ietf-smime
[Top] [All Lists]

RE: cmsalg-02 RSA OID Proposal

2001-09-05 16:06:32

I don't disagree with your comments, John -- the only thing I was trying to
point out is that there was a precedent in the original design that we chose
to ignore for DSA, and just trying to justify the original separation the
identification of "digest" and "the other thing that isn't the digest but
which takes as input the digest and the public key and maybe other stuff in
order to make sure the signature is valid" (which I think would make a lousy
name for this field in the ASN.1, but I don't have a good substitute).

If all we're trying to do is change it to satisfy our own guilt that this
inconsistency exists, I might suggest that we simply get over it and put a
big apologetic paragraph in here that indicates our remorse, and not risk
breaking implementations with this change.

Blake

-----Original Message-----
From: Pawling, John [mailto:John(_dot_)Pawling(_at_)GetronicsGov(_dot_)com]
Sent: Wednesday, September 05, 2001 2:11 PM
To: ietf-smime(_at_)imc(_dot_)org
Subject: RE: cmsalg-02 RSA OID Proposal



Blake,

I agree with your point that the RFC 2630 specification of
signatureAlgorithm OIDs is inconsistent (i.e. id-dsa-with-sha1 is
inconsistent with rsaEncryption).  However, I disagree with your statement
that id-dsa-with-sha1 doesn't work as an indicator of "what identifier from
a certificate's SubjectPublicKeyInfo would be required to verify this
signature".  It is straightforward to develop CMS implementations to
recognize that the presence of the id-dsa-with-sha1 OID in the signedData
signerInfo signatureAlgorithm field indicates that the certificate required
to verify the signature must contain the id-dsa OID in the
SubjectPublicKeyInfo algorithm field.  Similarly, it is straightforward to
develop CMS implementations to recognize that the presence of either the
md5WithRSAEncryption or sha1WithRSAEncryption OID in the signedData
signerInfo signatureAlgorithm field indicates that the certificate required
to verify the signature must contain the rsaEncryption OID in the
SubjectPublicKeyInfo algorithm field.

===========================================
John Pawling, John(_dot_)Pawling(_at_)GetronicsGov(_dot_)com
Getronics Government Solutions, LLC
===========================================