Re: Policy 2.4 Proposal: Update required version number of Baseline Requirements to 1.3.7
On Wed, Jan 18, 2017 at 7:16 AM, Gervase Markhamwrote: > On 17/01/17 23:32, Ryan Sleevi wrote: > > BRs 1.3.0 ( https://cabforum.org/wp-content/uploads/CAB-Forum-BR- > 1.3.0.pdf > > ) already include the clause (in Section 2.2) that: > > "The CA SHALL publicly give effect to these Requirements and represent > > that it will adhere to the latest published version." > > Hmm. I was not aware of that. I wonder how many CAs are aware that > according to the BRs, any changes to the BRs by default come in > immediately the motion is passed and the document is updated. Perhaps > I'm the only person who didn't know this. > I am surprised, since we discussed it during the Scottsdale F2F in the CA/Browser Forum :) See https://cabforum.org/2016/02/17/2016-02-17-minutes-of-f2f-meeting-37/#Compliance-Assessment-Coordination-with-auditors-and-browsers "Some CAs say they were audited to an older version of the BRs; therefore it would appear that the CA only thinks the need to comply with the older version of the BRs. This is not correct as the BRs state that the CA has to have a CPS statement that they comply with the latest version of the BRs." > You'll need to give me a more specific reference; I don't remember any > such question, and a quick scan back through top-level posts from Apple > employees hasn't revealed it. > There's not some single top-level post that states this - but it was the whole purpose of Ballots 180/181, which is the view that the Forum did not follow its Bylaws (with respect to either voting vs IP review period or with respect to the formation of a PAG following disclosures, going back to the first adoption of the BRs) therefore invalidates the adoption. Which is why we had Ballots 180/181 - to ensure that all members were happy that 'the process' was followed and all documents produced were done so in a way consistent with the Bylaws. It's procedural administrivia, certainly, and not one anyone advanced until Apple raised concerns, largely in part do to the proposed ways of resolving the lack of disclosure notices being provided during much of Dean's chairing. > So the suggestion is that we just update our policy to require adherence > to the latest version of the BRs, on the basis that this is what the BRs > require anyway? Yes. That gets you to Ballot 181 / v 1.4.2. However, you would then need to clarify (at least for 1.4.2) that the only acceptable forms of any other are the 169 methods. ___ dev-security-policy mailing list dev-security-policy@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-security-policy
Re: Policy 2.4 Proposal: Update required version number of Baseline Requirements to 1.3.7
On 17/01/17 23:32, Ryan Sleevi wrote: > BRs 1.3.0 ( https://cabforum.org/wp-content/uploads/CAB-Forum-BR-1.3.0.pdf > ) already include the clause (in Section 2.2) that: > "The CA SHALL publicly give effect to these Requirements and represent > that it will adhere to the latest published version." Hmm. I was not aware of that. I wonder how many CAs are aware that according to the BRs, any changes to the BRs by default come in immediately the motion is passed and the document is updated. Perhaps I'm the only person who didn't know this. > So despite Mozilla Policy requiring 1.3.0, up until the passage of > Ballots 180/181, CAs were already on the hook and expected to comply > with the BRs 1.4.1 - meaning implementing the methods of Ballot 169 by > 1 March 2017. > > Up until the questions by Apple in the Forum, there had not been any > debate or disagreement about what the 'latest published version' was, > either within the Forum or within mozilla.dev.security.policy. You'll need to give me a more specific reference; I don't remember any such question, and a quick scan back through top-level posts from Apple employees hasn't revealed it. > It's > unclear whether Mozilla shares Apple's interpretation about the > legitimacy of all versions of the BRs prior to 1.4.2, but based on > your replies, it does not seem you agree on substance. That is, BRs > 1.3.0 were/are a valid version of the BRs, at least within the spirit > and intent of the Mozilla policy, and so too by that logic are > versions 1.3.7, 1.4.1, and 1.4.2, which were passed in the same > manner. Yes, that's my view. > and it > seems that the reference to a specific version in Policy 2.4 is > perhaps superflous, so long as the Section 2.2 remains in force in the > BRs. So the suggestion is that we just update our policy to require adherence to the latest version of the BRs, on the basis that this is what the BRs require anyway? Gerv ___ dev-security-policy mailing list dev-security-policy@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-security-policy
Re: Policy 2.4 Proposal: Update required version number of Baseline Requirements to 1.3.7
On Mon, Jan 16, 2017 at 3:30 AM, Gervase Markhamwrote: > On 13/01/17 01:56, Ryan Sleevi wrote: >> Notably, 1.3.7 also has IP encumbrances - and uncertainty - the same >> as 1.4.1, so presumably, Mozilla is OK with having encumbered methods >> included. Considering some of these exclusions have existed since the >> BR's adoption, that doesn't seem an unreasonable conclusion. > > This is a fair point, to a degree. IP issues with the documented methods > are not avoided just because they are documented in a document which > existed before the disclosures were finally properly filed. > >> So what's the difference between 1.3.7 and 1.4.1? >> - A few methods were introduced which may or may not be encumbered >> - The ability for the CA to select anything they want to argue is >> equivalent is removed >> >> I would presume that your contention with 139 is not because new >> methods were (potentially) encumbered, but on the basis that it >> removes the any other method. > > Yes, I think that's more it. I was nervous about removing that "escape > hatch" for CAs before the IPR questions were more fully settled. > >> Given that there are other methods >> available (as reaffirmed by Ballot 181) that have no encumbrances by >> CA/B Forum members, and given that potentially *any or all* of the >> methods used may be encumbered by non-Forum members (who have no >> obligation to disclose), it does not seem that it creates any new >> meaningful risk for Mozilla to impose 1.4.1 upon CAs. > > It's true that the passing of 181, which had not happened when I > prepared my original thoughts for this Github issue, does clarify > matters. And it includes at least one 'unencumbered' automatable method. > >> Given that you can always revisit it if a CA can provide demonstrable >> evidence of concern and proposed alternatives, without waiting on the >> CA/Browser Forum, I'd like to encourage you that 1.4.1 is no worse >> than 1.3.7, either technically or from an IP encumbrance perspective, >> but is significantly and substantially better. > > The GoDaddy incident is relevant, I agree. > > OK, sold. Many recent changes to the BRs have been > additionally-permissive, and so don't require a lead time for CAs to > adopt. The most recent change which is not so is ballot 174, which > changed the requirements about what CAs must put in their CPs and CPSes > when their issuance practices conflict with the BRs due to local law. > That passed on 29th August 2016, 4.5 months ago, and led to version > 1.3.9. By the time we ship policy 2.4, I expect it'll be nearly 6 months. > > So we could move to 1.3.7, 1.4.1, or 1.4.2. So the real policy question > here is: > > "Does Mozilla want to leave the 'any other method' option open to CAs > and, if so, for how long?" > > If the answer is no, we should adopt 1.4.1 and require strict compliance > (i.e. you can't actually use 1.4.2 or later) for section 3.2.2.4, > although we might allow any later version for all other sections. > > If the answer is yes, we should adopt 1.4.2, and permit compliance to > later versions. When we decide to stop permitting "any other method", we > either update the version number to a later version of the BRs if it has > that option removed by then, or we simply state that we don't allow it > any more. > > Google's policy, AIUI, is that they are requiring strict 10-method > compliance by the original effective date of ballot 169, which is 1st > March 2017. I suspect we won't ship CA Policy 2.4 before that, although > I'd like to ship soon after. Because of Google's requirement, I expect > most CAs will be working towards that date anyway. So we could probably > piggy-back on that. > > Comments? It depends on what your view is with respect to the 'minimum' version that is required under the existing Mozilla Policy. BRs 1.3.0 ( https://cabforum.org/wp-content/uploads/CAB-Forum-BR-1.3.0.pdf ) already include the clause (in Section 2.2) that: "The CA SHALL publicly give effect to these Requirements and represent that it will adhere to the latest published version." So despite Mozilla Policy requiring 1.3.0, up until the passage of Ballots 180/181, CAs were already on the hook and expected to comply with the BRs 1.4.1 - meaning implementing the methods of Ballot 169 by 1 March 2017. Up until the questions by Apple in the Forum, there had not been any debate or disagreement about what the 'latest published version' was, either within the Forum or within mozilla.dev.security.policy. It's unclear whether Mozilla shares Apple's interpretation about the legitimacy of all versions of the BRs prior to 1.4.2, but based on your replies, it does not seem you agree on substance. That is, BRs 1.3.0 were/are a valid version of the BRs, at least within the spirit and intent of the Mozilla policy, and so too by that logic are versions 1.3.7, 1.4.1, and 1.4.2, which were passed in the same manner. I mention all of this to highlight that, up until the passage of Ballots
Re: Policy 2.4 Proposal: Update required version number of Baseline Requirements to 1.3.7
On 13/01/17 01:56, Ryan Sleevi wrote: > Notably, 1.3.7 also has IP encumbrances - and uncertainty - the same > as 1.4.1, so presumably, Mozilla is OK with having encumbered methods > included. Considering some of these exclusions have existed since the > BR's adoption, that doesn't seem an unreasonable conclusion. This is a fair point, to a degree. IP issues with the documented methods are not avoided just because they are documented in a document which existed before the disclosures were finally properly filed. > So what's the difference between 1.3.7 and 1.4.1? > - A few methods were introduced which may or may not be encumbered > - The ability for the CA to select anything they want to argue is > equivalent is removed > > I would presume that your contention with 139 is not because new > methods were (potentially) encumbered, but on the basis that it > removes the any other method. Yes, I think that's more it. I was nervous about removing that "escape hatch" for CAs before the IPR questions were more fully settled. > Given that there are other methods > available (as reaffirmed by Ballot 181) that have no encumbrances by > CA/B Forum members, and given that potentially *any or all* of the > methods used may be encumbered by non-Forum members (who have no > obligation to disclose), it does not seem that it creates any new > meaningful risk for Mozilla to impose 1.4.1 upon CAs. It's true that the passing of 181, which had not happened when I prepared my original thoughts for this Github issue, does clarify matters. And it includes at least one 'unencumbered' automatable method. > Given that you can always revisit it if a CA can provide demonstrable > evidence of concern and proposed alternatives, without waiting on the > CA/Browser Forum, I'd like to encourage you that 1.4.1 is no worse > than 1.3.7, either technically or from an IP encumbrance perspective, > but is significantly and substantially better. The GoDaddy incident is relevant, I agree. OK, sold. Many recent changes to the BRs have been additionally-permissive, and so don't require a lead time for CAs to adopt. The most recent change which is not so is ballot 174, which changed the requirements about what CAs must put in their CPs and CPSes when their issuance practices conflict with the BRs due to local law. That passed on 29th August 2016, 4.5 months ago, and led to version 1.3.9. By the time we ship policy 2.4, I expect it'll be nearly 6 months. So we could move to 1.3.7, 1.4.1, or 1.4.2. So the real policy question here is: "Does Mozilla want to leave the 'any other method' option open to CAs and, if so, for how long?" If the answer is no, we should adopt 1.4.1 and require strict compliance (i.e. you can't actually use 1.4.2 or later) for section 3.2.2.4, although we might allow any later version for all other sections. If the answer is yes, we should adopt 1.4.2, and permit compliance to later versions. When we decide to stop permitting "any other method", we either update the version number to a later version of the BRs if it has that option removed by then, or we simply state that we don't allow it any more. Google's policy, AIUI, is that they are requiring strict 10-method compliance by the original effective date of ballot 169, which is 1st March 2017. I suspect we won't ship CA Policy 2.4 before that, although I'd like to ship soon after. Because of Google's requirement, I expect most CAs will be working towards that date anyway. So we could probably piggy-back on that. Comments? Gerv ___ dev-security-policy mailing list dev-security-policy@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-security-policy
Re: Policy 2.4 Proposal: Update required version number of Baseline Requirements to 1.3.7
Gerv, I'd like to push a little and suggest that the IP issues are not a significant reason for Mozilla not to formalize on 1.4.1 (e.g. with 169 included) Notably, 1.3.7 also has IP encumbrances - and uncertainty - the same as 1.4.1, so presumably, Mozilla is OK with having encumbered methods included. Considering some of these exclusions have existed since the BR's adoption, that doesn't seem an unreasonable conclusion. So what's the difference between 1.3.7 and 1.4.1? - A few methods were introduced which may or may not be encumbered - The ability for the CA to select anything they want to argue is equivalent is removed I would presume that your contention with 139 is not because new methods were (potentially) encumbered, but on the basis that it removes the any other method. Given that there are other methods available (as reaffirmed by Ballot 181) that have no encumbrances by CA/B Forum members, and given that potentially *any or all* of the methods used may be encumbered by non-Forum members (who have no obligation to disclose), it does not seem that it creates any new meaningful risk for Mozilla to impose 1.4.1 upon CAs. Indeed, if anything, the recent events shown with GoDaddy hopefully demonstrates to Mozilla that 1.4.1 (that is, with Ballot 169 included) provides better security to your users and the Internet at large, by formally prohibiting the use of methods outside of that list. Given that you can always revisit it if a CA can provide demonstrable evidence of concern and proposed alternatives, without waiting on the CA/Browser Forum, I'd like to encourage you that 1.4.1 is no worse than 1.3.7, either technically or from an IP encumbrance perspective, but is significantly and substantially better. On Thu, Jan 12, 2017 at 12:44 PM, Jeremy Rowleywrote: > I agree with this approach. Nothing of note was include after the domain > validation passed so making 1.3.7 the effective version makes sense. > > -Original Message- > From: dev-security-policy > [mailto:dev-security-policy-bounces+jeremy.rowley=digicert.com@lists.mozilla > .org] On Behalf Of Gervase Markham > Sent: Thursday, January 12, 2017 9:45 AM > To: mozilla-dev-security-pol...@lists.mozilla.org > Subject: Policy 2.4 Proposal: Update required version number of Baseline > Requirements to 1.3.7 > > Point 12 of the Inclusion section requires conformance to the Baseline > Requirements version 1.3, released on 16th April 2015. The current version > is 1.4.1. > > I propose changing to version 1.3.7. This is the one before the version > which updated the domain validation requirements and which has had to be > walked back due to the IPR issues. Once the dust settles, we can look at > updating again. See the bug for more info on the logic here. > > This is: https://github.com/mozilla/pkipolicy/issues/30 > > --- > > This is a proposed update to Mozilla's root store policy for version 2.4. > Please keep discussion in this group rather than on Github. Silence is > consent. > > Policy 2.3 (current version): > https://github.com/mozilla/pkipolicy/blob/2.3/rootstore/policy.md > Update process: > https://wiki.mozilla.org/CA:CertPolicyUpdates > ___ > dev-security-policy mailing list > dev-security-policy@lists.mozilla.org > https://lists.mozilla.org/listinfo/dev-security-policy > > ___ > dev-security-policy mailing list > dev-security-policy@lists.mozilla.org > https://lists.mozilla.org/listinfo/dev-security-policy > ___ dev-security-policy mailing list dev-security-policy@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-security-policy
RE: Policy 2.4 Proposal: Update required version number of Baseline Requirements to 1.3.7
I agree with this approach. Nothing of note was include after the domain validation passed so making 1.3.7 the effective version makes sense. -Original Message- From: dev-security-policy [mailto:dev-security-policy-bounces+jeremy.rowley=digicert.com@lists.mozilla .org] On Behalf Of Gervase Markham Sent: Thursday, January 12, 2017 9:45 AM To: mozilla-dev-security-pol...@lists.mozilla.org Subject: Policy 2.4 Proposal: Update required version number of Baseline Requirements to 1.3.7 Point 12 of the Inclusion section requires conformance to the Baseline Requirements version 1.3, released on 16th April 2015. The current version is 1.4.1. I propose changing to version 1.3.7. This is the one before the version which updated the domain validation requirements and which has had to be walked back due to the IPR issues. Once the dust settles, we can look at updating again. See the bug for more info on the logic here. This is: https://github.com/mozilla/pkipolicy/issues/30 --- This is a proposed update to Mozilla's root store policy for version 2.4. Please keep discussion in this group rather than on Github. Silence is consent. Policy 2.3 (current version): https://github.com/mozilla/pkipolicy/blob/2.3/rootstore/policy.md Update process: https://wiki.mozilla.org/CA:CertPolicyUpdates ___ dev-security-policy mailing list dev-security-policy@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-security-policy smime.p7s Description: S/MIME cryptographic signature ___ dev-security-policy mailing list dev-security-policy@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-security-policy