OpenSSL in squeeze is v. 0.9.8o-4? http://packages.debian.org/squeeze/openssl
C'mon, guys, the latest current OpenSSL is already at v. 1.2.2
This is why I'm switching to ubuntu.
OpenSSL in squeeze is v. 0.9.8o-4? http://packages.debian.org/squeeze/openssl
C'mon, guys, the latest current OpenSSL is already at v. 1.2.2
This is why I'm switching to ubuntu.
Many people don't get Debian. This is a released aimed for servers and stable workstations. If you want or need bleeding edge stuff you can use Debian testing/unstable or Ubuntu as you suggested.
Post-release, OpenSSL 1.0.0 will now be migrated to unstable, and then any problems that causes or exposes can be found and fixed on a more generous schedule.
edit: the latest OpenSSL release is 1.0.0, not 1.2.2. And development on the 0.9.8-series seems to be still active, as latest version on it was released on the same day that 1.0.0c
edit2: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=578376
I reckon I prefer stability and predictability over modernity in my encryption and signing libraries.
Seems like a flaw in the PCI requirements to me, do they really demand the "latest" version instead of the stable, time-tested one?
It certainly can't be in the spirit of these audits to encourage people to move from Debian stable to a distro that's based on Debian unstable...
That was easy, but then rebuilding other components which were linked to the original v. 0.9.x was a major PITA.
This is the problem, according to the auditor:
Vulnerability in OpenSSL 0.9.8g Severity: Critical Problem CVE: CVE2008-0891 CVE-2008-1672 CVE-2008-5077 CVE-2009-0590 CVE-2009-0789 CVE-20091377 CVE-2009-1378 CVE-2009-1379 CVE-2009-1386 CVE-2009-3245 CVE-2009-3555 CVE-2010-0433 CVE-2010-0740 Impact: A remote attacker could execute arbitrary commands , cause a buffer overflow, bypass security or create a denial of service. Resolution OpenSSL shouldbe [http://www.openssl.org/source/] upgraded to 1.0.0a or higher.
Those CVE ("Common Vulnerabilities and Exposures") items are explained in more detail at NIST: http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2008-089... http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2009-137... http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2010-074... http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2010-043... http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2009-324...
Version numbers are not supposed to change after the fact in a stable-release, hence security fixes get backported (every distro has a security-team for this).
If PCI requires a less tested newer version over a battle-scarred (patched up) older one then PCI is working against its own stated goal.
It doesn't take much wisdom to realize that it's less likely for new bugs to crop up in the 0.9.8 openssl that Debian ships than in the 1.0.0c that RHEL6 bundles (just one month after release!).
New software has bugs. Old software has less bugs.
Another thing is they release the same distrib version on 9 different architectures, not only i386/amd64.
Considering this, and the fact they're volunteers, I don't think they're that far behind.
The debian unstable->testing->stable cycle is vicious, and on a production system it's actually very sensible - by the time a package is allowed to reach stable, it will have been rigourously tested and actually be properly stable.