Re: Question about veth_xmit()

2017-01-23 Thread Cong Wang
On Mon, Jan 23, 2017 at 2:05 PM, Eric Dumazet wrote: > On Mon, 2017-01-23 at 13:46 -0800, Xiangning Yu wrote: >> On Mon, Jan 23, 2017 at 12:56 PM, Cong Wang wrote: >> > On Mon, Jan 23, 2017 at 10:46 AM, Xiangning Yu >> > wrote: >> >> Hi netdev folks, >> >> >> >> It looks like we call dev_forwar

Re: Question about veth_xmit()

2017-01-23 Thread Eric Dumazet
On Mon, 2017-01-23 at 13:46 -0800, Xiangning Yu wrote: > On Mon, Jan 23, 2017 at 12:56 PM, Cong Wang wrote: > > On Mon, Jan 23, 2017 at 10:46 AM, Xiangning Yu > > wrote: > >> Hi netdev folks, > >> > >> It looks like we call dev_forward_skb() in veth_xmit(), which calls > >> netif_rx() eventually

Re: Question about veth_xmit()

2017-01-23 Thread Xiangning Yu
On Mon, Jan 23, 2017 at 12:56 PM, Cong Wang wrote: > On Mon, Jan 23, 2017 at 10:46 AM, Xiangning Yu wrote: >> Hi netdev folks, >> >> It looks like we call dev_forward_skb() in veth_xmit(), which calls >> netif_rx() eventually. >> >> While netif_rx() will enqueue the skb to the CPU RX backlog befo

Re: Question about veth_xmit()

2017-01-23 Thread Cong Wang
On Mon, Jan 23, 2017 at 10:46 AM, Xiangning Yu wrote: > Hi netdev folks, > > It looks like we call dev_forward_skb() in veth_xmit(), which calls > netif_rx() eventually. > > While netif_rx() will enqueue the skb to the CPU RX backlog before the > actual processing takes place. So, this actually me

Re: Question about veth_xmit()

2017-01-23 Thread Xiangning Yu
On Mon, Jan 23, 2017 at 11:07 AM, Eric Dumazet wrote: > On Mon, 2017-01-23 at 10:46 -0800, Xiangning Yu wrote: >> Hi netdev folks, >> >> It looks like we call dev_forward_skb() in veth_xmit(), which calls >> netif_rx() eventually. >> >> While netif_rx() will enqueue the skb to the CPU RX backlog b

Re: Question about veth_xmit()

2017-01-23 Thread Eric Dumazet
On Mon, 2017-01-23 at 10:46 -0800, Xiangning Yu wrote: > Hi netdev folks, > > It looks like we call dev_forward_skb() in veth_xmit(), which calls > netif_rx() eventually. > > While netif_rx() will enqueue the skb to the CPU RX backlog before the > actual processing takes place. So, this actually

Question about veth_xmit()

2017-01-23 Thread Xiangning Yu
Hi netdev folks, It looks like we call dev_forward_skb() in veth_xmit(), which calls netif_rx() eventually. While netif_rx() will enqueue the skb to the CPU RX backlog before the actual processing takes place. So, this actually means a TX skb has to wait some un-related RX skbs to finish. And thi