Thank you, Joel!

I can confirm that with this change the signature does indeed validate in
my code tests.

-derek

On Mon, April 25, 2022 1:32 pm, Joel Höglund wrote:
> A big thank you Derek, for your observation of the non validating
> signature!
>
> The source of the bug was as mundane as a copy-paste error between draft
> versions quite a while back, that obviously had gone unnoticed until now.
> A
> correct signature for the current sample native C509 certificate is:
>
> R = 4EFCEE214E2F3EB94BBDAE43C1948D144140233FC52FAC37428ED4BD8A050C5D
> S = F980DF51B71C70D32B3E92D56A6C7AB0D2671B2B865A35B2E2CA421D44C5C93F
>
> <=>
>
> 0,
> h'01f50d',
> "RFC test CA",
> 1577836800,
> 1612224000,
> h'0123456789AB',
> 1,
> h'02B1216AB96E5B3B3340F5BDF02E693F16213A04525ED44450
>   B1019C2DFD3838AB',
> 1,
> 0,
> h'4EFCEE214E2F3EB94BBDAE43C1948D144140233FC52FAC3742
>   8ED4BD8A050C5DF980DF51B71C70D32B3E92D56A6C7AB0D267
>   1B2B865A35B2E2CA421D44C5C93F'
>
> I apologize to you and anyone who spent time on this, updates with
> corrections will soon be on their way. And should you find anything else
> that seems strange we are very interested to know.
>
> Best Regards
>
> Joel Höglund
>
> On Mon, 25 Apr 2022 at 14:38, Derek Atkins <[email protected]> wrote:
>
>> Hi All,
>>
>> I am working on an implementation of C509 based on
>> draft-ietf-cose-cbor-encoded-cert-03 and I wanted to report the issues
>> I've found.
>>
>> First, there is a bug in the draft.  In Figure 7, C509 Attributes in
>> section 11.2, there are two entries with tag 21, the Inc. Country and
>> the
>> Domain Component.  I suspect DC should be '22'?
>>
>> The second issue I'm having is that the certificate example in Appendix
>> A.1.2 does not validate using the keys in Appendix A.1.3.   I have
>> verified that the isserPublicKey is valid, and it's exactly what I
>> create
>> when making the public key from the issuerPrivateKey.  However, when I
>> attempt to validate the A.1.2 certificate using the issuerPublicKey the
>> signature is not valid.
>>
>> I am computing the SHA256 hash of the TBSCertificate sequence as
>> b5bca215e1d1478d2fe7728a54089f2032a4a1a245fafb5bd21d9eeb9d076aed --
>> however I have no way to validate if this is correct.
>>
>> Has anyone written a C509 parser and successfully validated this
>> certificate?  If so, could you report what hash value you get from the
>> TBSCertificate sequence?
>>
>> Thanks,
>>
>> -derek
>>
>> PS: I reported this directly to the authors last week, but figured I
>> would
>> reach out to a wider audience for validation of my issue(s).
>>
>> --
>>        Derek Atkins                 617-623-3745
>>        [email protected]             www.ihtfp.com
>>        Computer and Internet Security Consultant
>>
>> _______________________________________________
>> COSE mailing list
>> [email protected]
>> https://www.ietf.org/mailman/listinfo/cose
>>
>


-- 
       Derek Atkins                 617-623-3745
       [email protected]             www.ihtfp.com
       Computer and Internet Security Consultant

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

Reply via email to