int		cl_max_connect; /* max number of transports not to the same IP */
 };
 
 /*
@@ -136,6 +137,7 @@ struct rpc_create_args {
 	char			*client_name;
 	struct svc_xprt		*bc_xprt;	/* NFSv4.1 backchannel */
 	const struct cred	*cred;
+	unsigned int		max_connect;
 };
 
 struct rpc_add_xprt_test {
diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index 451ac7d031db..f056ff931444 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -2787,6 +2787,15 @@ int rpc_clnt_test_and_add_xprt(struct rpc_clnt *clnt,
 	struct rpc_cb_add_xprt_calldata *data;
 	struct rpc_task *task;
 
+	if (xps->xps_nunique_destaddr_xprts + 1 > clnt->cl_max_connect) {
+		rcu_read_lock();
+		pr_warn("SUNRPC: reached max allowed number (%d) did not add "
+			"transport to server: %s\n", clnt->cl_max_connect,
+			rpc_peeraddr2str(clnt, RPC_DISPLAY_ADDR));
+		rcu_read_unlock();
+		return -EINVAL;
+	}
+
 	data = kmalloc(sizeof(*data), GFP_NOFS);
 	if (!data)
 		return -ENOMEM;
-- 
cgit v1.2.3