macvtap: avoid hash calculating for single queue

We decide the rxq through calculating its hash which is not necessary
if we only have one rx queue. So this patch skip this and just return
queue 0. Test shows 22% improving on guest rx pps.

Before: 1201504 pkts/s
After:  1472731 pkts/s

Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

diff --git a/drivers/net/macvtap.c b/drivers/net/macvtap.c
index 95a13321f68826f6e741ee2fe834bb44a641005e..2476923b424d88ccffbff4028fec812e620d551c 100644 (file)
--- a/drivers/net/macvtap.c
+++ b/drivers/net/macvtap.c
@@ -299,6 +299,9 @@ static struct macvtap_queue *macvtap_get_queue(struct net_device *dev,
        if (!numvtaps)
                goto out;
 
+       if (numvtaps == 1)
+               goto single;
+
        /* Check if we can use flow to select a queue */
        rxq = skb_get_hash(skb);
        if (rxq) {
@@ -316,6 +319,7 @@ static struct macvtap_queue *macvtap_get_queue(struct net_device *dev,
                goto out;
        }
 
+single:
        tap = rcu_dereference(vlan->taps[0]);
 out:
        return tap;