Re: [tsvwg] draft-ietf-udp-options issues from IETF 104

[Tom Herbert <[e-mail]>]

On Sat, Jul 20, 2019 at 6:50 PM Black, David <[e-mail]> wrote:
>
> Still as an individual ...
>
> > What I was actually advocating was that in the presence of any option marked
> > "drop if unknown" there MUST NOT be any legacy data, i.e, such options
> > could not appear in the trailer format. All payload would have to be in the
> > surplus area (which requires the header format). In that case, "drop the
> > packet" and "drop the surplus" would amount to the same thing.
>
> That would "freeze" LITE - if we ever found anything problematic with LITE and wanted to replace it with LITE2, that replacement would be prohibited by such a requirement.
>
David,

This is why having a version number, variant number, or the like in
the surplus area header is so critical.

Tom

> NB: *replace* is crucial in this context - this is not about trying to use both LITE and LITE2 in the same packet.
>
> Thanks, --David
>
> > -----Original Message-----
> > From: tsvwg <[e-mail]> On Behalf Of C. M. Heard
> > Sent: Saturday, July 20, 2019 8:51 PM
> > To: Derek Fawcus
> > Cc: tsvwg
> > Subject: Re: [tsvwg] draft-ietf-udp-options issues from IETF 104
> >
> >
> > [EXTERNAL EMAIL]
> >
> > On Sat, Jul 20, 2019 at 2:56 PM Derek Fawcus wrote:
> > > I suspect we should interpret "drop the packet" here for options marked
> > > as "drop if not known" to actually mean "discard the surplus".
> > >
> > > i.e. if a non-legacy receiver in processing the surplus finds such an
> > > option, it then switches to operating in legacy mode for that packet.
> >
> > I do not agree.
> >
> > > Which means that like the legacy receiver, it will either see a zero length
> > > UDP packet (if UDP payload length was zero) or see whatever legacy UDP data
> > > was passed in the non-surplus area.
> >
> > What I was actually advocating was that in the presence of any option marked
> > "drop if unknown" there MUST NOT be any legacy data, i.e, such options
> > could not appear in the trailer format. All payload would have to be in the
> > surplus area (which requires the header format). In that case, "drop the
> > packet" and "drop the surplus" would amount to the same thing.
> >
> > Mike
>