[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <6668bd2632a87_f6b0e294fd@willemb.c.googlers.com.notmuch>
Date: Tue, 11 Jun 2024 17:09:58 -0400
From: Willem de Bruijn <willemdebruijn.kernel@...il.com>
To: YiFei Zhu <zhuyifei@...gle.com>,
netdev@...r.kernel.org,
bpf@...r.kernel.org
Cc: Björn Töpel <bjorn@...nel.org>,
Magnus Karlsson <magnus.karlsson@...el.com>,
Maciej Fijalkowski <maciej.fijalkowski@...el.com>,
Jonathan Lemon <jonathan.lemon@...il.com>,
Alexei Starovoitov <ast@...nel.org>,
Daniel Borkmann <daniel@...earbox.net>,
"David S . Miller" <davem@...emloft.net>,
Jakub Kicinski <kuba@...nel.org>,
Jesper Dangaard Brouer <hawk@...nel.org>,
John Fastabend <john.fastabend@...il.com>,
Andrii Nakryiko <andrii@...nel.org>,
Stanislav Fomichev <sdf@...gle.com>,
Willem de Bruijn <willemb@...gle.com>
Subject: Re: [RFC PATCH net-next 3/3] selftests: drv-net: Add xsk_hw AF_XDP
functionality test
YiFei Zhu wrote:
> Run tools/testing/selftest/bpf/xsk_hw.c as part of drv-net.
>
> A matrix of TX/RX, copy/zerocopy, and driver mode / skb mode, are
> tested. Additionally, it tests some edge cases such as:
> - Zerocopy TX with and without attaching an XDP prog.
> - Zerocopy RX where binding happens before fillq gets filled.
>
> TX and RX are tested separately, and the remote side always runs
> the basic AF_PACKET handler rather than AF_XDP, in order to
> isolate potential causes of test failures.
>
> Currently the next-hop MAC address of each side must be manually
> specified via LOCAL_NEXTHOP_MAC & REMOTE_NEXTHOP_MAC. It's probably
> doable to detect these addresses automatically, but it's future work,
> and probably library code since it is also applicable to csum.py.
>
> Signed-off-by: YiFei Zhu <zhuyifei@...gle.com>
Reviewed-by: Willem de Bruijn <willemb@...gle.com>
Powered by blists - more mailing lists