Re: AF_XDP metadata/hints

From: Alexei Starovoitov <alexei.starovoitov@gmail.com>
To: Alexander Lobakin <alexandr.lobakin@intel.com>
Cc: "John Fastabend" <john.fastabend@gmail.com>,
	"Toke Høiland-Jørgensen" <toke@redhat.com>,
	"Jesper Dangaard Brouer" <brouer@redhat.com>,
	"Saeed Mahameed" <saeed@kernel.org>,
	"Raczynski, Piotr" <piotr.raczynski@intel.com>,
	"Zhang, Jessica" <jessica.zhang@intel.com>,
	"Kubiak, Marcin" <marcin.kubiak@intel.com>,
	"Joseph, Jithu" <jithu.joseph@intel.com>,
	"kurt@linutronix.de" <kurt@linutronix.de>,
	"Maloor, Kishen" <kishen.maloor@intel.com>,
	"Gomes, Vinicius" <vinicius.gomes@intel.com>,
	"Brandeburg, Jesse" <jesse.brandeburg@intel.com>,
	"Swiatkowski, Michal" <michal.swiatkowski@intel.com>,
	"Plantykow, Marta A" <marta.a.plantykow@intel.com>,
	"Ong, Boon Leong" <boon.leong.ong@intel.com>,
	"Desouza, Ederson" <ederson.desouza@intel.com>,
	"Song, Yoong Siang" <yoong.siang.song@intel.com>,
	"Czapnik, Lukasz" <lukasz.czapnik@intel.com>,
	bpf <bpf@vger.kernel.org>,
	xdp-hints@xdp-project.net
Subject: Re: AF_XDP metadata/hints
Date: Wed, 26 May 2021 09:41:52 -0700	[thread overview]
Message-ID: <CAADnVQL3uuKY4kY3v60Wzjh1QPT+k4+jVnN+Y3a_SBF3DFbwWg@mail.gmail.com> (raw)
In-Reply-To: <20210526155402.172-1-alexandr.lobakin@intel.com>

On Wed, May 26, 2021 at 8:57 AM Alexander Lobakin
<alexandr.lobakin@intel.com> wrote:
> >
> >Well likely libbpf would do the rewrite I think.
>
> So your proposal is to not compose metadata according to the prog's
> request, but rather reprogram the prog itself to access metadata
> accordingly? Sounds very nice.
>
> If follow this path, is it something like this?
>
> 1. Driver exposes the fields layout (e.g. Rx/Tx descriptor fields)
> via BTF to the BPF layer.
> 2. When an XDP prog is attached, BPF reprograms it to look for the
> required fields at the right offset.

The driver doesn't need to expose it directly via ndo.
There is already generic support for BTF in modules
and support for encoding btf_id for further use inside verifier
and other components.
I think the driver can simply do:
BTF_ID_LIST(known_packet_fields)
and the bpf core will pick it from there.
While libbpf will do a CO-RE style re-write when driver layout changes.
Ideally bpf core doesn't need to be involved and it's done completely in libbpf.