RFC Errata


Errata Search

 
Source of RFC  
Summary Table Full Records

Found 2 records.

Status: Verified (1)

RFC 8765, "DNS Push Notifications", June 2020

Source of RFC: dnssd (int)

Errata ID: 8560
Status: Verified
Type: Technical
Publication Format(s) : TEXT, PDF, HTML

Reported By: Ilya Kulakov
Date Reported: 2025-08-31
Verifier Name: Éric Vyncke
Date Verified: 2025-11-06

Section 6.5 says:

For other types of DNS server, the RECONFIRM operation is currently
undefined and SHOULD result in a NOERROR response, but it need not
cause any other action to occur.

It should say:

For other types of DNS server, the RECONFIRM operation is currently
undefined and MUST be silently ignored on reception.

Notes:

RECONFIRM TLV is only allowed as a Primary TLV in a Unidirectional DSO message. Per RFC 8490, no response must be generated for Unidirectional DSO messages: server either processes, ignores or aborts the connection. Therefore it's wrong to recommend to "a NOERROR response".

--- Verifier note (Éric Vyncke) ---

See also: https://mailarchive.ietf.org/arch/msg/dnssd/FP00REBOxT0PsxuqKZmQ1XvbaNw/

Status: Reported (1)

RFC 8765, "DNS Push Notifications", June 2020

Source of RFC: dnssd (int)

Errata ID: 8631
Status: Reported
Type: Technical
Publication Format(s) : HTML

Reported By: Miek Gieben
Date Reported: 2025-11-12

Section 6.3.1 says:

n/a

See notes below

It should say:

n/a

Notes:

Hello,

This concerns the text in section 6.3.1. PUSH Message in the RFC. I'm puzzled about the text concerning the compression. I'll layout my concerns (without providing alternative text) and then point out some factual mistakes (I think) in this section.

RFC 3597 section 4 states that only well-known RRs may be compressed, those are the RRs from 1035.

Specifically:

> Future specifications for new RR types that contain domain names
> within their RDATA MUST NOT allow the use of name compression for
> those names, and SHOULD explicitly state that the embedded domain
> names MUST NOT be compressed.

> As noted in [RFC1123], the owner name of an RR is always eligible for compression.

Taking this in consideration and the DSO is defining new "RR" types, where other RRs are the RDATA, the case can be made no compression should be allowed. RFC 6762 doesn't says it's updating RFC 3597, so that can not apply here.

Taking this to the extreme I believe no compression should be allowed for any of these news types.

Some factial observations:

The list of RR that require compressions is not correct:

> NS, CNAME, PTR, DNAME, SOA, MX, AFSDB, RT, KX, RP, PX, SRV, NSEC

DNAME, NSEC should not be in this list. SRV is doubtful, but I guess also not.


And later, and more worrying:

> Servers may generate PUSH messages up to a maximum DNS message length of 16,382 bytes,

A DNS message has 16 bits for the length, giving it 64K of maximum length, compression is only up to 14 bits I think the two are confused here.

Kind Regards,
Miek GIeben

Report New Errata



Advanced Search