[IETF-PKIX] PKIX Part 1: more clarity needed on AltNames

[Paul Hoffman / IMC <phoffman@xxxxxxx>]

issuerAltName and subjectAltName seem squishy to me. In reading the spec, I
could not determine what to do if I was looking at a cert that had various
combinations of non-AltNames and AltNames. I don't believe that this is a
fatal flaw in the design, but I am sure we need more wording in the draft
to explain what to do in various cases.

In section 4.1.2.4, it says "The issuer identity may be carried in the
issuer name field and/or the issuerAltName extension." The same kind of
wording appears in 4.1.2.6 about subject names. There is a description of
how to put something in just issuerName or in just issuerAltName. However,
nowhere does it say how to interpret a certificate that has names in both
fields.

If I have a DN in the issuerName and an rfc822Name in the issuerAltName,
does it mean that either name is definitive? Or does it mean that only the
combination of the two names is definitive?

What does it mean if the subjectName is filled in and there is a
subjectAltName with the criticality bit set? Again, does this mean either
is definitive, or the combination is definitive?

If the answer to the above is "only the combination", what do you do with a
cert that has two AltNames, one that is set critical and one that isn't? If
the answer to the above is "either is definitive", does that answer still
hold true if the AltName is marked critical?

This also affects the name chain checking in section 6.

Again, I believe that the spec probably works fine, but the interpretation
of more cases need to spelled out.


--Paul Hoffman, Director
--Internet Mail Consortium