Virtualization - Apr 2022 - [PATCH net-next] net/af_packet: add VLAN support for AF_PACKET SOCK_RAW GSO

On Sat, Apr 23, 2022 at 10:31 PM Hangbin Liu <liuhangbin at gmail.com>
wrote:
>
> On Fri, Apr 22, 2022 at 05:39:48PM -0400, Willem de Bruijn wrote:
> > > If we split skb_probe_transport_header() from
packet_parse_headers() and
> > > move it before calling virtio_net_hdr_* function in packet_snd().
Should
> > > we do the same for tpacket_snd(), i.e. move
skb_probe_transport_header()
> > > after the virtio_net_hdr_* function?
> >
> > That sounds like the inverse: "move after" instead of
"move before"?
>
> That's for "split packet_parse_headers()" option.
>
> >
> > But I thought the plan was to go back to your last patch which brings
> > packet_snd in line with tpacket_snd by moving packet_parse_headers in
> > its entirety before virtio_net_hdr_*?
>
> Yes, exactly.
>
> > > So my conclusion is. There is no need to split
packet_parse_headers(). Move
> > > packet_parse_headers() before calling virtio_net_hdr_* function
in packet_snd()
> > > should be safe.
> >
> > Ack. Sorry if my last response was not entirely clear on this point.
>
> Thanks a lot for your review. Do you think if I need to re-post the patch?
> Or will you give an Acked-by for this one?
Please resubmit. And then please also add the comment about tap having
the same path.
>
> Hangbin