Messages in this thread | | | Subject | Re: [PATCH net-next] virtio-net: switch to use build_skb() for small buffer | From | Jason Wang <> | Date | Wed, 22 Feb 2017 11:38:01 +0800 |
| |
On 2017年02月22日 11:17, Jason Wang wrote: > > > On 2017年02月22日 11:06, Michael S. Tsirkin wrote: >> On Wed, Feb 22, 2017 at 10:58:08AM +0800, Jason Wang wrote: >>> >>> On 2017年02月21日 22:37, Michael S. Tsirkin wrote: >>>> On Tue, Feb 21, 2017 at 04:46:28PM +0800, Jason Wang wrote: >>>>> This patch switch to use build_skb() for small buffer which can have >>>>> better performance for both TCP and XDP (since we can work at page >>>>> before skb creation). It also remove lots of XDP codes since both >>>>> mergeable and small buffer use page frag during refill now. >>>>> >>>>> Before | After >>>>> XDP_DROP(xdp1) 64B : 11.1Mpps | 14.4Mpps >>>>> >>>>> Tested with xdp1/xdp2/xdp_ip_tx_tunnel and netperf. >>>>> >>>>> Signed-off-by: Jason Wang<jasowang@redhat.com> >>>> Thanks! >>>> I had a similar patch for mergeable too, though it's trickier there >>>> as host has a lot of flexibility in sizing buffers. >>>> Looks like a good intermediate step to me. >>> Yes, I think it's more tricky for the case of mergeable buffer: >>> >>> 1) we need reserve NET_SKB_PAD + NET_IP_ALIGN for each buffer, this >>> will >>> break rx frag coalescing >>> 2) need tailroom for skb_shinfo, so it won't work for all size of >>> packet >>> >>> Thanks >> Have you seen my prototype? > > Not yet, please give me a pointer.
FYI, I posted a prototype during EWMA discussion in 2013:
https://lists.linuxfoundation.org/pipermail/virtualization/2013-October/025556.html
| |