Re: [tsvwg] Fwd: New Version Notification for draft-touch-tsvwg-udp-options-05.txt

[Derek Fawcus <[e-mail]>]

On Tue, Feb 28, 2017 at 01:16:07p.m. -0800, Joe Touch wrote:
> > Sorry,  I've lost the thread here.
> >
> > I was not suggesting altering how the normal UDP checksum works.
> >
> > I was mainly driving at how in implementing this in a host stack I'd
> > like to verify the options.  Namely that in the absense of the Lite
> > option,  it would allow one to use the existing checksum validation
> > routine to run across the whole of the surplus area, and if it summed
> > to the usual (1's complement) zero,  it is good.
> I'm not sure I see that.
> The existing checksum runs over the IP pseudoheader, the UDP header, and
> the UDP body.
> 
> OCS runs over just the option area.
> 
> Even in the absence of LITE, the UDP checksum wouldn't cover the option
> area.

We're at cross purposes here...

I'm not suggesting altering the UDP checksum,  I'm referring to replacing
(discarding) the OCS option.

So in section 4, figure 3 where you describe the surplus area, which is
essentially an array of bytes used for options.  I'm proposing instead to
give it a structure:

    +------+-------+
    |   checksum   |
    +------+-------+
    ... options ...
    +------+-------+

Where options is zero or more options as you describe (omitting OCS).

Where checksum is the usual checksum algorith over the whole of the
surplus area.  i.e. such that when the receiver sums the whole of the
area checksum through to the end of options, it should sum to 1s-complement
zero.  (taking in to consideration the length of options being odd).


Then for your LITE header scheme,  still require that the LITE header
be the first option,  and rather than swap 4 bytes (the LITE header)
as you describe,  instead swap 6 bytes (the checksum followed by the
LITE header).  Hence why I was the referring to partial sums.

DF