Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×
Encryption Networking Security IT

OpenSSL Security Update Less Critical Than Expected, Still Recommended 64

An anonymous reader writes As announced on Monday, the OpenSSL project team has released new versions of the cryptographic library that fix a number of security issues. The announcement created a panic within the security community, who were dreading the discovery of another Heartbleed-type bug, but as it turns out, the high severity issue fixed is a bug than can be exploited in a DoS attack against servers. Other issues fixed are mostly memory corruption and DoS flaws of moderate and low severity.
This discussion has been archived. No new comments can be posted.

OpenSSL Security Update Less Critical Than Expected, Still Recommended

Comments Filter:
  • by Anonymous Coward on Thursday March 19, 2015 @11:56AM (#49292869)

    For those unaware, the OpenBSD team forked OpenSSL a while back and started a huge cleanup of ugly existing codebase. Their project is named LibreSSL, and is available here: https://github.com/libressl-portable/portable

    So how did they do?

    CVEs that don't effect LibreSSL:
    OpenSSL 1.0.2 ClientHello sigalgs DoS (CVE-2015-0291) - Severity: High
    Reclassified: RSA silently downgrades to EXPORT_RSA [Client] (CVE-2015-0204) - Severity: High
    Base64 decode (CVE-2015-0292) - Severity: Moderate
    Multiblock corrupted pointer (CVE-2015-0290) - Severity: Moderate
    Segmentation fault in DTLSv1_listen (CVE-2015-0207) - Severity: Moderate
    Segmentation fault for invalid PSS parameters (CVE-2015-0208) - Severity: Moderate
    DoS via reachable assert in SSLv2 servers (CVE-2015-0293) - Severity: Moderate
    Empty CKE with client auth and DHE (CVE-2015-1787) - Severity: Moderate
    Handshake with unseeded PRNG (CVE-2015-0285) - Severity: Low
    CVEs that effect LibreSSL:
    Segmentation fault in ASN1_TYPE_cmp (CVE-2015-0286) - Severity: Moderate
    ASN.1 structure reuse memory corruption (CVE-2015-0287) - Severity: Moderate
    PKCS7 NULL pointer dereferences (CVE-2015-0289) - Severity: Moderate
    Use After Free following d2i_ECPrivatekey error (CVE-2015-0209) - Severity: Low
    X509_to_X509_REQ NULL pointer deref (CVE-2015-0288) - Severity: Low

    So LibreSSL had already avoided 9 of these issues as a result of their code cleanup. This includes all CVEs labelled as high severity. This is just another reminder to use LibreSSL.

    Sources:

    https://marc.info/?l=openbsd-announce&m=142677546015662
    https://www.reddit.com/r/openbsd/comments/2zl6y4/no_highseverity_issues_from_openssl_were_present/

    • by Noryungi ( 70322 )

      Yup, I have the feeling that LibreSSL is going to replace OpenSSL like OpenSSH replaced SSH as ''the'' standard.

      The fact that both LibreSSL and OpenSSH are OpenBSD project is not a coincidence...

      More details on Undeadly [undeadly.org].

      • Maybe

        With ssh the original project had moved to a propietary license so linux distros that only accepted free software had to go with a fork or stick with a very outdated version. With openssl the original project is still alive. So the developers of linux distros will have to have a big argument over whether the reduced security exposure outweighs the reduced feature set.

        • Theres also a thorny license issue, some projects released under the GPL make a exception for openssl and it's not always clear whether that would apply to forks of openssl.

          • by Noryungi ( 70322 )

            AFAIK, OpenSSL is Apache Licensed and LibreSSL is, well... BSD-Licensed.

            If you accept an Apache-style license, I really don't see why LibreSSL's BSD is a problem.

            You had a better argument when it came to the fact that OpenSSL is still active. Or, at least, that there is activity in the project, including some projects to audit the whole thing.

            • by MSG ( 12810 ) on Thursday March 19, 2015 @01:19PM (#49293649)

              No, OpenSSL is not Apache licensed. It has its own license, similar to BSD-with-attribution license. And the thorny issue is that this license is not compatible with the GPL. That's why projects have to modify the GPL to make a specific exception for it.

              It's also why Red Hat started work to standardize on Mozilla's NSS as the one true SSL library. However, I'm not sure what the status of that project is.

              • And the thorny issue is that this license is not compatible with the GPL. That's why projects have to modify the GPL to make a specific exception for it.

                Exactly and in most cases the exception says "openssl". Does a slightly patched version from a distro still count as "openssl"? Does a forked and renamed version with substantial changes still count as "openssl"?

    • So LibreSSL had already avoided 9 of these issues as a result of their code cleanup.

      5 of them at least a result of forking before relevant code/feature existed.

      CVE-2015-0208, CVE-2015-0207, CVE-2015-0290, CVE-2015-0285 and CVE-2015-0291

      This includes all CVEs labelled as high severity. This is just another reminder to use LibreSSL.

      I think having other forks and more people working a project is ultimately great for everyone. The tit-for-tat elitism and misleading hyperbole is not productive.

      • The libressl fork was 11 months ago. They managed to add 5 (at a minimum) critical vulnerabilities in the past 11 months? Jeezus fucking christ.

        When you're in a hole, stop digging.

        • The libressl fork was 11 months ago. They managed to add 5 (at a minimum) critical vulnerabilities in the past 11 months?

          Probably a *lot* more than that. These are only bugs having been caught thus far.

          Jeezus fucking christ.

          OpenSSL is currently offering and maintaining four separate release trains for download from the bleeding edge to ancient versions lacking TLS 1.1/1.2 support.

          Hard to get excited about DOS/crash shit limited to a new immature branch only a dufus would select for production use... or in other words ...OMFG the sky is falling..

    • by Burz ( 138833 )

      That is not such a big difference, considering most installations are still using OpenSSL (more eyes...).

      LibreSSL is still valued for their efforts, but they and most of the IT community waited until a major crisis occurred before taking action. Now that OpenSSL has been in the spotlight and finally received decent funding to do their own reviews and cleanup, I'm not sure where that leaves LibreSSL.

    • by Elessar ( 8997 )

      This is not a fair comparison.

      LibreSSL forked OpenSSL 1.0.1. Therefore LibreSSL would never have been vulnerable to issues that did not affect 1.0.1 - since those arose after the codebases split. A fairer comparison would be to compare issues that affected OpenSSL 1.0.1 with LibreSSL. You also should not include CVE-2015-0204 since that is just a reclassification of a previously fixed defect. Simillarly CVE-2015-0292 was a historic issue not in recent versions of OpenSSL so also should not be included. By t

"The vast majority of successful major crimes against property are perpetrated by individuals abusing positions of trust." -- Lawrence Dalzell

Working...