* The TERENA SSL CA 3 subordinate has misissued a number of certificates [3], most of which are not revoked.
- We can revoke these. I have no issue remediating them. I didn’t realize
these were an ongoing concern.
* DigiCert’s response in this bug states “We were under the impression from
previous communications with Mozilla that certain types of errors identified
did not require certificate revocation. It would help if Mozilla could indicate
which certificate errors are believed to require revocation. We will then
review the lists to see which certificates need to be revoked.” I do not
believe that Mozilla should create such a list, and we have set a precedent for
requiring revocation for at least some of the errors that are identified - e.g.
metadata in subject fields [4].
- That’s fine, but the answer still stands as why we didn’t revoke them before.
Gerv told us we didn’t need to revoke them because they didn’t represent an
actual security concern. I can go find the email if that helps.
* In addition, DigiCert previously reported that they had addressed the problem
of metadata in subject fields for certificates issued by the Terena subordinate
[5].
- Yes. This should be addressed. Unless you are expecting them to be revoked
now?
* Linters identify a large number of misissued certificates under the DigiCert
SHA2 Secure Server CA intermediate [6]. Many of these are false positives (e.g.
ZLint expects CN and SAN fields to be lowercased), but some are not and of
those many are not revoked - e.g. [7].
Thanks a ton for the breakdown. We’ll get these taken care of where it’s not a
false positive. I think there are only 2-3 that are not false positives, with 2
not previously discussed.
Here’s the breakdown:
FATAL: x509: RSA modulus is not a positive number
Bad reading of the BRs. The BRs say the range should be between 2^16+1 and
2^256-1. The team implementing this thought saw SHOULD and thought it was
optional. They missed the sentence before which says the public exponent MUST
be equal to 3 or more. I’ll file and incident report on this.
FATAL: asn1: structure error: integer not minimally-encoded
I think this one is a false positive? It’s caused by padded zeros which aren’t
expressly prohibited. Want us to revoke these?
ERROR: The common name field in subscriber certificates must include only
names from the SAN extension
This one is a false positive
ERROR: DNSNames must have a valid TLD
This is a false positive
ERROR: The 'Organization Name' field of the subject MUST be less than 64
characters
This is one of the issues mentioned above. We can revoke all of these. We
didn’t know Mozilla was waiting on that.
ERROR: CAs MUST NOT issue subscriber certificates with validity periods longer
than 39 months regardless of circumstance.
False positive
ERROR: The country name field MUST contain the two-letter ISO code for the
country or XX
Grr.. I dislike this one (XK was used instead of XX). Although recognized as a
temporary code for Kosovo, technically XK is not an official ISO code so it
violates the strict letter of the law. We’ll revoke these.
ERROR: Subject name fields must not contain '.','-',' ' or any other indication
that the field has been omitted
False positive. The BRs say “All other optional attributes, when present within
the subject field, MUST contain information that has been verified by the CA.
Optional attributes MUST NOT contain metadata such as ‘.’, ‘-‘, and ‘ ‘ (i.e.
space) characters, and/or any other indication that the value is absent,
incomplete, or not applicable.” With the strict wording policy that seems in
effect, organizationalUnit is not “All other optional attributes”. It’s a
defined attribute and thus cannot be “other”.
ERROR: Explicit text has a maximum size of 200 characters
False positive I think….
ERROR: Subscriber Certificates issued after 1 July 2016 but prior to 1 March
2018 MUST have a Validity Period no greater than 39 months.
False positive
ERROR: Subscriber Certificate: subject:localityName MUST appear if
subject:organizationName, subject:givenName, or subject:surname fields are
present but the subject:stateOrProvinceName field is absent.
Wow. I hadn’t seen this. It’ll be revoked and we’ll look at what happened.
* CPS section 3.2.2 did not, in my opinion, adequately specify the procedures
employed to perform email address verification as required by Mozilla policy
section 2.2(2). The latest update addressed this.
- This is an issue related to the fact there is no policy on s/MIME. We updated
it with more specificity, of course, but I’d love to see a real s/MIME policy.
Thanks Wayne. I can confirm we will revoke all mis-issued certs.
From: Wayne Thayer <[email protected]>
Sent: Thursday, December 13, 2018 5:34 PM
To: Jeremy Rowley <[email protected]>
Cc: Ryan Sleevi <[email protected]>; mozilla-dev-security-policy
<[email protected]>
Subject: Re: DigiCert Assured ID Root CA and Global Root CA EV Request
My main concern with this request is the misissued certificates identified by
linters that have not been revoked - I have included my comments and links from
the original message below.
It appears that DigiCert is not planning to remediate these certificates - can
a representative from DigiCert confirm that?
If these certificates are not revoked, I feel that it would be consistent with
our treatment of other CAs to deny this request. I would appreciate everyone's
opinion on that, and also if you think that the amount of misissuance is reason
enough to deny this request, even if the misissuance is remediated.
=============================
[3] https://crt.sh/?caid=1687
<https://crt.sh/?caid=1687&opt=cablint,zlint,x509lint&minNotBefore=2017-01-01>
&opt=cablint,zlint,x509lint&minNotBefore=2017-01-01
[4] https://crt.sh/?id=629259396 <https://crt.sh/?id=629259396&opt=cablint>
&opt=cablint
[5] https://bugzilla.mozilla.org/show_bug.cgi?id=1397958
[6]https://crt.sh/?caid=1191
<https://crt.sh/?caid=1191&opt=cablint,zlint,x509lint&minNotBefore=2017-01-01>
&opt=cablint,zlint,x509lint&minNotBefore=2017-01-01
[7] https://crt.sh/?id=286404787 <https://crt.sh/?id=286404787&opt=zlint>
&opt=zlint
=============================
On Thu, Nov 29, 2018 at 4:17 PM Wayne Thayer <[email protected]
<mailto:[email protected]> > wrote:
I would appreciate it if we could move the discussion of exceptions to the
deadline for revoking certificates containing underscores to a new thread.
As it relates to this request, any failure to meet the revocation deadline
would trigger the creation of an incident bug. (that is unless we as a
community decide otherwise)
I am not of the opinion that the existence of such a bug would change the
outcome of this discussion. If others feel that it might, I am not opposed to
holding the discussion open. Meanwhile, i'd suggest we stick to discussing the
current facts of this request.
- Wayne
smime.p7s
Description: S/MIME cryptographic signature
_______________________________________________ dev-security-policy mailing list [email protected] https://lists.mozilla.org/listinfo/dev-security-policy
