xprtrdma: Don't wake pending tasks until disconnect is done

Transport disconnect processing does a "wake pending tasks" at
various points.

Suppose an RPC Reply is being processed. The RPC task that Reply
goes with is waiting on the pending queue. If a disconnect wake-up
happens before reply processing is done, that reply, even if it is
good, is thrown away, and the RPC has to be sent again.

This window apparently does not exist for socket transports because
there is a lock held while a reply is being received which prevents
the wake-up call until after reply processing is done.

To resolve this, all RPC replies being processed on an RPC-over-RDMA
transport have to complete before pending tasks are awoken due to a
transport disconnect.

Callers that already hold the transport write lock may invoke
->ops->close directly. Others use a generic helper that schedules
a close when the write lock can be taken safely.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

1 files changed, 10 insertions, 7 deletions

diff --git a/net/sunrpc/xprtrdma/transport.c b/net/sunrpc/xprtrdma/transport.c
index 91c476a8f51c..134aae2ee779 100644
--- a/net/sunrpc/xprtrdma/transport.c
+++ b/net/sunrpc/xprtrdma/transport.c
@@ -437,8 +437,7 @@ out1:
  * Caller holds @xprt's send lock to prevent activity on this
  * transport while the connection is torn down.
  */
-static void
-xprt_rdma_close(struct rpc_xprt *xprt)
+void xprt_rdma_close(struct rpc_xprt *xprt)
 {
 	struct rpcrdma_xprt *r_xprt = rpcx_to_rdmax(xprt);
 	struct rpcrdma_ep *ep = &r_xprt->rx_ep;
@@ -453,13 +452,13 @@ xprt_rdma_close(struct rpc_xprt *xprt)
 
 	if (test_and_clear_bit(RPCRDMA_IAF_REMOVING, &ia->ri_flags)) {
 		rpcrdma_ia_remove(ia);
-		return;
+		goto out;
 	}
+
 	if (ep->rep_connected == -ENODEV)
 		return;
 	if (ep->rep_connected > 0)
 		xprt->reestablish_timeout = 0;
-	xprt_disconnect_done(xprt);
 	rpcrdma_ep_disconnect(ep, ia);
 
 	/* Prepare @xprt for the next connection by reinitializing
@@ -467,6 +466,10 @@ xprt_rdma_close(struct rpc_xprt *xprt)
 	 */
 	r_xprt->rx_buf.rb_credits = 1;
 	xprt->cwnd = RPC_CWNDSHIFT;
+
+out:
+	++xprt->connect_cookie;
+	xprt_disconnect_done(xprt);
 }
 
 /**
@@ -717,7 +720,7 @@ xprt_rdma_send_request(struct rpc_rqst *rqst)
 #endif	/* CONFIG_SUNRPC_BACKCHANNEL */
 
 	if (!xprt_connected(xprt))
-		goto drop_connection;
+		return -ENOTCONN;
 
 	if (!xprt_request_get_cong(xprt, rqst))
 		return -EBADSLT;
@@ -749,8 +752,8 @@ failed_marshal:
 	if (rc != -ENOTCONN)
 		return rc;
 drop_connection:
-	xprt_disconnect_done(xprt);
-	return -ENOTCONN;	/* implies disconnect */
+	xprt_rdma_close(xprt);
+	return -ENOTCONN;
 }
 
 void xprt_rdma_print_stats(struct rpc_xprt *xprt, struct seq_file *seq)