[ovs-dev] iperf tcp issue on veth using afxdp

[Eelco Chaudron]

On 22 Nov 2019, at 19:50, Ilya Maximets wrote:

> On 22.11.2019 18:51, William Tu wrote:
>> Hi Ilya and Eelco,
>>
>> Yiyang reports very poor TCP performance on his setup and I can
>> also reproduce it on my machine. Somehow I think this might be a
>> kernel issue, but I don't know where to debug this. Need your 
>> suggestion
>> about how to debug.
>>
>> So the setup is like the system-traffic, creating 2 namespaces and
>> veth devices and attach to OVS. I do remember to turn off tx offload
>> and ping, UDP, nc (tcp-mode) works fine.
>>
>> TCP using iperf drops to 0Mbps after 4 seconds.
>
> The key questions are:
> * What is your kernel version?
> * Does it work in generic mode?
> * Does it work if iperf generates UDP traffic?
>   ex. iperf3 -c 10.1.1.1 -t 3600 -u -b 10G/64 -l 1460
> * It seems like a ring breakage or a umem memory leak.
>   So, what are the afxdp related coverage counters?
> * Does OVS forward packets, i.e. if there something received/sent?
>
> Best regards, Ilya Maximets.

This looks very much like the issues we had earlier with tap interfaces, 
this issue was related to OVS not managing the umem ring correctly (and 
hence running out of buffers). Guess the coverage counters suggested 
above should show give you a good starting point for debugging.

//Eelco