Re: CVE-2014-6271: remote code execution through bash

[Tavis Ormandy <taviso () google com>]

On Thu, Sep 25, 2014 at 12:18 PM, Chet Ramey <chet.ramey () case edu> wrote:
> On 9/25/14, 12:15 PM, Solar Designer wrote:
>
> > What do you think of distros' going with Florian's prefix-suffix patch
> > right now?  I think it breaks function imports/exports between
> > pre-patch and post-patch bash versions, but keeps them intact for
> > patched versions.  Right?  If so, this sounds acceptable for immediate
> > use by distros.  Do you agree?
>
> I haven't looked at that particular patch in detail yet, but I am wondering
> why adding both a prefix and a suffix is better than just adding a prefix.

I think it's just paranoia, it demonstrates significant enough control
over the environment that it's unlikely someone could construct it via
a naive cgi environment or similar. If someone can create variables
with an arbitrary suffix and prefix, then there are likely bigger
problems to worry about than what bash is doing. FWIW, I like
Florian's approach.

Tavis.