On Mon, Oct 12, 2026 at 10:03 AM <[email protected]> wrote:

> The following errata report has been submitted for RFC7901,
> "CHAIN Query Requests in DNS"
>
> --------------------------------------
> You may review the report below and at:
> https://errata.rfc-editor.org/eid9214/
>
> --------------------------------------
> Type: Technical
> Reported by: Ensar Sarajčić <[email protected]>
>
> Section 8.1 says:
>
> Original Text
> -------------
> o  The upstream recursive resolver adds the CHAIN option to the DNS
>       response as follows:
>
>       *  Option-code, set to 13
>
>       *  Option-length, set to 5
>
>       *  The closest trust point is set to "com." (0x03 0x63 0x6f 0x6d
>          0x00)
>
> Corrected Text
> --------------
> o  The upstream recursive resolver adds the CHAIN option to the DNS
>       response as follows:
>
>       *  Option-code, set to 13
>
>       *  Option-length, set to 13
>
>       *  The closest trust point is set to "example.com." (0x07 0x65 0x78
> 0x61 0x6d 0x70 0x6c 0x65 0x03 0x63 0x6f 0x6d 0x00)
>
> Notes
> -----
> The provided example seems inconsistent with:
> https://datatracker.ietf.org/doc/html/rfc7901#section-5.4
> Specifically:
>    If a CHAIN answer would be bigger than the recursive resolver is
>    willing to serve, it SHOULD send a partial chain starting with the
>    data closest to the top of the chain.  The client MAY resend the
>    query with an updated closest trust point until it has received the
> full chain.
>    The CHAIN response will contain the lowest closest trust
>    point that was included in the CHAIN answer.
>
> I may be confusing the wording here or misunderstanding things, but I
> understood that the lowest closest trust point in the example response is "
> example.com." and not "com.". This is especially important to enable
> clients to use the CHAIN option value instead of checking records in the
> authority section.
>
>
No this is incorrect. Think of it this way. It sends the part of the chain
that builds on top of the chain the requester already indicated it knows,
and extends it. If it doesn't fit, the requester now has a longer verified
chain (even if incomplete) but it can on the 2nd query extend its known
trust trust path.

The requester is basically indicating what DNSSEC records from the root
down it already has, that does not need to be included in the CHAIN.

I agree that "top" here is confusing, but think of it as "top level
domain", as in towards the root.

So lets say you are asking for www.basement.toronto.nohats.ca. (and all
dots are zone cuts with DS records), and you already have the DS for ca in
your cache, then you signal passing "ca" as trust point, so the chain
starts at the DS for nohats.ca. If the entre chain doesn't fit, you build
it from this DS further down. So you try to see if the DS of
toronto.nohats.ca. fits, and if so, you then try to see if the DS for
basement.toronto.nohats.ca. fits. Let's say this doesn't  fit, so you stop
adding to the chain. . Now the requester can ask the exact same question as
before, but sends the trustpoint as "toronto.nohats.ca". Now the chain
starts lower and you could fit the DS for basement.toronto.nohats.ca.

Paul
_______________________________________________
DNSOP mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to