jamal writes:
> > Here we process (drop) about 13% packets more when skb'a get reued.
> >
> Very cool.
> Robert, it would be interesting to see something more interesting
> (longer code path) such as udp. You can still use a packetgen to shoot
> at the box (netperf is a wimp), send it to udp port 9 which just
> swallows packets.
Neither UDP or TCP kfree's the skb in it's normal path today so it cannot
be resused. ICMP packets seems get reused as-is. Dave and Andi some possible
improvement for TCP.
For UDP I just took a quick look. Only in case errors the skb gets freed.
So ie sending UDP pkt to socket which is not open would make the skb's reused.
And someting we could test...
The normal path from what I understand is queue packets in into the socket
receive queue.
skb_queue_tail(&sk->sk_receive_queue, skb);
Which could be an excellent place for UDP copy break code... which then
makes skb reuse possible.
Cheers.
--ro
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
