lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite for Android: free password hash cracker in your pocket
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAOLP8p4q_bB=9OVrDBFbC09eWz9UKEwZP1=54da2OmKSWj6HLA@mail.gmail.com>
Date: Thu, 23 Apr 2015 16:38:16 -0700
From: Bill Cox <waywardgeek@...il.com>
To: "discussions@...sword-hashing.net" <discussions@...sword-hashing.net>
Subject: Re: [PHC] (not) protecting password length from side-channels (Re:
 [PHC] Argon2 modulo division)

On Thu, Apr 23, 2015 at 3:57 PM, Ben Harris <ben@...rr.is> wrote:

> On 24/04/2015 1:38 am, "Bill Cox" <waywardgeek@...il.com> wrote:
> >
> > On Wed, Apr 22, 2015 at 5:47 PM, Solar Designer <solar@...nwall.com>
> wrote:
> >>
> >> On Wed, Apr 22, 2015 at 09:13:21AM -0700, Bill Cox wrote:
> >> > It turns out that there was not a single entry in the competition
> that is
> >> > power rail analysis resistant
> >>
> >> ... with respect to password length.
> >>
> >> I'm with Thomas on this.  It is futile for PHC candidates to fully
> >> protect the password length.
>
> Isn't the best solution to just hash the password at the source so it is
> the same length (it is also incompressible which prevents against another
> class of attacks)?
>
I prefer this solution.  If you compute H(salt || domain || password) when
talking to a remote URL, and send an ASCII hash digest instead of a
password, you gain a lot of protection, regardless of what hashing
algorithm the remote server uses.  If you do the hashing in hardware, you
gain even more protection.  I would love to see hardware backed password
managers that do this.

As for being incompressible, I don't thiink this defeats all compression
related attacks.  If I'm transmitting the password over a channel where an
attacker can inject data and also measure the compression, he can still
tell if his injected data matches a portion of the incompressible
password.  The only solution for this I can think of is to stop the lower
level code from trying to compress security related data.  This is doable,
but would be difficult in existing SSL/TLS implementations.  One way would
be  to embed commands to the SSL layer along with application data, so you
can surround passwords with commands to disable compression.

> Do any of the second round candidates seem like their power usage side
> channel is significant enough for attacks like this
> http://www.tau.ac.il/~tromer/acoustic/
>
I don't think an audio analysis would have enough resolution, but I could
be wrong.  If a slow microcontroller were doing the hashing, then maybe.

Bill

Content of type "text/html" skipped

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ