[Servercert-wg] [EXTERNAL]-Re: SCXX Ballot proposal: Debian Weak keys

Wed Oct 27 17:43:25 UTC 2021

> Hi Jaime.  Ooh, you're right!  The affected OpenSSL versions generate the same predictable moduli regardless of the public exponent value.

Yes, that’s great to know; thanks for pointing it out.

> What's the best way to capture all this in the ballot?

I think we can merely state that CAs must check for Debian weak keys for all RSA modulus lengths and exponents that they accept. Using a comparison of the modulus (or its hash) is essentially an implementation detail that we don’t need to explicitly mandate.

Thanks,

Corey

From: Rob Stradling <rob at sectigo.com> 
Sent: Wednesday, October 27, 2021 5:31 AM
To: Jaime Hablutzel <jhablutz at WISEKEY.COM>; CA/B Forum Server Certificate WG Public Discussion List <servercert-wg at cabforum.org>
Cc: Corey Bonnell <Corey.Bonnell at digicert.com>; Christopher Kemmerer <chris at ssl.com>
Subject: Re: [EXTERNAL]-Re: [Servercert-wg] SCXX Ballot proposal: Debian Weak keys

Hi Jaime.  Ooh, you're right!  The affected OpenSSL versions generate the same predictable moduli regardless of the public exponent value.

So yes, the optimal approach seems to be for CAs to use Debian weak key blocklists that are based on only the RSA modulus.

Corey's point applies if a CA chooses instead to implement a Debian weak key blocklist of (for example) SubjectPublicKeyInfos with public exponent 65537.

What's the best way to capture all this in the ballot?

  _____  

From: Jaime Hablutzel
Sent: Sunday, October 24, 2021 23:25
To: Rob Stradling; CA/B Forum Server Certificate WG Public Discussion List
Cc: Corey Bonnell; Christopher Kemmerer
Subject: Re: [EXTERNAL]-Re: [Servercert-wg] SCXX Ballot proposal: Debian Weak keys 

Hi, I might be (very) wrong here, but, shouldn’t blocklists be based only on the RSA modulus for different key sizes so validation implementations match the module only irrespective of whatever the public exponent is? or does the affected prime generation random source seed from the public exponent too?

On 22 Oct 2021, at 08:58, Rob Stradling via Servercert-wg <servercert-wg at cabforum.org> wrote:

> ...my opinion is that we should introduce a new requirement such that CAs must check for Debian weak keys for all RSA modulus lengths and exponents that they accept. CAs are uniquely positioned to prevent the usage of these weak keys in the web PKI, so there is a security benefit in mandating such universal checks.

Hi Corey.  Yeah, OK.  You've persuaded me.

FWIW, my tools at https://github.com/CVE-2008-0166 <https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_CVE-2D2008-2D0166&d=DwMGaQ&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=-bX5hBm1IdRDykQ-dBR8tsFRCM4v1VXUyG7RZa2WqPY&m=fMDCutmeJbXlHHWIZLMy2UAZB79bm_AVGAAADmUsNAE&s=gZAtYdIgwjZ_F9FpjPlUFmh9SQve9WXOyzZCTDLhsH4&e=>  only support 65537 at the moment.  I guess I'll just have to wait and see if anyone asks for other public exponent values to be supported.  🙂

  _____  

From: Corey Bonnell
Sent: Tuesday, October 19, 2021 19:48
To: Rob Stradling; Christopher Kemmerer; CA/B Forum Server Certificate WG Public Discussion List
Subject: RE: [Servercert-wg] SCXX Ballot proposal: Debian Weak keys 

Hi Rob,

Comments inline.

> AFAICT, in the affected Debian OpenSSL versions:

  - "openssl req -newkey" had a hardcoded public exponent of 65537 (see  <https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_openssl_openssl_blob_OpenSSL-5F0-5F9-5F8f_apps_req.c-23L768&d=DwMGaQ&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=-bX5hBm1IdRDykQ-dBR8tsFRCM4v1VXUyG7RZa2WqPY&m=fMDCutmeJbXlHHWIZLMy2UAZB79bm_AVGAAADmUsNAE&s=Vu5UXlPv7euZNJXCO15ReMLK_k5MyC3YaUliVn6DQcU&e=> https://github.com/openssl/openssl/blob/OpenSSL_0_9_8f/apps/req.c#L768).

  - "openssl genrsa" defaulted to 65537, but provided a "-3" command-line option to use a public exponent of 3 instead (see  <https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_openssl_openssl_blob_OpenSSL-5F0-5F9-5F8f_apps_genrsa.c&d=DwMGaQ&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=-bX5hBm1IdRDykQ-dBR8tsFRCM4v1VXUyG7RZa2WqPY&m=fMDCutmeJbXlHHWIZLMy2UAZB79bm_AVGAAADmUsNAE&s=MXbwubefERoNQfWd4kC0f7rxRrBl5yB1YZ2Y3OmPQoo&e=> https://github.com/openssl/openssl/blob/OpenSSL_0_9_8f/apps/genrsa.c).

As you point out, the command-line tooling bundled with OpenSSL 0,9.8 generally restricted the allowed exponent. However, the RSA key generation API allowed any exponent to be specified [1], so it is possible that a custom application passed exponent values besides 3 or 65537 to the RSA key generation function.

> Are there any good reasons to continue to permit the public exponent 3 ?

Judging from Censys, it appears that there are some publicly trusted certificates containing RSA keys with an exponent of 3, so there will presumably be a (minor) ecosystem impact if an exponent value of 3 were banned. That being said, exponents smaller than 65537 are outside the SHOULD-level exponent range since BR v1.1.3 (now in section 6.1.6) so perhaps it’s time to consider strengthening the SHOULD to a MUST. Probably such a change would be outside the scope of this ballot, though.

> The "openssl-vulnkey" tool that Debian used to ship only provided blocklists for keys with public exponents of 65537, so should we take that as a sign that CAs needn't perform a Debian weak key check when the public exponent is anything other than 65537 ?

While the precedent set by accepted remediations for incidents surrounding Debian weak keys has been for CAs to check the lists distributed in the openssl-blacklist Debian package, my opinion is that we should introduce a new requirement such that CAs must check for Debian weak keys for all RSA modulus lengths and exponents that they accept. CAs are uniquely positioned to prevent the usage of these weak keys in the web PKI, so there is a security benefit in mandating such universal checks.

Thanks,

Corey

[1] https://github.com/openssl/openssl/blob/OpenSSL_0_9_8f/crypto/rsa/rsa_gen.c#L78 <https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_openssl_openssl_blob_OpenSSL-5F0-5F9-5F8f_crypto_rsa_rsa-5Fgen.c-23L78&d=DwMGaQ&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=-bX5hBm1IdRDykQ-dBR8tsFRCM4v1VXUyG7RZa2WqPY&m=fMDCutmeJbXlHHWIZLMy2UAZB79bm_AVGAAADmUsNAE&s=BZt9wGuErHLlj4PgA-Q_BWX-TmBE7NrL_QZcjyFCmLs&e=> 

From: Rob Stradling <rob at sectigo.com <mailto:rob at sectigo.com> > 
Sent: Tuesday, October 19, 2021 11:31 AM
To: Christopher Kemmerer <chris at ssl.com <mailto:chris at ssl.com> >; CA/B Forum Server Certificate WG Public Discussion List <servercert-wg at cabforum.org <mailto:servercert-wg at cabforum.org> >; Corey Bonnell <Corey.Bonnell at digicert.com <mailto:Corey.Bonnell at digicert.com> >
Subject: Re: [Servercert-wg] SCXX Ballot proposal: Debian Weak keys

Hi Corey.

AFAICT, in the affected Debian OpenSSL versions:

  - "openssl req -newkey" had a hardcoded public exponent of 65537 (see  <https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_openssl_openssl_blob_OpenSSL-5F0-5F9-5F8f_apps_req.c-23L768&d=DwMGaQ&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=-bX5hBm1IdRDykQ-dBR8tsFRCM4v1VXUyG7RZa2WqPY&m=fMDCutmeJbXlHHWIZLMy2UAZB79bm_AVGAAADmUsNAE&s=Vu5UXlPv7euZNJXCO15ReMLK_k5MyC3YaUliVn6DQcU&e=> https://github.com/openssl/openssl/blob/OpenSSL_0_9_8f/apps/req.c#L768).

  - "openssl genrsa" defaulted to 65537, but provided a "-3" command-line option to use a public exponent of 3 instead (see  <https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_openssl_openssl_blob_OpenSSL-5F0-5F9-5F8f_apps_genrsa.c&d=DwMGaQ&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=-bX5hBm1IdRDykQ-dBR8tsFRCM4v1VXUyG7RZa2WqPY&m=fMDCutmeJbXlHHWIZLMy2UAZB79bm_AVGAAADmUsNAE&s=MXbwubefERoNQfWd4kC0f7rxRrBl5yB1YZ2Y3OmPQoo&e=> https://github.com/openssl/openssl/blob/OpenSSL_0_9_8f/apps/genrsa.c).

Are there any good reasons to continue to permit the public exponent 3 ?

The "openssl-vulnkey" tool that Debian used to ship only provided blocklists for keys with public exponents of 65537, so should we take that as a sign that CAs needn't perform a Debian weak key check when the public exponent is anything other than 65537 ?

  _____  

From: Servercert-wg <servercert-wg-bounces at cabforum.org <mailto:servercert-wg-bounces at cabforum.org> > on behalf of Corey Bonnell via Servercert-wg <servercert-wg at cabforum.org <mailto:servercert-wg at cabforum.org> >
Sent: 19 October 2021 15:31
To: Christopher Kemmerer <chris at ssl.com <mailto:chris at ssl.com> >; CA/B Forum Server Certificate WG Public Discussion List <servercert-wg at cabforum.org <mailto:servercert-wg at cabforum.org> >
Subject: Re: [Servercert-wg] SCXX Ballot proposal: Debian Weak keys

CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.

Hi Chris,

Apologies for the late reply. I noticed that the current proposed language has no guidance regarding RSA exponents. I think it would be useful to specify the expectations in this regard (whether the CA must check for weak keys for all key lengths and exponent combinations accepted/supported by the CA, or if checking weak key lists for only exponents 3 and 65537 is sufficient, etc.).

Thanks,

Corey

From: Servercert-wg <servercert-wg-bounces at cabforum.org <mailto:servercert-wg-bounces at cabforum.org> > On Behalf Of Christopher Kemmerer via Servercert-wg
Sent: Friday, October 15, 2021 10:33 AM
To: Rob Stradling <rob at sectigo.com <mailto:rob at sectigo.com> >; Dimitris Zacharopoulos (HARICA) <dzacharo at harica.gr <mailto:dzacharo at harica.gr> >; CA/B Forum Server Certificate WG Public Discussion List <servercert-wg at cabforum.org <mailto:servercert-wg at cabforum.org> >; Jacob Hoffman-Andrews <jsha at letsencrypt.org>
Subject: Re: [Servercert-wg] SCXX Ballot proposal: Debian Weak keys

Thank you, Rob, and shall watch for that update. Meanwhile we are doing a final-final pass through our draft language for clarity and will send it early next week.

Chris K

Meanwhile, we've cycled our draft language through  another review and have made IIRC only one or two minor edits for clarity (h/t BenW).

On 10/14/2021 9:49 AM, Rob Stradling wrote:

Today I rediscovered that I'd previously generated the RSA-8192 blocklists back in December 2009, and that they're still available at https://secure.sectigo.com/debian_weak_keys/.  When I compared the old and new RSA-8192 blocklists, I found that ~0.8% of the "rnd" keys are different.  It looks like, for reasons unknown, the "OpenSSL random file state" misbehaved occasionally over the 8 month run that ended recently.

I'll report back once I've regenerated and verified the problematic keys.

  _____  

From: Rob Stradling  <mailto:rob at sectigo.com> <rob at sectigo.com>
Sent: 23 September 2021 19:17
To: Christopher Kemmerer  <mailto:chris at ssl.com> <chris at ssl.com>; Dimitris Zacharopoulos (HARICA)  <mailto:dzacharo at harica.gr> <dzacharo at harica.gr>; CA/B Forum Server Certificate WG Public Discussion List  <mailto:servercert-wg at cabforum.org> <servercert-wg at cabforum.org>; Jacob Hoffman-Andrews  <mailto:jsha at letsencrypt.org> <jsha at letsencrypt.org>; Rob Stradling <mailto:rob at sectigo.com> <rob at sectigo.com>
Subject: Re: [Servercert-wg] SCXX Ballot proposal: Debian Weak keys

> BTW, in case it helps, I'm about half way through generating a full set of RSA-8192 Debian weak keys, which (when complete) I'll add to the https://github.com/CVE-2008-0166 repositories.

It took nearly 8 months (using just a single core of a fairly modest CPU), but it finally finished!  Repositories updated.

  _____  

From: Servercert-wg  <mailto:servercert-wg-bounces at cabforum.org> <servercert-wg-bounces at cabforum.org> on behalf of Rob Stradling via Servercert-wg  <mailto:servercert-wg at cabforum.org> <servercert-wg at cabforum.org>
Sent: 13 May 2021 15:42
To: Christopher Kemmerer  <mailto:chris at ssl.com> <chris at ssl.com>; Dimitris Zacharopoulos (HARICA)  <mailto:dzacharo at harica.gr> <dzacharo at harica.gr>; CA/B Forum Server Certificate WG Public Discussion List  <mailto:servercert-wg at cabforum.org> <servercert-wg at cabforum.org>; Jacob Hoffman-Andrews  <mailto:jsha at letsencrypt.org> <jsha at letsencrypt.org>
Subject: Re: [Servercert-wg] SCXX Ballot proposal: Debian Weak keys

CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.

> iii) All RSA Public Key lengths supported by the CA up to and including 4096 bits;

> ...

> For Debian weak keys not covered above, the CA SHALL take actions to minimize the probability of certificate issuance.

Hi Christopher.  What sort of "actions" are envisaged here?  If a CA is processing a certificate request that contains a (for example) RSA-4088 public key (i.e., a key size not covered by an available Debian weak list), either the CA is going to issue the cert or they're not.  What, concretely, does "minimize the probability of certificate issuance" actually mean?

Why not remove that "SHALL" sentence and change point iii to: "iii) All RSA Public Key lengths supported by the CA." ?

BTW, in case it helps, I'm about half way through generating a full set of RSA-8192 Debian weak keys, which (when complete) I'll add to the https://github.com/CVE-2008-0166 repositories.

  _____  

From: Christopher Kemmerer  <mailto:chris at ssl.com> <chris at ssl.com>
Sent: 13 May 2021 15:12
To: Rob Stradling  <mailto:rob at sectigo.com> <rob at sectigo.com>; Dimitris Zacharopoulos (HARICA)  <mailto:dzacharo at harica.gr> <dzacharo at harica.gr>; CA/B Forum Server Certificate WG Public Discussion List  <mailto:servercert-wg at cabforum.org> <servercert-wg at cabforum.org>; Jacob Hoffman-Andrews  <mailto:jsha at letsencrypt.org> <jsha at letsencrypt.org>
Subject: Re: [Servercert-wg] SCXX Ballot proposal: Debian Weak keys

CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.

Hello,

We deeply appreciate the useful discussion in this thread regarding this issue. We especially applaud the efforts of HARICA and Sectigo to independently generate more comprehensive lists of potentially affected Debian weak keys. As Rob Stradling observed through his crt.sh research (20210107, https://gist.github.com/robstradling/a5590b6a13218fe561dcb5d5c67932c5) of the five most utilized algorithm/key size populations, two are ECC (so not impacted by the Debian weak key issue) and three are RSA (2048, 4096, and 3072 bit length, in that order).

As of their most recent messages it appears that these two organizations have independently generated comprehensive lists identifying all RSA-2048 and -4096 bit length keys. (We understand RSA-3072 length keys are also available.) This offers the possibility that complete lists, if accepted as authoritative, could be accessed by the community to help prevent exploitation of this vulnerability.

It was also noted (by the representative from Let's Encrypt) that the ROCA vulnerability is presently identified through use of a tool supported externally. It was suggested that this resource be archived in a manner that ensures availability. (Our proposed language points to "https://github.com/crocs-muni/roca or equivalent.")

We think our present ballot language (reproduced at the end of this message) provides appropriately focused guidance to CAs. If available, we'd certainly like to also see the HARICA/Sectigo lists (which CAs could use for the majority of Debian weak key use cases) captured somewhere in this ballot language. We are agnostic as to 1) where exactly these resources might be maintained and 2) where this ballot places directions to these resources - an annex to the current requirements, a separate CA/BF guidance document or within Sections 4.9.1.1/6.1.1.3.

Our intent is to ensure that 1) clear, accurate guidance on CA expectations is provided and 2) any resources assisting CAs in meeting these expectations are fully described, publicly available (somewhere) and with reliable links provided. The language below, we feel, meets the first requirement. We'd appreciate input on how to best meet the second. (Note that SSL.com <http://ssl.com/>  would be happy to support the community by hosting any of these as publicly accessible resources, whether solo or alongside other organizations.)

Chris K 

SSL.com <http://ssl.com/> 

===== 

--- Motion Begins --- 

This ballot modifies the “Baseline Requirements for the Issuance and Management of Publicly-Trusted Certificates” as follows, based on Version 1.7.4: 

Proposed ballot language: 

4.9.1.1 Reasons for Revoking a Subscriber Certificate 

Replace: 

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) 

With: 

4. The CA is made aware of a demonstrated or proven method that can easily compute the Subscriber’s Private Key (such as those identified in 6.1.1.3(4)). 

--- 

6.1.1.3. Subscriber Key Pair Generation 

Replace: 

The CA SHALL reject a certificate request if one or more of the following conditions are met: 

1. The Key Pair does not meet the requirements set forth in Section 6.1.5 and/or Section 6.1.6; 

2. There is clear evidence that the specific method used to generate the Private Key was flawed; 

3. The CA is aware of a demonstrated or proven method that exposes the Applicant's Private Key to compromise; 

4. The CA has previously been made aware that the Applicant's Private Key has suffered a Key Compromise, such as through the provisions of Section 4.9.1.1; 

5. The CA is aware of a demonstrated or proven method to easily compute the Applicant's Private Key based on the Public Key (such as a Debian weak key, see https://wiki.debian.org/SSLkeys). 

With: 

The CA SHALL reject a certificate request if one or more of the following occurs: 

1) The requested Public Key does not meet the requirements set forth in Sections 6.1.5 and/or 6.1.6; 

2) The CA is aware of a demonstrated or proven method that exposes the Subscriber's Private Key to compromise; 

3) The CA has previously been made aware that the Subscriber's Private Key has suffered a Key Compromise, such as through the provisions of Section 4.9.1.1; 

4) The Public Key corresponds to an industry demonstrated weak Private Key, in particular: 

a) In the case of ROCA vulnerability, the CA SHALL reject keys identified by the tools available at https://github.com/crocs-muni/roca or equivalent. 

b) In the case of Debian weak keys (https://wiki.debian.org/SSLkeys), the CA SHALL reject at least keys generated by the flawed OpenSSL version with the combination of the following parameters: 

i) Big-endian 32-bit, little-endian 32-bit, and little-endian 64-bit architecture; 

ii) Process ID of 0 to 32767, inclusive; 

iii) All RSA Public Key lengths supported by the CA up to and including 4096 bits; 

iv) rnd, nornd, and noreadrnd OpenSSL random file state. 

For Debian weak keys not covered above, the CA SHALL take actions to minimize the probability of certificate issuance. 

--- Motion Ends ---

On 1/18/2021 3:34 PM, Rob Stradling wrote:

> I'm mid-way through generating the RSA-4096 keys.

The RSA-4096 private keys and blocklists are now in https://github.com/CVE-2008-0166/private_keys andhttps://github.com/CVE-2008-0166/openssl_blocklists.

The RSA-2048 and RSA-4096 private keys in https://github.com/HARICA-official/debian-weak-keys (which only covers 2 of the 3 word size / endianness combinations) are identical to the equivalents in https://github.com/CVE-2008-0166/private_keys.

  _____  

From: Dimitris Zacharopoulos (HARICA)  <mailto:dzacharo at harica.gr> <dzacharo at harica.gr>
Sent: 14 January 2021 18:39
To: Rob Stradling  <mailto:rob at sectigo.com> <rob at sectigo.com>; CA/B Forum Server Certificate WG Public Discussion List  <mailto:servercert-wg at cabforum.org> <servercert-wg at cabforum.org>; Jacob Hoffman-Andrews  <mailto:jsha at letsencrypt.org> <jsha at letsencrypt.org>; Christopher Kemmerer  <mailto:chris at ssl.com> <chris at ssl.com>
Subject: Re: [Servercert-wg] SCXX Ballot proposal: Debian Weak keys

CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.

On 14/1/2021 12:30 π.μ., Rob Stradling wrote:

Thanks Dmitris.

So far I've generated the RSA-2048 and RSA-3072 keys using https://github.com/CVE-2008-0166/key_generator and uploaded them to https://github.com/CVE-2008-0166/private_keys, and I've generated the corresponding blocklists and uploaded them to https://github.com/CVE-2008-0166/openssl_blocklists.  My RSA-2048 blocklists exactly match the ones from the original Debian openssl-blacklist package.

I'm mid-way through generating the RSA-4096 keys.

Let's compare keys when we're both done.  🙂

Certainly :-) the RSA-2048 keys already match the fingerprints from the openssl-blacklist Debian package.

We did this work several months ago but never found the time to make it publicly available. We managed to break down the big task and run jobs in parallel which made things a bit more interesting.

It's nice we did this independently, I guess it increases the accuracy level of the resulted keys :)

Cheers,
Dimitris.

  _____  

From: Dimitris Zacharopoulos (HARICA)  <mailto:dzacharo at harica.gr> <dzacharo at harica.gr>
Sent: 13 January 2021 21:49
To: Rob Stradling  <mailto:rob at sectigo.com> <rob at sectigo.com>; CA/B Forum Server Certificate WG Public Discussion List  <mailto:servercert-wg at cabforum.org> <servercert-wg at cabforum.org>; Jacob Hoffman-Andrews  <mailto:jsha at letsencrypt.org> <jsha at letsencrypt.org>; Christopher Kemmerer  <mailto:chris at ssl.com> <chris at ssl.com>
Subject: Re: [Servercert-wg] SCXX Ballot proposal: Debian Weak keys

CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.

Dear friends,

HARICA has generated the weak keys (RSA 2048 and 4096 bit lengths) from the vulnerable openssl package. We will generate 3072 bit keys as well and add them soon. The methodology is described in the following GitHub repo along with the produced keys:

*	https://github.com/HARICA-official/debian-weak-keys

Please review and let us know if you spot any issues or problems with our approach and methodology.

As always, please use other people's work at your own risk.

Dimitris.

On 7/1/2021 2:25 μ.μ., Rob Stradling via Servercert-wg wrote:

I've used crt.sh to produce a survey of key algorithms/sizes in currently unexpired, publicly-trusted server certificates:

https://gist.github.com/robstradling/a5590b6a13218fe561dcb5d5c67932c5

The four most popular choices are no surprise: RSA-2048, P-256, RSA-4096, and P-384.  openssl-blacklist covers RSA-2048 and RSA-4096, and ECC keys are implicitly not Debian weak keys.

Fifth most popular is RSA-3072, with over 3 million unexpired, publicly-trusted server certs.  openssl-blacklist doesn't cover RSA-3072, but ISTM that this is a key size that CAs will want to permit.

Some of the lesser used key sizes are mostly likely due to Subscriber typos (e.g., 2408 and 3048 were probably intended to be 2048, 4048 was probably intended to be either 2048 or 4096, etc), but some of the other ones look like they were deliberately chosen (e.g., 2432 is 2048+384).  Is it worth generating Debian weak keys/blocklists for any of these key sizes?

https://nvlpubs.nist.gov/nistpubs/SpecialPublications/NIST.SP.800-57pt1r5.pdf (Table 4, p59) permits RSA-2048 until the end of 2030, whereas https://www.sogis.eu/documents/cc/crypto/SOGIS-Agreed-Cryptographic-Mechanisms-1.2.pd f permits RSA-2048 only until the end of 2025.  It is of course possible that quantum computing will render RSA obsolete before Subscribers need to think about which larger RSA keysize they want to migrate to; however, it seems prudent to also plan for the possibility that RSA will survive and that some other RSA keysize(s) might become popular.

  _____  

From: Servercert-wg  <mailto:servercert-wg-bounces at cabforum.org> <servercert-wg-bounces at cabforum.org> on behalf of Rob Stradling via Servercert-wg  <mailto:servercert-wg at cabforum.org> <servercert-wg at cabforum.org>
Sent: 06 January 2021 16:08
To: Jacob Hoffman-Andrews  <mailto:jsha at letsencrypt.org> <jsha at letsencrypt.org>; Christopher Kemmerer  <mailto:chris at ssl.com> <chris at ssl.com>; CA/B Forum Server Certificate WG Public Discussion List  <mailto:servercert-wg at cabforum.org> <servercert-wg at cabforum.org>
Subject: Re: [Servercert-wg] SCXX Ballot proposal: Debian Weak keys

CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.

Jacob wrote:

> Lastly, I think we should archive openssl-blacklist, and include in the BRs: "A CA may reject the full set of Debian weak keys by rejecting this superset of the Debian weak keys:

> 

> - All RSA public keys with modulus lengths other than 2048 or 4096, and

> - All RSA public keys with exponents other than 65537, and

Hi Jacob.  65537 (aka 0x10001) is hard-coded here...

https://github.com/openssl/openssl/blob/OpenSSL_0_9_8f/apps/req.c#L768

Would it therefore be fair to say that keys with public exponents other than 65537 are implicitlynot Debian weak keys?

> - All RSA public keys that are detected as vulnerable by the openssl-vulnkey program in the openssl-blacklist package version 0.5-3 (see addendum), or an equivalent program."

  _____  

From: Servercert-wg  <mailto:servercert-wg-bounces at cabforum.org> <servercert-wg-bounces at cabforum.org> on behalf of Jacob Hoffman-Andrews via Servercert-wg  <mailto:servercert-wg at cabforum.org> <servercert-wg at cabforum.org>
Sent: 12 December 2020 02:21
To: Christopher Kemmerer  <mailto:chris at ssl.com> <chris at ssl.com>; CA/B Forum Server Certificate WG Public Discussion List  <mailto:servercert-wg at cabforum.org> <servercert-wg at cabforum.org>
Subject: Re: [Servercert-wg] SCXX Ballot proposal: Debian Weak keys

CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.

Thanks for your continued efforts to improve this part of the BRs! Let's Encrypt is in theory interested in endorsing, but I think it still needs a bit of work. Thanks for incorporating my most recent comments on endianness and word size vs 11 platforms.

Goals: We want CAs to consistently not issue certificates for weak keys in general, and also in the specific case of Debian and ROCA keys. We want the definition of Debian and ROCA keys to be clear and actionable for as long as possible - say, at least twenty years.

We have three ways to specify Debian and ROCA keys: With a list, with a tool, or with an algorithm*. The original revision of this ballot proposed to use a list (https://lists.cabforum.org/pipermail/servercert-wg/2020-April/001821.html). There were two objections:

 - The list (openssl-blacklist) is subject to change or removal.
 - The list only covers 2048 and 4096 bit keys.

The current draft proposes specifying a tool for ROCA (https://github.com/crocs-muni/roca) and an algorithm for Debian keys.

The ROCA tool is subject to change or removal, just like the openssl-blacklist package. I propose we instead specify ROCA detection in terms of the paper (https://crocs.fi.muni.cz/public/papers/rsa_ccs17) and ask for permission from the authors to archive an unchanging copy as an addendum to the BRs.

For Debian keys, what looks like an algorithm specification is actually a tool + algorithm specification. The tool is "OpenSSL 0.9.8c-1 up to versions before 0.9.8g-9 on Debian-based operating systems" (per CVE-2008-01666 - https://cve.mitre.org/cgi-bin/cvename.cgi?name=2008-0166). To ensure an unchanging copy of that, we should archive 3 copies of Debian, for the 3 word size + endianness combinations.

The algorithm also needs an additional line: "v) using the command 'openssl req -nodes -subj / -newkey rsa:<Public Key length>'" (adapted from https://sources.debian.org/data/main/o/openssl-blacklist/0.5-3/examples/gen_certs.sh). Other tools that linked OpenSSL, like openvpn and openssh, generated different sets of keys. We can include or exclude openvpn and openssh keys, but should thoroughly specify.

Lastly, I think we should archive openssl-blacklist, and include in the BRs: "A CA may reject the full set of Debian weak keys by rejecting this superset of the Debian weak keys:

 - All RSA public keys with modulus lengths other than 2048 or 4096, and
 - All RSA public keys with exponents other than 65537, and
 - All RSA public keys that are detected as vulnerable by the openssl-vulnkey program in the openssl-blacklist package version 0.5-3 (see addendum), or an equivalent program."

My reasoning: Given the difficulty of correctly setting up old Debian versions and generating weak keys for sizes that are not part of openssl-blacklist, I expect most CAs will choose this path. Given that, we should just say what we mean: the pregenerated list is fine if you restrict key sizes, but you don't *have* to restrict key sizes, so long as you have an alternate method to ensure you're not issuing for Debian weak keys at other sizes.

*I'm considering specifying an algorithm to be functionally equivalent to specifying an "outcome," though I recognize this may be too hand-wavy.

_______________________________________________
Servercert-wg mailing list
Servercert-wg at cabforum.org <mailto:Servercert-wg at cabforum.org> 
https://lists.cabforum.org/mailman/listinfo/servercert-wg

_______________________________________________
Servercert-wg mailing list
 <mailto:Servercert-wg at cabforum.org> Servercert-wg at cabforum.org
 <https://urldefense.proofpoint.com/v2/url?u=https-3A__lists.cabforum.org_mailman_listinfo_servercert-2Dwg&d=DwICAg&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=-bX5hBm1IdRDykQ-dBR8tsFRCM4v1VXUyG7RZa2WqPY&m=fMDCutmeJbXlHHWIZLMy2UAZB79bm_AVGAAADmUsNAE&s=z9o7_4hZP712S4Pmo87r-EDp1vhMXmppIaYtKPbuJWw&e=> https://urldefense.proofpoint.com/v2/url?u=https-3A__lists.cabforum.org_mailman_listinfo_servercert-2Dwg&d=DwICAg&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=-bX5hBm1IdRDykQ-dBR8tsFRCM4v1VXUyG7RZa2WqPY&m=fMDCutmeJbXlHHWIZLMy2UAZB79bm_AVGAAADmUsNAE&s=z9o7_4hZP712S4Pmo87r-EDp1vhMXmppIaYtKPbuJWw&e=

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cabforum.org/pipermail/servercert-wg/attachments/20211027/ddbcf234/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 4990 bytes
Desc: not available
URL: <http://lists.cabforum.org/pipermail/servercert-wg/attachments/20211027/ddbcf234/attachment-0001.p7s>