Web lists-archives.com

Re: openssl/libssl1 in Debian now blocks offlineimap?




On Tue, Aug 15, 2017 at 05:04:50PM +0200, Kurt Roeckx wrote:
> My problem is that if we don't do something, TLS 1.0 will be used
> for an other 10 year, and that's just not acceptable.

My problem is that the cause you're fighting, while laudable, should not
be fought in Debian.

Debian is a general-purpose operating system, not a security-focused
one. If we were, people would be right to expect that there might
sometimes be problems with the software they're using in support of the
"security" goal. Since we're not, people would rightly be upset if the
complex environments they're trying to support are suddenly broken in
the name of "security".

> So I would like to do something so that hopefully by the time Buster
> releases you can disable TLS 1.0 by default, and that almost no users
> would need to enable it again.
> 
> Having TLS 1.0 (and 1.1) enabled by default itself is not a
> problem, it's actually using it that's a problem. There are
> clearly still too many that don't support TLS 1.2, but it's
> getting better.

So ship a version of OpenSSL that ships with TLS1.0 and TLS1.1 (aka
TLS1.old) disabled, but *allow people to re-enable it* if things break
(without requiring them to compile their own version). By shipping a
version of OpenSSL that has TLS1.old not even compiled in, you're not
doing that.

I have seen suggestions that the goal of this would be to re-enable
TLS1.old before buster releases. If that is true, I fail to see the
point; I think it makes *much* more sense to ship buster with TLS1.old
disabled by default, but to allow for re-enabling it on a
per-application basis. Having a policy configuration that can be
modified by the system administrator through a mechanism inside libssl
that cannot be disabled and that can be configured on a per-application
basis would seem to be a much more interesting way to do this.
 
> Disabling the protocols is the only way I know how to identify
> all the problems. And I would like to encourage everybody to
> contact the other side if things break and get them to upgrade.

That latter part is certainly a good idea; and while "get the other side
to upgrade" is often a reasonable course of action, just as often it is
not, and then you need to have a way out. By not even compiling in the
code to support TLS1.old, you're depriving people of that "way out".

-- 
Could you people please use IRC like normal people?!?

  -- Amaya Rodrigo Sastre, trying to quiet down the buzz in the DebConf 2008
     Hacklab