phc-discussions - Re: [PHC] Mechanical tests (was: POMELO fails the dieharder tests)

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]

Date: Sat, 5 Apr 2014 15:57:02 +0100
From: Peter Maxwell <peter@...icient.co.uk>
To: "discussions@...sword-hashing.net" <discussions@...sword-hashing.net>
Subject: Re: [PHC] Mechanical tests (was: POMELO fails the dieharder tests)

On 5 April 2014 15:34, Poul-Henning Kamp <phk@....freebsd.dk> wrote:

>
>
> [snip]

>
>
> We are not, we are in the business of making sure that entropy is
> not lost, and we do not care if an algorithm spits out 100 bits
> with full entropy or 1000 bits each with only 1/10th bit of entropy.
>

So, essentially, we're looking for injective mappings?  That, in theory
at least, should be easy enough to check by hand for a specific algorithm.

>
> So one of my ideas for "mechanical tests" was to use compressors
> as measurement entropy, and run sanity-checking test along the
> general scheme of:
>
>         Generate_test_data > file1
>         candidate_algorithm < file1 > file2
>         gzip -9v file1 file2
>
> Any algorithm where file2.gz can be smaller than file1.gz is toast.
>

That won't tell you whether you lose entropy but rather whether a series of
results tested is amenable to a specific compression algorithm.

If the result of a single password hash has low hamming weight and
compresses easily, it tells you nothing about whether the password hashing
function is an injection (if it is an injective mapping then you're
guaranteed not to have collisions).

In a very contrived example, a password hash function operating on, say, an
input space that is strictly much smaller than the output space can still
be injective yet have all its output populate low hamming weight elements
of the output space.  You could measure as many samples as you wish, find
they all compress well, but yet have not lost entropy or be in any danger
of the function creating a collision.

Content of type "text/html" skipped