Re: [TLS] WGLC for draft-ietf-tls-flags

[Yoav Nir <[e-mail]>]

Thanks for the review. Comments inline.

> On 19 Jul 2021, at 2:26, Michael StJohns <[e-mail]> wrote:
> 
> On 7/16/2021 7:55 PM, Christopher Wood wrote:
>> This is the second working group last call for the "A Flags Extension for TLS 1.3" draft, available here:
>> 
>>     https://datatracker.ietf.org/doc/draft-ietf-tls-tlsflags/
>> 
>> Please review this document and send your comments to the list by July 30, 2021. The GitHub repository for this draft is available here:
>> 
>>     https://github.com/tlswg/tls-flags
>> 
>> Thanks,
>> Chris, on behalf of the chairs
>> 
>> _______________________________________________
>> TLS mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/tls
> 
> 
> Hi - I have not followed the discussion of this document on the mailing list so this review is only against the document itself. It's possible that these concerns have already been discussed.
> 
> 
> Section 2 requires  (MUST) the generation of fatal illegal_parameter alert upon reception of a mal-encoded extension (e.g. any trailing zero bytes), but compare and contrast this with section 3 which is full of MUST and MUST NOT declarations but with no concrete actions to be taken.  E.g. if I (server or client) send 0x01 0x10, and receive 0x01 0x11 from the client or server, wouldn't that be an illegal value as I've added a bit not sent to me?   Should that cause the same fatal illegal_parameter alert? Alternately, "receiver MUST ignore received bits that weren't sent" language could clean this up.

I prefer the hard error, but this should be discussed in the WG. Perhaps a PR is in order?

> 
> Section 4 is a bit painful to read in that it took me three read-throughs to understand that what the document is asking for is a monolithic registry which requires "expert review" for all registrations, but where the experts are responsible for the sub range determinations.   Usually, that's not the way the IANA works.  If a registry has distinct set of ranges, each range normally has a specific registration procedure that the IANA follows before placing a parameter in that registry.
> 

The way it works with all the TLS registries is that IANA asks the expert for advice on which numbers to assign regardless of how the registry is segmented.  So it’s up to the experts anyway.

Yoav