Re: [Intel-wired-lan] FW: [PATCH bpf-next 2/4] xsk: allow AF_XDP sockets to receive packets directly from a queue
From: Alexei Starovoitov <hidden>
Date: 2019-10-31 23:15:49
Also in:
bpf, intel-wired-lan
From: Alexei Starovoitov <hidden>
Date: 2019-10-31 23:15:49
Also in:
bpf, intel-wired-lan
On Thu, Oct 31, 2019 at 3:38 PM Samudrala, Sridhar [off-list ref] wrote:
Alexei, Jakub Do you think it will be possible to avoid this overhead when mitigations are turned ON?
yes
The other part of the overhead is going through the redirect path. Can i assume that your silence as an indication that you are now okay with optional bypass flag as long as it doesn't effect the normal XDP datapath. If so, i will respin and submit the patches against the latest bpf-next
I'm still against it. Looks like the only motivation for it is extra overhead due to retpolines. imo it's not a good reason to introduce a bunch of extra code helping single kernel feature. We will have proper solution for indirect calls.