chainntnfs/neutrino: start concurrent queues prior to connection

Similar to what was done for btcd, just to make sure we won't be blocked
on any incoming notifications, start the queues first thing during
startup.

chainntnfs/neutrinonotify/neutrino.go

@@ -116,12 +116,20 @@ func (n *NeutrinoNotifier) Start() error {
 		return nil
 	}
 
+	// Start our concurrent queues before starting the rescan, to ensure
+	// onFilteredBlockConnected and onRelavantTx callbacks won't be
+	// blocked.
+	n.chainUpdates.Start()
+	n.txUpdates.Start()
+
 	// First, we'll obtain the latest block height of the p2p node. We'll
 	// start the auto-rescan from this point. Once a caller actually wishes
 	// to register a chain view, the rescan state will be rewound
 	// accordingly.
 	startingPoint, err := n.p2pNode.BestBlock()
 	if err != nil {
+		n.txUpdates.Stop()
+		n.chainUpdates.Stop()
 		return err
 	}
 	n.bestBlock.Hash = &startingPoint.Hash
@@ -160,9 +168,6 @@ func (n *NeutrinoNotifier) Start() error {
 	)
 	n.rescanErr = n.chainView.Start()
 
-	n.chainUpdates.Start()
-	n.txUpdates.Start()
-
 	n.wg.Add(1)
 	go n.notificationDispatcher()