xen-netback: immediately wake tx queue when guest rx queue has space

When an skb is removed from the guest rx queue, immediately wake the
tx queue, instead of after processing them.
Signed-off-by: NDavid Vrabel <david.vrabel@citrix.com>
[re-based]
Signed-off-by: NPaul Durrant <paul.durrant@citrix.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

drivers/net/xen-netback/rx.c

@@ -92,25 +92,19 @@ static struct sk_buff *xenvif_rx_dequeue(struct xenvif_queue *queue)
 	spin_lock_irq(&queue->rx_queue.lock);
 
 	skb = __skb_dequeue(&queue->rx_queue);
-	if (skb)
+	if (skb) {
 		queue->rx_queue_len -= skb->len;
+		if (queue->rx_queue_len < queue->rx_queue_max) {
+			struct netdev_queue *txq;
 
-	spin_unlock_irq(&queue->rx_queue.lock);
-
-	return skb;
-}
-
-static void xenvif_rx_queue_maybe_wake(struct xenvif_queue *queue)
-{
-	spin_lock_irq(&queue->rx_queue.lock);
-
-	if (queue->rx_queue_len < queue->rx_queue_max) {
-		struct net_device *dev = queue->vif->dev;
-
-		netif_tx_wake_queue(netdev_get_tx_queue(dev, queue->id));
+			txq = netdev_get_tx_queue(queue->vif->dev, queue->id);
+			netif_tx_wake_queue(txq);
+		}
 	}
 
 	spin_unlock_irq(&queue->rx_queue.lock);
+
+	return skb;
 }
 
 static void xenvif_rx_queue_purge(struct xenvif_queue *queue)
@@ -585,8 +579,6 @@ int xenvif_kthread_guest_rx(void *data)
 		 */
 		xenvif_rx_queue_drop_expired(queue);
 
-		xenvif_rx_queue_maybe_wake(queue);
-
 		cond_resched();
 	}