SUNRPC: Avoid an unnecessary task reschedule on ENOTCONN

If the socket is unconnected, and xprt_transmit() returns ENOTCONN, we
currently give up the lock on the transport channel. Doing so means that
the lock automatically gets assigned to the next task in the xprt->sending
queue, and so that task needs to be woken up to do the actual connect.

The following patch aims to avoid that unnecessary task switch.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

1 files changed, 18 insertions, 8 deletions

diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index 836f15c0c4a3..07e9b05321e6 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -1105,14 +1105,24 @@ static void
 call_transmit_status(struct rpc_task *task)
 {
         task->tk_action = call_status;
-        /*
-         * Special case: if we've been waiting on the socket's write_space()
-         * callback, then don't call xprt_end_transmit().
-         */
-        if (task->tk_status == -EAGAIN)
-                return;
-        xprt_end_transmit(task);
-        rpc_task_force_reencode(task);
+        switch (task->tk_status) {
+        case -EAGAIN:
+                break;
+        default:
+                xprt_end_transmit(task);
+                /*
+                 * Special cases: if we've been waiting on the
+                 * socket's write_space() callback, or if the
+                 * socket just returned a connection error,
+                 * then hold onto the transport lock.
+                 */
+        case -ECONNREFUSED:
+        case -ENOTCONN:
+        case -EHOSTDOWN:
+        case -EHOSTUNREACH:
+        case -ENETUNREACH:
+                rpc_task_force_reencode(task);
+        }
 }
 
 /*