[CIFS] Make socket retry timeouts consistent between blocking and nonblocking cases

We have used approximately 15 second timeouts on nonblocking sends in the past, and
also 15 second SMB timeout (waiting for server responses, for most request types).
Now that we can do blocking tcp sends,
make blocking send timeout approximately the same (15 seconds).

Signed-off-by: Steve French <sfrench@us.ibm.com>

2 files changed, 21 insertions, 3 deletions

diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index 7419576228f..a3537a90a9d 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -1802,7 +1802,7 @@ ipv4_connect(struct TCP_Server_Info *server)
 	 *  user space buffer
 	 */
 	socket->sk->sk_rcvtimeo = 7 * HZ;
-	socket->sk->sk_sndtimeo = 3 * HZ;
+	socket->sk->sk_sndtimeo = 5 * HZ;
 
 	/* make the bufsizes depend on wsize/rsize and max requests */
 	if (server->noautotune) {
@@ -1953,7 +1953,7 @@ ipv6_connect(struct TCP_Server_Info *server)
 	 * user space buffer
 	 */
 	socket->sk->sk_rcvtimeo = 7 * HZ;
-	socket->sk->sk_sndtimeo = 3 * HZ;
+	socket->sk->sk_sndtimeo = 5 * HZ;
 	server->ssocket = socket;
 
 	return rc;
diff --git a/fs/cifs/transport.c b/fs/cifs/transport.c
index 2c7efd26992..0ad3e2d116a 100644
--- a/fs/cifs/transport.c
+++ b/fs/cifs/transport.c
@@ -199,7 +199,25 @@ smb_sendv(struct TCP_Server_Info *server, struct kvec *iov, int n_vec)
 				    n_vec - first_vec, total_len);
 		if ((rc == -ENOSPC) || (rc == -EAGAIN)) {
 			i++;
-			if (i >= 14) {
+			/* if blocking send we try 3 times, since each can block
+			   for 5 seconds. For nonblocking  we have to try more
+			   but wait increasing amounts of time allowing time for
+			   socket to clear.  The overall time we wait in either
+			   case to send on the socket is about 15 seconds.
+			   Similarly we wait for 15 seconds for
+			   a response from the server in SendReceive[2]
+			   for the server to send a response back for
+			   most types of requests (except SMB Write
+			   past end of file which can be slow, and
+			   blocking lock operations). NFS waits slightly longer
+			   than CIFS, but this can make it take longer for
+			   nonresponsive servers to be detected and 15 seconds
+			   is more than enough time for modern networks to
+			   send a packet.  In most cases if we fail to send
+			   after the retries we will kill the socket and
+			   reconnect which may clear the network problem.
+			*/
+			if ((i >= 14) || (!server->noblocksnd && (i > 2))) {
 				cERROR(1,
 				   ("sends on sock %p stuck for 15 seconds",
 				    ssocket));