ceph: preserve seq # on requeued messages after transient transport e…

…rrors

If the tcp connection drops and we reconnect to reestablish a stateful
session (with the mds), we need to resend previously sent (and possibly
received) messages with the _same_ seq # so that they can be dropped on
the other end if needed.  Only assign a new seq once after the message is
queued.

Signed-off-by: Sage Weil <sage@newdream.net>

fs/ceph/messenger.c

@@ -492,7 +492,14 @@ static void prepare_write_message(struct ceph_connection *con)
 		list_move_tail(&m->list_head, &con->out_sent);
 	}
 
-	m->hdr.seq = cpu_to_le64(++con->out_seq);
+	/*
+	 * only assign outgoing seq # if we haven't sent this message
+	 * yet.  if it is requeued, resend with it's original seq.
+	 */
+	if (m->needs_out_seq) {
+		m->hdr.seq = cpu_to_le64(++con->out_seq);
+		m->needs_out_seq = false;
+	}
 
 	dout("prepare_write_message %p seq %lld type %d len %d+%d+%d %d pgs\n",
 	     m, con->out_seq, le16_to_cpu(m->hdr.type),
@@ -1986,6 +1993,8 @@ void ceph_con_send(struct ceph_connection *con, struct ceph_msg *msg)
 
 	BUG_ON(msg->front.iov_len != le32_to_cpu(msg->hdr.front_len));
 
+	msg->needs_out_seq = true;
+
 	/* queue */
 	mutex_lock(&con->mutex);
 	BUG_ON(!list_empty(&msg->list_head));

fs/ceph/messenger.h

@@ -86,6 +86,7 @@ struct ceph_msg {
 	struct kref kref;
 	bool front_is_vmalloc;
 	bool more_to_follow;
+	bool needs_out_seq;
 	int front_max;
 
 	struct ceph_msgpool *pool;