1146026 - Distrust MSCHOLDING intermediate certificate

Status: RESOLVED FIXED

(NSS :: CA Certificates Code, task)

Description

[Richard Barnes [:rbarnes]]

An intermediate certificate under the CNNIC root issued a certificate for google.com.  We should distrust that intermediate in NSS.

The only thing that gives me pause about this change is that the intermediate expires on 6 Apr 2015, so the benefit would be rather short-lived.

Comment 1

[Richard Barnes [:rbarnes]]

Attachment: MSCHOLDING intermediate certificate

Comment 2

[Richard Barnes [:rbarnes]]

Attachment: Mis-issued Google EE certificate

Comment 3

[Richard Barnes [:rbarnes]]

Attachment: bug-1146026.patch

Comment 4

[Kai Engert (:KaiE:)]

Yes, given the short lived nature of this intermediate, I wonder how much benefit it would have.

Has the private key been published for this intermediate?

Would Mozilla be willing to push out a chemspill release for this?

Comment 5

[Kai Engert (:KaiE:)]

Comment on attachment 8581175 [details] [diff] [review]
bug-1146026.patch

Richard, for the purpose of confirming your patch, I tried to reproduce your patch by using the intermediate you have attached, and running the addbuiltin utility.

I'm confused, because my patch differs from your's in the encoding of the serial number.

Even more confusing, there are two sections, one for the certificate, one for the trust record. Your patch has different encodings in the sections, but they should be identical.

How did you create the patch?

I'll attach the patch that I've created using:

cat /tmp/int_orig.pem| atob| addbuiltin -t p,p,p -n "Explicitly Distrusted MCSHOLDING CA"

Comment 6

[Kai Engert (:KaiE:)]

Attachment: patch v2

Comment 7

[Richard Barnes [:rbarnes]]

(In reply to Kai Engert (:kaie) from comment #5)
> Comment on attachment 8581175 [details] [diff] [review]
> bug-1146026.patch
>
> How did you create the patch?
> 
> I'll attach the patch that I've created using:
> 
> cat /tmp/int_orig.pem| atob| addbuiltin -t p,p,p -n "Explicitly Distrusted
> MCSHOLDING CA"

I hacked it by hand, so your version should be the right one.

I'm still not totally convinced that this is worth landing.  Mozilla is not planning a chemspill for this.  The public key has not been leaked, and MSC Holdings claims to have deleted their copy.

Comment 8

[Kai Engert (:KaiE:)]

(In reply to Richard Barnes [:rbarnes] from comment #7)
> I'm still not totally convinced that this is worth landing.  Mozilla is not
> planning a chemspill for this.  The public key has not been leaked, and MSC
> Holdings claims to have deleted their copy.

We discussed it today in the call. We thought it would be reasonable to:
- don't rush
- push out the intermediate anyway.

Richard, if you r+ed my patch, I'd to the following:

- land both this bug and bug 1145270
- release it as a new root CA list (version 2.4)
- create a NSS minibranch release 3.18.0.1 which has these two changes

Anyone who wants this NSS update could then easily pick it up.

We could offer to Firefox drivers to take it for Firefox 38, which already uses 3.18, and taking 3.18.0.1 would be a non-code change, only CA list change, so very low risk.

Comment 9

[Richard Barnes [:rbarnes]]

Comment on attachment 8582599 [details] [diff] [review]
patch v2

Review of attachment 8582599 [details] [diff] [review]:
-----------------------------------------------------------------

It's fine with me if you want to proceed with this.  But it doesn't seem hugely important.  Given that the certificate is already expired, the only real benefit is that clients with clocks that are badly wrong.

Comment 10

[Kai Engert (:KaiE:)]

can we open this bug?

Comment 11

[Kai Engert (:KaiE:)]

https://hg.mozilla.org/projects/nss/rev/c1bd3c0eaa34

Comment 12

[Kai Engert (:KaiE:)]

Landed into NSS_3_18_BRANCH
https://hg.mozilla.org/projects/nss/rev/624fd149bd3a

Comment 13

[Kai Engert (:KaiE:)]

(In reply to Kai Engert (:kaie) from comment #10)
> can we open this bug?

Comment 14

[Ryan VanderMeulen [:RyanVM]]

I'm guessing this is wontfix for esr31? Unless we end up taking 3.19.1 there as well.

Comment 15

[Kai Engert (:KaiE:)]

(In reply to Ryan VanderMeulen [:RyanVM UTC-4] from comment #14)
> I'm guessing this is wontfix for esr31?

Agreed.


> Unless we end up taking 3.19.1 there as well.

We'll take a 3.19.1 that's modified to keep the old set of root CA certificates (pre 1024-bit removal).

Comment 16

[Christiane Ruetten [INACTIVE]]

Al, what's the advisory status for Firefox on this?

Comment 17

[Al Billings [:abillings - ex-MoCo]]

(In reply to Christiane Ruetten [:cr] from comment #16)
> Al, what's the advisory status for Firefox on this?

We don't write advisories for certificate changes.