openssl CVE-2014-0224 fix broke tls_session_secret_cb and EAP-FAST

Thanks for reporting this!

1.0.1 commit:

https://git.openssl.org/gitweb/?p=openssl.git;a=commit;h=0d4d2e02eb55f3a03e2a8e39b723b2a2ba436584

0.9.8 is not affected.

This bug was fixed in the package openssl - 1.0.1f-1ubuntu5

---------------
openssl (1.0.1f-1ubuntu5) utopic; urgency=medium

  * SECURITY UPDATE: regression with tls_session_secret_cb (LP: #1329297)
    - debian/patches/CVE-2014-0224.patch: set the CCS_OK flag when using
      tls_session_secret_cb for session resumption in ssl/s3_clnt.c.
 -- Marc Deslauriers <email address hidden> Thu, 12 Jun 2014 08:23:12 -0400

This bug was fixed in the package openssl - 1.0.1-4ubuntu5.15

---------------
openssl (1.0.1-4ubuntu5.15) precise-security; urgency=medium

  * SECURITY UPDATE: regression with tls_session_secret_cb (LP: #1329297)
    - debian/patches/CVE-2014-0224.patch: set the CCS_OK flag when using
      tls_session_secret_cb for session resumption in ssl/s3_clnt.c.
 -- Marc Deslauriers <email address hidden> Thu, 12 Jun 2014 08:30:56 -0400

This bug was fixed in the package openssl - 1.0.1e-3ubuntu1.5

---------------
openssl (1.0.1e-3ubuntu1.5) saucy-security; urgency=medium

  * SECURITY UPDATE: regression with tls_session_secret_cb (LP: #1329297)
    - debian/patches/CVE-2014-0224.patch: set the CCS_OK flag when using
      tls_session_secret_cb for session resumption in ssl/s3_clnt.c.
 -- Marc Deslauriers <email address hidden> Thu, 12 Jun 2014 08:30:03 -0400

This bug was fixed in the package openssl - 1.0.1f-1ubuntu2.3

---------------
openssl (1.0.1f-1ubuntu2.3) trusty-security; urgency=medium

  * SECURITY UPDATE: regression with tls_session_secret_cb (LP: #1329297)
    - debian/patches/CVE-2014-0224.patch: set the CCS_OK flag when using
      tls_session_secret_cb for session resumption in ssl/s3_clnt.c.
 -- Marc Deslauriers <email address hidden> Thu, 12 Jun 2014 08:29:16 -0400

Thanks! Would not have believed this could get released so quickly :-)

Hi,

I'm not sure if this is the correct forum for this question, but this issue seems to have been posted as a security problem ("USN-2232-2: OpenSSL regression"). It looks like this is just a functional issue. Can anyone clarify if this is a security issue, and if so, what the implications are?

Thanks

USN-2232-1 is a security fix
USN-2232-2 is a regression fix for USN-2232-1 that has no security impact

I agree with this not being an independent security issue. There is a (mostly theoretical) potential security impact based on how applications or users react to the case where session ticket unexpectedly cannot be used. That could, at least in theory, result in trying the authentication handshake again with reduced security (e.g., EAP-FAST anonymous provisioning) even when there would be a valid session ticket still available. I don't think this would really result in practical security issues, i.e., the impact is in previously working functionality not working anymore and connections not being established. That said, it is useful to get this regression addressed in a way that makes it more likely for devices to get the update since the regression was caused by a high priority security fix that was likely applied to most devices immediately.

Thanks for the replies and clarification. That helps!