Re: [dhcwg] VIV*O options.

On Wednesday 17 May 2006 22:12, David W. Hankins wrote:

> On Wed, May 17, 2006 at 12:17:26PM -0700, Andre Kostur wrote:
> > As long as you start after the Enterprise OID (and data length), that
> > shouldn't be an issue.
>
> That's precisely the rub: this would use the same loop to process
> VIV*O options (and also their contents, but that's not the point).
>
> That is, VIV*O options (DHCP codes 124 and 125) are merely an
> encapsulated space whose tag width is 4, and whose length width is
> 1, and which generally contain different encapsulated spaces.  Compare
> DHCP which is 1/1, DHCPv6 which is 2/2.
>
> I'm suggesting that the generic functions which implement that would
> merely accept widths of 1, 2, 4 for code fields, and 1, 2 for length
> fields - any combination.  1/2 is legal, 2/1 is legal.

What you suggest is how we do it. The width of the tag, len and type fields 
are primarily defined by a "Protocol Attributes" class, but the meta-data for 
any given option can override those (individually) if need be.

So I'd describe the option in 3925 as a regular DHCPv4 option, meaning 1/1, 
but then provide these settings for the option's meta-data:

     type = subencoded
     subtag_width = 4
     sublen_width = 1

This allows us to iterate over the same loop, reading tag/len fields of any 
size.

Using a default of 1/1 or 2/2, but providing option-specific overrides seems 
to work well in practice.

- Bud

Bud Millwood
Weird Solutions, Inc.
http://www.weird-solutions.com
tel: +46 8 758 3700
fax: +46 8 758 3687
mailto:budm@weird-solutions.com

_______________________________________________
dhcwg mailing list
dhcwg@ietf.org
https://www1.ietf.org/mailman/listinfo/dhcwg