Thanks, Peter.

Doesn’t this mean that the OSPFv3 draft needs to create its own registry for 
the flags, then?

—John

P.S.: I see 8 instances of “ISIS” in version 19 of isis-srv6, but maybe you’re 
looking at the RFCEd version.

> On Dec 15, 2022, at 6:29 AM, Peter Psenak <[email protected]> wrote:
> 
> John,
> 
> On 14/12/2022 22:59, John Scudder wrote:
>> Hi Authors, WG,
>> 
>> As part of my review of draft-ietf-idr-bgpls-srv6-ext-12 I was looking at 
>> these documents and came up with a few comments that would otherwise become 
>> part of my AD review for ospfv3-srv6-extensions, so I thought I’d share them 
>> now.
>> 
>> draft-ietf-lsr-isis-srv6-extensions-19 is currently in RFC-EDITOR state so 
>> it may or may not be reasonable to make any changes, but I’m going to 
>> mention them below anyway. I’m sorry I didn’t notice this before we got 
>> through IESG review.
>> 
>> draft-ietf-lsr-ospfv3-srv6-extensions-08 Section 2 defines the flags field 
>> in a way that (as per usual) conveniently happens to be identical to how the 
>> IS-IS document defines it. However, I don’t see any language in 
>> draft-ietf-lsr-ospfv3-srv6-extensions-08 saying that the IANA Registry from 
>> the IS-IS draft MUST be used for the assignment of flags for the OSPFv3 
>> field. Shouldn’t you say this?
> 
> we kept the ISIS and OSPF SRv6 capabilities separate. One day they may
> diverge.
> 
>> 
>> Also, the registry should probably refer back to both (all three, if we 
>> include the BGP-LS one, but that's another story) specs that make 
>> assignments from it, shouldn’t it? So where the IS-IS document says,
>> 
>> "This document requests a new IANA registry be created under the IS-IS TLV 
>> Codepoints registry to control the assignment of bits 0 to 15 in the Flags 
>> field of the ISIS SRv6 Capabilities sub-TLV specified in this document 
>> (Section 2):"
>> 
>> Maybe it should add something like “… as well as the corresponding field 
>> defined in Section 2 of draft-ietf-lsr-ospfv3-srv6-extensions-08.” (And 
>> maybe “and Section 3.1 of draft-ietf-idr-bgpls-srv6-ext-12”?). It might be 
>> easier to drop that change into the IS-IS draft now, before it exits 
>> RFC-EDITOR state (if agreed of course) but we could also do it by patching 
>> the registry text in the OSPFv3 (and BGP-LS?) draft’s IANA section.
>> 
>> Also, a minor point: draft-ietf-lsr-isis-srv6-extensions-19 uses 
>> inconsistent hyphenation for the protocol name — “ISIS” or “IS-IS”. My own 
>> preference is “IS-IS” as both more accurate and not as subject to conflation 
>> with other uses of “Isis”, but anyway please pick one and stick with it? I 
>> guess the RFC Editor will take care of this, but I apologize for not 
>> catching it during IESG review.
> 
> I see one occurrence of "ISIS" in the latest version of the draft, I
> will ask editors to fix that.
> 
>> 
>> (This also leads me to wonder why we even put this registry, that’s shared 
>> by two (three?) protocols, under the IS-IS TLV Codepoints registry instead 
>> of under Interior Gateway Protocol (IGP) Parameters, but given the 
>> publication stage we’re at it may not be worth trying to change that.)
> 
> no, the flags are kept independent in each protocol.
> 
> thanks,
> Peter
>> 
>> Thanks,
>> 
>> —John

_______________________________________________
Lsr mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/lsr

Reply via email to