Michael Richardson wrote:
"I think that this means that there is CBOR definite(?) array of bytes."

COSE_X509 and COSE_C509 are both defines as

bstr / [ 2*bstr ]

which means they they are either a single CBOR byte string or an CBOR array of 
CBOR bytes strings.

While Section 9 of RFC 9052 requires that "Encoding MUST be done using definite 
lengths, and the length of the (encoded) argument MUST be the minimum possible 
length" and draft-ietf-cose-cbor-encoded-cert states that "When referring to 
CBOR, this specification always refers to Deterministically Encoded CBOR as 
specified in Sections 4.2.1 and 4.2.2 of [RFC8949], I cannot find any text in 
RFC 9360 restricting CBOR to deterministic CBOR.

- I think you could read RFC 9360 as indefinate arrays and non-minumum length 
encodings are allowed.
- I think you could also see the lack of deterministic encoding requirements in 
RFC 9360 as an errata.

Cheers,
John

From: Michael Richardson <[email protected]>
Date: Friday, 3 October 2025 at 18:29
To: Thomas Fossati <[email protected]>
Cc: cose <[email protected]>
Subject: [COSE] Re: extending application/cose-x509

Thomas Fossati <[email protected]> wrote:
    > We want to transport DICE [0] certificate chains in CMWs [1], and for
    > that, we need a media type.

    > Note that DICE certificate chains differ semantically from standard
    > X.509 certificate chains in that they also represent attestation
    > Evidence [2].  Therefore, using
    > * application/pkcs7-mime; smime-type="certs-only"
    > * application/cose-x509; usage=chain, and
    > * application/pkix-pkipath
    > would provide too coarse typing information, so we'd like to improve this.

    > One way would be to extend the application/cose-x509 "usage" parameter
    > to include the value "dice-chain", i.e., application/cose-x509;
    > usage=dice-chain.

cose-x509.  I was thinking this is from cbor-encoded-cert, but it defines 
cose-c509-cert.
And that definition has usage=chain, so was this a typo?  NOPE.
cose-x509 is RFC9360... and COSE_X509 is a CBOR sequence of bstr wrapped
DER-encoded PKIX certificates.
I think that this means that there is CBOR definite(?) array of bytes.

So this becomes a dice-chain.
And after you do CoAP/Content-Format registration, you get an integer for the
CBOR CMW, so any verbosity of the media type is a moot point.

    > Would that be acceptable?  If so, what steps need to be taken to
    > register the new parameter value?
    > Do we need a specification, and if so, what kind? Or is a request to
    > the media-types list sufficient?

I understand that an email to [email protected] with the template is
enough.  However, I find that one has to poke the reviewers.
I'm hoping IANA's new DE RT system will get help..

--
Michael Richardson <[email protected]>   . o O ( IPv6 IøT consulting )
           Sandelman Software Works Inc, Ottawa and Worldwide
_______________________________________________
COSE mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to