netdev - Re: [PATCH net-next v2 05/12] flow_dissector: Parse vxlan in UDP

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

Message-ID: <CALx6S36RJ9poHZBo+T0ZT1YB0Z6gEV5iiC+-C-SJDrBbd1wWvA@mail.gmail.com>
Date: Tue, 20 Aug 2024 11:15:56 -0700
From: Tom Herbert <tom@...bertland.com>
To: Jakub Kicinski <kuba@...nel.org>
Cc: davem@...emloft.net, edumazet@...gle.com, netdev@...r.kernel.org, 
	felipe@...anda.io, willemdebruijn.kernel@...il.com
Subject: Re: [PATCH net-next v2 05/12] flow_dissector: Parse vxlan in UDP

On Fri, Aug 16, 2024 at 6:09 PM Jakub Kicinski <kuba@...nel.org> wrote:
>
> On Thu, 15 Aug 2024 14:45:20 -0700 Tom Herbert wrote:
> > +     if (hdr->vx_flags & VXLAN_F_GPE) {
>
> sparse points out that VXLAN_F_GPE needs to be byte swapped
>
> not 100% sure I'm following the scheme but it appears _F_ flags are in
> CPU byte order, and _HF_ flags are in network

Hmmm, VXLAN does seem to be a bit squirrely :-) AFAICT, a VXLAN-GPE
packet isn't self identifying but requires a VNI lookup which is what
the vxlan driver. We don't have access to that table in flow
dissector, but I think we can safely infer a VXLAN-GPE packet if the
next proto-applied flag bit is set (it's always set for VXLAN-GPE,
never for VXLAN).

Tom

> --
> pw-bot: cr