[Servercert-wg] Proposal to Incorporate Mozilla's CRL Revocation Reason Code Requirements into the BRs

Ben Wilson bwilson at mozilla.com
Wed Sep 21 04:15:58 UTC 2022


Hi Corey,

See responses below.

On Wed, Sep 14, 2022 at 11:38 AM Corey Bonnell <Corey.Bonnell at digicert.com>
wrote:

> Hi Ben,
>
> It appears the ballot text has potential divergences from the published
> MRSP:
>
>
>
1. This ballot prohibits other CRLReasons from appearing in CRLs. This is
> meaningfully different from MRSP, where the new requirements are applicable
> solely to revocations that occur on or after the effective date.
>

 I think this can be fixed with some language changes.

2. There is no requirement to document reason codes in the Subscriber
> Agreement, whereas there is in MRSP. Is this change intentional?
>

Not exactly an intentional elimination of the requirement, but I can make
the ballot consistent with the MRSP with some language changes as well. My
idea was to suggest that CAs could incorporate the necessary information
"by reference" so that the CRL reason code explanations wouldn't have to
appear fully in Subscriber Agreements or Terms of Use.


> 3. Regarding 24-hour revocation reason #5: it appears that
> privilegeWithdrawn is now allowed. According to MRSP, only superseded is
> appropriate for this case.
>

For consistency, I'll change this to superseded only.


> 4. Regarding 5-day revocation reason #9: this is not a scenario listed in
> MRSP. In other words, this revocation scenario must be denoted as
> “unspecified” as the CRLReason under MRSP. Therefore, it is not possible to
> satisfy both the proposed BR text and MRSP.
>

That's probably the approach to take - thanks.  Another possibility is to
move this revocation reason down to 4.9.1.2 - CAs should revoke the
intermediate CA certificate(s) rather than all end entity certificates.

5.  Regarding 5-day revocation reason #10: this appears to be like scenario
> #7, but it is different in that revocation may be required even if there’s
> no violation of the CP/CPS. I don’t think this scenario is enumerated in
> MRSP, so it is not possible to specify a reason code that satisfies both
> MRSP and this ballot for this scenario.
>

Kathleen and I think that this reason is in the MRSP under the section for
the superseded CRLReason -  "the CA operator has revoked the certificate
for compliance reasons such as the certificate does not comply with this
policy, the CA/Browser Forum's Baseline Requirements, or the CA operator’s
CP or CPS".


>
> More generally, the Defined Term “Certificate” should be used throughout
> the ballot for consistency.
>

Agreed.  Thanks.


>
>
> Thanks,
>
> Corey
>

Thanks,
Ben

>
>
> *From:* Servercert-wg <servercert-wg-bounces at cabforum.org> *On Behalf Of *Ben
> Wilson via Servercert-wg
> *Sent:* Tuesday, September 13, 2022 11:37 PM
> *To:* Ben Wilson <bwilson at mozilla.com>; CA/B Forum Server Certificate WG
> Public Discussion List <servercert-wg at cabforum.org>
> *Subject:* Re: [Servercert-wg] Proposal to Incorporate Mozilla's CRL
> Revocation Reason Code Requirements into the BRs
>
>
>
> Here is the most current comparison:
>
>
>
>
> https://github.com/cabforum/servercert/compare/bbca71465ed8a8a76383086039f52c750009286a..1699612e5157423f607d67cc8ab9dc3a1d52b318
>
>
>
> Ben
>
>
>
> On Mon, Sep 12, 2022 at 11:00 AM Ben Wilson <bwilson at mozilla.com> wrote:
>
> Here is another edit that tries to make minimal changes to BR section
> 4.9.1.1.
>
>
> <http://goog_144053405>
>
>
> https://github.com/BenWilson-Mozilla/servercert/commit/94a07d08855cf489a2bdddff7d8a9490969d5d06
>
>
>
> Ben
>
>
>
> On Mon, Sep 12, 2022 at 9:51 AM Ben Wilson via Servercert-wg <
> servercert-wg at cabforum.org> wrote:
>
> Thanks, Dimitris. I'll work on that approach and get something back to you
> soon.
>
> Ben
>
>
>
> On Mon, Sep 12, 2022 at 2:56 AM Dimitris Zacharopoulos (HARICA) <
> dzacharo at harica.gr> wrote:
>
> Hi Ben,
>
> After a quick reading, I noticed that the subsections are not symmetrical
> and a bit inconsistent. For example, some of them contain the statement
> "the CA SHOULD revoke a certificate within 24 hours and MUST revoke a
> Certificate within 5 days", some do not.
>
> Other examples:
>
>    - 4.9.1.1.1, is labeled "Subscriber-Requested Revocation", however
>    there are other subsections that are also "Subscriber-Requested". This
>    separation seems confusing.
>    - 4.9.1.1.4 is about unreliable validation but most of the remaining
>    subsections are titled after the RFC 5280 revocation reasons.
>
> Finally, it's not very clear when the "unspecified (0)" reason must be
> used because of section 4.9.1.1.8 (Other Circumstances) which doesn't point
> to a revocation reason.
>
> From my perspective, I'm not sure if breaking down each subsection is more
> helpful for reading the revocation requirements than the current listing. I
> understand there is a desire to copy the MRSP language as much as possible
> but perhaps we need to consider a less "intrusive" set of changes to a
> section that CAs already have a difficult time reading and implementing.
>
> IMO we either need to describe the revocation scenario and point to the
> RFC 5280 revocation reason (closer to what the BRs have today), or start
> with the RFC 5280 revocation reasons and enumerate the revocation scenarios
> (closer to what MRSP has today). I find it confusing to mix the two
> approaches.
>
>
> Thanks,
> Dimitris.
>
> On 12/9/2022 6:32 π.μ., Ben Wilson wrote:
>
> For review - here is another proposal that takes BR section 4.9.1.1 and
> puts the 24-hour and 5-day revocation times into subsections that match the
> CRL reason codes.
>
>
>
>
> https://github.com/BenWilson-Mozilla/servercert/commit/b185a28fcc20d5853747e4506103823e3dc7c282
>
>
>
> Ben
>
>
>
> On Thu, Sep 8, 2022 at 12:05 PM Dimitris Zacharopoulos (HARICA) <
> dzacharo at harica.gr> wrote:
>
> Good point.
>
> s/
>
>
> *expected/shall use/*
>
> On 8/9/2022 8:26 μ.μ., Tim Hollebeek wrote:
>
> I would prefer standard 2119 language instead of an “expectation”.  There
> are no documented rules for what it means for a CRLReason to be expected to
> be a certain value.
>
>
>
> -Tim
>
>
>
> *From:* Servercert-wg <servercert-wg-bounces at cabforum.org>
> <servercert-wg-bounces at cabforum.org> *On Behalf Of *Dimitris
> Zacharopoulos (HARICA) via Servercert-wg
> *Sent:* Thursday, September 8, 2022 3:21 AM
> *To:* Ben Wilson <bwilson at mozilla.com> <bwilson at mozilla.com>; CA/B Forum
> Server Certificate WG Public Discussion List <servercert-wg at cabforum.org>
> <servercert-wg at cabforum.org>
> *Subject:* Re: [Servercert-wg] Proposal to Incorporate Mozilla's CRL
> Revocation Reason Code Requirements into the BRs
>
>
>
>
>
> On 7/9/2022 8:22 μ.μ., Ben Wilson wrote:
>
> Good suggestion. I can re-work a proposal that re-writes BR sec. 4.9.1.1
> to re-group the revocation reasons into the reason codes that should be
> used. Is that what you were thinking?
>
>
> Yes. We should also try to keep the current BRs prioritization. The
> section begins with the cases where the Certificate(s) need to be revoked
> within 24h and then moves to the 5-day revocation cases.
>
> We could walk this list down making sure that all Mozilla cases are listed
> (add the ones that are not) and add the expected revocationReason for each
> case. For example:
>
> *The CA SHALL revoke a Certificate within 24 hours if one or more of the
> following occurs:*
>
>    1. *The Subscriber requests in writing that the CA revoke the
>    Certificate (expected CRLReason:unspecified);*
>    2. *The Subscriber notifies the CA that the original certificate
>    request was not authorized and does not retroactively grant authorization
>    (expected CRLReason:privilegeWithdrawn);*
>    3. *The CA obtains evidence that the Subscriber's Private Key
>    corresponding to the Public Key in the Certificate suffered a Key
>    Compromise (expected CRLReason:keyCompromise);*
>    4. *The CA is made aware of a demonstrated or proven method that can
>    easily compute the Subscriber's Private Key based on the Public Key in the
>    Certificate (such as a Debian weak key, see *
>    *https://wiki.debian.org/SSLkeys* <https://wiki.debian.org/SSLkeys>*)
>    (expected CRLReason:keyCompromise);*
>    5. *The CA obtains evidence that the validation of domain
>    authorization or control for any Fully-Qualified Domain Name or IP address
>    in the Certificate should not be relied upon (expected CRLReason:
>    superseded).*
>
> and so on.
>
> Does that work?
>
> Dimitris.
>
> Thanks,
>
> Ben
>
>
>
> On Wed, Sep 7, 2022 at 6:01 AM Dimitris Zacharopoulos (HARICA) via
> Servercert-wg <servercert-wg at cabforum.org> wrote:
>
> Hi Ben,
>
> I believe the proposal, as written, causes confusion in regards to
> 4.9.1.1. Some of the reasons described in your proposal are already
> mentioned in 4.9.1.1. Perhaps we should work some more to "unify" the two
> sections.
>
> My proposal would be to update 4.9.1.1 and include the expected CRLReason
> after each case.
>
>
> Thoughts?
> Dimitris.
>
> On 6/9/2022 8:13 μ.μ., Ben Wilson via Servercert-wg wrote:
>
> All,
>
> I'm looking for one more endorser.
>
> Thanks,
>
> Ben
>
>
>
> On Fri, Jul 29, 2022 at 12:40 PM Ben Wilson via Servercert-wg <
> servercert-wg at cabforum.org> wrote:
>
> All,
>
>
>
> I have created a proposal in Github to incorporate Mozilla's CRL
> Revocation Reason Code requirements into the Baseline Requirements.
>
>
>
> See https://github.com/cabforum/servercert/issues/377
>
>
>
>
> https://github.com/BenWilson-Mozilla/servercert/commit/52a480803beff1f96d61c4b6d76570ac7adff4d5
>
>
>
> I'm looking for comments, suggestions, and two endorsers.
>
>
>
> Thanks,
>
>
>
> Ben
>
> _______________________________________________
> Servercert-wg mailing list
> Servercert-wg at cabforum.org
> https://lists.cabforum.org/mailman/listinfo/servercert-wg
>
>
>
> _______________________________________________
>
> Servercert-wg mailing list
>
> Servercert-wg at cabforum.org
>
> https://lists.cabforum.org/mailman/listinfo/servercert-wg
>
>
>
> _______________________________________________
> Servercert-wg mailing list
> Servercert-wg at cabforum.org
> https://lists.cabforum.org/mailman/listinfo/servercert-wg
>
>
>
>
>
>
>
> _______________________________________________
> Servercert-wg mailing list
> Servercert-wg at cabforum.org
> https://lists.cabforum.org/mailman/listinfo/servercert-wg
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cabforum.org/pipermail/servercert-wg/attachments/20220920/5e544001/attachment-0001.html>


More information about the Servercert-wg mailing list