Hi Guy, > On Dec 5, 2025, at 1:56 PM, Guy Harris <[email protected]> wrote: > > On Dec 5, 2025, at 12:34 PM, Mahesh Jethanandani <[email protected]> > wrote: > >> What did we decide for grandfathered entries that do not have a reference? >> Did we not decide to refer to the document itself? E.g., LinkType Value 4, >> 5, etc. > > For LinkType value 4, what does adding a reference to a document that says > > Name LINKTYPE_PRONET > Number 4 > Description Proteon PRONet Token Ring > > have over a Descrption field in the registry that says "Proteon PRONet Token > Ring"? (There's no reference because I haven't found a document giving the > detailed octet-by-octet format of the link-layer header for PRONet; if > somebody knows of such a document, I'd be happy to point to that as a > reference.) > > For LinkType value 5, we *do* have a reference - > http://www.bitsavers.org/pdf/mit/ai/AIM-628_chaosnet.pdf, which is Dave > Moon's AI lab memo on Chaosnet. > For LinkType value 6, the reference would either be IEEE Std 802.5, as the > Description field says "IEEE 802.5 Token Ring", or would be a > LINKTYPE_IEEE802_5 document on tcpdump.org giving more details, indicating > that the frame format is from 802.5 and that it doesn't include the Starting > Delimiter, Ending Delimiter, or Frame Status. For 802.5, we can probably > specifically refer to IEEE Std 802.5-1998, as I don't anticipate any further > 802.5 work. In any case, I'm not sure how citing referring back to the > I-D/RFC itself is useful other than filling a hole in the registry. > For LinkType value 7, the problem is that I haven't spent enough time staring > at the BSD ARCnet code to figure out everything weird about it. For example, > it appears that for fragmented frames as per RFC 1201/ATA 878.2 Draft ARCNET > Packet Fragmentation Standard, the fragments and the reassembled packet might > both be delivered. > So my inclination for those entries is to fill in the Reference column in the > text with "Reference not available". That was something IANA was I believe not comfortable with. In that case, I believe Amanda suggested that the reference point to the document itself. > Are registry entries mutable, so that, if a reference for a "Reference not > available" entry later becomes available, it could be added? The answer is yes, the entries are mutable. The DE(s) can decide whether the reference can be updated. And as Eliot suggested, it can be added in the “Guidance for Designated Experts” section.
Thanks Mahesh Jethanandani [email protected]
_______________________________________________ OPSAWG mailing list -- [email protected] To unsubscribe send an email to [email protected]
