It is not worth making policy qualifiers more general, and if I never had to 
think about either of the existing ones ever again it would make me very happy. 
 Neither of them is particularly well specified or useful.

-Tim

From: COSE <[email protected]> On Behalf Of Russ Housley
Sent: Wednesday, April 12, 2023 11:55 AM
To: Joel Höglund <[email protected]>
Cc: cose <[email protected]>
Subject: Re: [COSE] Opinions on C509 / cbor-encoded-cert open issues

Joel:

I have argued against the use of policy qualifiers since the original work that 
lead to RFC 2459.  However, there are always a few people that want to keep 
them.  So, they never get deprecated.  I support anything that makes the use of 
policy qualifiers more painful.

Russ



On Apr 12, 2023, at 7:59 AM, Joel Höglund 
<[email protected]<mailto:[email protected]>> wrote:

Hi all!

We are working on the remaining open issues regarding 
draft-ietf-cose-cbor-encoded-cert. Below are two areas in which we are 
interested in hearing the opinions of the community before making decisions on 
which updates to make:

1. The encoding of the NameConstraints extension, applicable only for CA 
certificates (originally defined in 
https://www.rfc-editor.org/rfc/rfc5280#section-4.2.1.10)

NameConstraint contains GeneralSubtree, which in turn, in the current version 
of the draft is defined as:


GeneralSubtree = [ GeneralName, minimum: uint, ? maximum: uint ]

Due to current limitations of how GeneralName is encoded, we need to either 
update the GeneralName encoding (making it slightly longer). Alternatively, 
limit the scope of the draft to certificates that don’t contain explicit 
min-max specifications.
- Is the usage of min-max considered an important aspect which needs to be 
supported, or an acceptable restriction to impose?

2. Limitations regarding the certificate policies extension; the current 
version of the draft states the following restrictions:


“If noticeRef is not used and any explicitText are encoded as UTF8String, the 
extension value can be CBOR encoded.”
With the following CDDL:


PolicyIdentifier = int / ~oid
PolicyQualifierInfo = (
  policyQualifierId: int / ~oid,
  qualifier: text,
)
CertificatePolicies = [
  + ( PolicyIdentifier, ? [ + PolicyQualifierInfo ] )
]

The RFC5280 defines the two PolicyQualifierIds CPS Pointer and User Notice 
(https://www.rfc-editor.org/rfc/rfc5280#section-4.2.1.4).

The question to the community becomes:
- Are you aware of other PolicyQualifierInfo used, which include their own 
non-text qualifiers? That is, is it worth the effort to make the qualifier 
encoding more general, dependent on the policyQualifierId type.
A related issue is that there exist certificates using text types other than 
the recommended UTF8String encoding for explicitText. This we suggest accepting 
as a limitation, but are interested in hearing if there are other opinions.

The current latest version 05 can be found here:
https://www.ietf.org/archive/id/draft-ietf-cose-cbor-encoded-cert-05.html

Looking forward to your input!

Joel Höglund

_______________________________________________
COSE mailing list
[email protected]<mailto:[email protected]>
https://www.ietf.org/mailman/listinfo/cose

_______________________________________________
COSE mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/cose

Reply via email to