john-users - Re: Cracking a GPG keyring

Openwall GNU/*/Linux 3.0 - a small security-enhanced Linux distro for servers

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date: Tue, 4 Dec 2012 22:14:36 +0400
From: Solar Designer <solar@...nwall.com>
To: john-users@...ts.openwall.com
Subject: Re: Cracking a GPG keyring

Dhiru -

On Tue, Dec 04, 2012 at 05:54:04PM +0100, magnum wrote:
> On 4 Dec, 2012, at 12:42 , Dhiru Kholia <dhiru.kholia@...il.com> wrote:
> > On Tue, Dec 4, 2012 at 5:07 PM, magnum <john.magnum@...hmail.com> wrote:
> >> On 4 Dec, 2012, at 12:29 , Dhiru Kholia <dhiru.kholia@...il.com> wrote:
> >>> Ensure that your key uses SHA1 based s2k function before using the
> >>> OpenCL gpg format.
> >> 
> >> Will gpg2john and/or the format's valid() reject it if unsupported? Or will it happily try to attack it without a chance?
> > 
> > This rejection should be done in format's valid method. However, I
> > have been lazy in doing it so far :(
> 
> In cases like this it's *really* important that it gets rejected somewhere in the chain (could be in gpg2john as well, or even as an assertion in in get_salt() but that is a crude solution for no reason) and IMHO you should place it in, well, the top line of your to-do list. Imagine someone spending literally months with 96 CPU cores just to find out your format gladly and silently accepted input that it simply can not crack.

I'm with magnum on this.  Implementing proper valid() functions for all
of your formats should be your first priority now.  It is a higher
priority task than adding more formats.

Thanks,

Alexander