[dns-operations] EDNS buffer size on TCP connections?

[Mark Andrews]

In message <85e24b02-3243-ec24-ce96-c488e5116111 at redhat.com>, Florian Weimer wr
ites:
> On 06/22/2016 06:03 PM, Ray Bellis wrote:
> > On 22/06/2016 16:48, Wessels, Duane wrote:
> >
> >> RFC 6891 section 6.2.4 describes its use by responders.
> >
> > This text appears to be lifted straight from section 4.5.4 of RFC 2671.
> >
> > IMHO, it's wrong headed, because it's inconsistent with the
> > message-by-message semantics of EDNS0 :(
> 
> I don't think EDNS is strictly per-message.  There is a negotiation 
> mechanism which requires preserving state across messages.
> 
> Florian

What's needed is a new extended rcode UDPTOBIG for when a UDP
messages is truncated at reception.  That allow the UDP client to
detect when the server has changed its EDNS buffer size.

If a truncated message is received over UDP then UDPTOBIG is sent
to a caller with a up to date EDNS buffer size.

The are socket extensions that do this or you can also just use
<edns_size+1> and see if recv/recvfrom returns a packet bigger than
edns_size.

Mark

> _______________________________________________
> dns-operations mailing list
> dns-operations at lists.dns-oarc.net
> https://lists.dns-oarc.net/mailman/listinfo/dns-operations
> dns-operations mailing list
> https://lists.dns-oarc.net/mailman/listinfo/dns-operations
-- 
Mark Andrews, ISC
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742                 INTERNET: marka at isc.org