On Thu, Jun 02, 2011 at 06:43:17PM +0300, Michael S. Tsirkin wrote:
> Current code might introduce a lot of latency variation
> if there are many pending bufs at the time we
> attempt to transmit a new one. This is bad for
> real-time applications and can't be good for TCP either.
>
> Free up just enough to both clean up all buffers
> eventually and to be able to xmit the next packet.
>
> Signed-off-by: Michael S. Tsirkin <mst@xxxxxxxxxx>
I've been testing this patch and it seems to work fine
so far. The following fixups are needed to make it
build though:
diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index b25db1c..77cdf34 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -529,11 +529,8 @@ static bool free_old_xmit_skb(struct virtnet_info *vi)
* virtqueue_add_buf will succeed. */
static bool free_xmit_capacity(struct virtnet_info *vi)
{
- struct sk_buff *skb;
- unsigned int len;
-
while (virtqueue_min_capacity(vi->svq) < MAX_SKB_FRAGS + 2)
- if (unlikely(!free_old_xmit_skb))
+ if (unlikely(!free_old_xmit_skb(vi)))
return false;
return true;
}
@@ -628,7 +625,7 @@ static netdev_tx_t start_xmit(struct sk_buff *skb, struct net_device *dev)
* Doing this after kick means there's a chance we'll free
* the skb we have just sent, which is hot in cache. */
for (i = 0; i < 2; i++)
- free_old_xmit_skb(v);
+ free_old_xmit_skb(vi);
if (likely(free_xmit_capacity(vi)))
return NETDEV_TX_OK;
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
