Re: [dns-privacy] I-D Action: draft-ietf-dprive-early-data-00.txt

[Ilari Liusvaara <[e-mail]>]

On Wed, Apr 22, 2020 at 06:41:39AM -0700, [e-mail] wrote:
> 
> A New Internet-Draft is available from the on-line Internet-Drafts directories.
> This draft is a work item of the DNS PRIVate Exchange WG of the IETF.
> 
>         Title           : Using Early Data in DNS over TLS
>         Author          : Alessandro Ghedini
> 	Filename        : draft-ietf-dprive-early-data-00.txt
> 	Pages           : 6
> 	Date            : 2020-04-22

That RRTYPE whitelist looks quite questionable. Any Data RRTYPE (numbers
1-127 and 256-61439) needs to be safe as QTYPE, or there are major
problems already (since servers MUST answer all of them).

Meta RRTYPEs (numbers 128-255) might be unsafe (and servers are
allowed to reject such queries already). Then there is the unassigned,
private use and reserved stuff (numbers 0, 61440-65535) and who knows
what is there.

Unfortunately there is the special snowflake that is OPT (number 41).
Despite being in DATA RRTYPE range, it is special (usually not even used
as QTYPE). Now, the base structure absolutely has to be allowed in
0-RTT. The problem with it is that it can carry its own extensions. I
have no idea of what most of those even do, and there are probably at
least some that are unsafe in 0-RTT, and at least some that are
actually useful in 0-RTT.

(As sidenote, I discovered that Unbound does not like OPT as QTYPE
and answers with FORMERR, but there are servers out there that answer
OPT queries like any other datatype they have no records for).


-Ilari