[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <867d5d17-753a-10f6-fba5-1c854b891b03@gmail.com>
Date: Fri, 24 Apr 2020 14:23:46 -0600
From: David Ahern <dsahern@...il.com>
To: Toke Høiland-Jørgensen <toke@...hat.com>,
David Ahern <dsahern@...nel.org>, netdev@...r.kernel.org
Cc: davem@...emloft.net, kuba@...nel.org,
prashantbhole.linux@...il.com, jasowang@...hat.com,
brouer@...hat.com, toshiaki.makita1@...il.com,
daniel@...earbox.net, john.fastabend@...il.com, ast@...nel.org,
kafai@...com, songliubraving@...com, yhs@...com, andriin@...com,
David Ahern <dahern@...italocean.com>
Subject: Re: [PATCH v2 bpf-next 06/17] net: Add IFLA_XDP_EGRESS for XDP
programs in the egress path
On 4/24/20 2:57 AM, Toke Høiland-Jørgensen wrote:
>
> I think this will be a problem: Since we don't reject unknown netlink
> attributes, userspace can't know if xdp_egress is supported. So maybe
Well known problem with networking and one Johannes, myself and others
discussed and worked on a solution for -- add strict_start_type to new
attributes so going forward we can detect that support directly vs
indirectly (probing). For link attributes that opportunity was missed in
December when IFLA_PERM_ADDRESS was added.
> it's better to just re-use IFLA_XDP with a new flag, instead of using a
> separate attribute?
After considering how the code has evolved over past few months, I
reworked the uapi to be more consistent with existing XDP attach modes.
Powered by blists - more mailing lists