Hi Corey,

Sincere apologies for missing this review! There has been an attention gap for 
some time.

Thanks for great comments. Please see responses inline below, with reference to 
PR#383:
https://github.com/cose-wg/CBOR-certificates/pull/383

For overview of the current state, here the Editor’s copy:
https://cose-wg.github.io/CBOR-certificates/draft-ietf-cose-cbor-encoded-cert.html


Thanks,
Göran


From: Corey Bonnell via Datatracker <[email protected]>
Date: Monday, 9 February 2026 at 19:46
To: [email protected] <[email protected]>
Cc: [email protected] <[email protected]>; 
[email protected] 
<[email protected]>; [email protected] 
<[email protected]>
Subject: draft-ietf-cose-cbor-encoded-cert-16 ietf last call Secdir review

Document: draft-ietf-cose-cbor-encoded-cert
Title: CBOR Encoded X.509 Certificates (C509 Certificates)
Reviewer: Corey Bonnell
Review result: Has Nits

Section 3.1.4
s/extension value/attribute value/
[GS: Right, fixed in the Editor’s copy]


Section 3.2.2
The ECDSA signature encoding process can reference RFC 9053, section 2.1.

[GS: Done]

Section 3.3

The ASN.1 definition of id-pkix-ocsp-nocheck indicates the value will always be
NULL, so "If the extension value is NULL" is not needed. Instead, the same text
used for "Precertificate Signing Certificate" can be used.

[GS: Done]

"Precertificate Signing Certificate" is not the name of the extension, but is
rather a certificate for a signer of pre-certificates. I suggest using
"Precertificate Critical Poison" instead.

[GS: Certificate Transparency was not considered prioritized and has been 
removed from the draft.]

Section 8.

"The CBOR encoding of X.509 certificates does not change the security
assumptions needed when deploying standard X.509 certificates but decreases the
number of fields transmitted, which reduces the risk for implementation
errors." Is this true? The number of fields in a C509 certificate appears to be
the same as a X.509 certificate. If anything, the C509 specification appears to
be more complex than ASN.1-based X.509 due to the variable encoding of elements
in certain situations to minimize size.

[GS: The text is updated. Regarding complexity, there are known issues with 
ASN.1, but perhaps too early
to say if CBOR will be better (which the proponents of the draft believes). ]

"The gateway solution described in Section 6 requires unencrypted certificates
and is not recommended." I think this needs to be fleshed out, because it
assumes that certificates are secret information. In scenarios where
certificates are not secret, it is unclear whether this SHOULD NOT is relevant.

[GS: This relates to identity protection and privacy, a clarification is added.]

It would be good if there is an explicit MUST that certification path
processing (as defined in RFC 5280, section 6) be performed on C509
certificates before they can be considered trusted.

[GS: Added.]

Section 9.13

The "Compressed subjectPublicKey" comment is potentially confusing, especially
for the EC Public Key types, as it might be read as a statement that compressed
point encoding is used for the coordinates.

[GS: Term removed and a reference to the relevant section is added.]

[GS: Do let us know if we did not appropriately address your concerns.

Again, thanks for your help!

Best regards,
Göran]

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

Reply via email to