netdev - Re: [PATCH bpf-next 1/2] bpf: add BPF_LWT_ENCAP_IP option to bpf_lwt_push

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <CAPNVh5exjDdstFC2tzHKrgdMPY0kaKnJDRrWLU_oBPZT3_ngQw@mail.gmail.com>
Date:   Mon, 3 Dec 2018 08:55:59 -0800
From:   Peter Oskolkov <posk@...gle.com>
To:     David Ahern <dsahern@...il.com>
Cc:     ast@...nel.org, daniel@...earbox.net, netdev@...r.kernel.org,
        posk.devel@...il.com
Subject: Re: [PATCH bpf-next 1/2] bpf: add BPF_LWT_ENCAP_IP option to bpf_lwt_push_encap

Our case is a bit different - it is more like using the SRH header in
IPv6 to route packets via a non-default intermediate hop. But I see
your point - I'll expand the patchset to cover IPv6 and the ingress
path.
On Mon, Dec 3, 2018 at 8:04 AM David Ahern <dsahern@...il.com> wrote:
>
> On 11/30/18 5:14 PM, Peter Oskolkov wrote:
> > On Fri, Nov 30, 2018 at 3:52 PM David Ahern <dsahern@...il.com> wrote:
> >>
> >> On 11/30/18 4:35 PM, Peter Oskolkov wrote:
> >>> Thanks, David! This is for egress only, so I'll add an appropriate
> >>> check. I'll also address your other comments/concerns in a v2 version
> >>> of this patchset.
> >>
> >> Why are you limiting this to egress only?
> >
> > I'm focusing on egress because this is a use case that we have and
> > understand well, and would like to have a solution for sooner rather
> > than later.
> >
> > Without understanding why anybody would want to do lwt-bpf encap on
> > ingress, I don't know, for example, what a good test would look like;
> > but I'd be happy to look into a specific ingress use case if you have
> > one.
> >
>
> We can not have proliferation of helpers for a lot of one off use cases.
> A little thought now makes this helper useful for more than just your 1
> use case. And, IPv6 parity should be a minimal requirement for helpers.
>
> Based on your route lookup I presume your use case is capturing certain
> local traffic, pushing a custom header and sending that packet else
> where. The same could be done on the ingress path.