phc-discussions - Re: [PHC] PolyPassHash PHS interface

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>] [day] [month] [year] [list]

Message-ID: <53497EE7.4030606@gmail.com>
Date: Sat, 12 Apr 2014 13:59:03 -0400
From: Santiago Torres <torresariass@...il.com>
To: discussions@...sword-hashing.net
CC: Justin Cappos <jcappos@....edu>, waywardgeek@...il.com
Subject: Re: [PHC] PolyPassHash PHS interface

Bill,

Yes, the output seems "random". This behavior lies in the operating
nature of the PolyPassHash scheme: we need to create a store
in order to produce a Hash.

After considering this, we decided to implement the PHS function in the
following way:

1) Create a new store after a call. This implies a new secret is
generated, hence the random output.
2) create a new polyhashed entry, this would belong to a user, but we
do not care about users now.
3) copy the value of the polyhashed entry to the output buffer
4) destroy the store.

We decided to do this in order to provide a time cost that should vary
with the input. The time cost, in this case, is the threshold value for
the store. We took this design decision based on the fact that vault
creation and unlocking is the only cost-intensive operation that resides
in the scheme; unlocking was not an option either since the store is not
tune-able at that point.

Taking this into consideration, integrating PolyPassHash into your test
vectors seems to be difficult, but not impossible. The cost of doing
this would be reduced secrecy, but it can be done; we can move to a
chroot environment and overwrite dev/urandom to do the testing:

mkdir pph_chroot
mkdir -p pph_chroot/bin
mount --bind /bin pph_chroot/bin
mkdir -p pph_chroot/usr
mount --bind /usr pph_chroot/usr
mkdir -p pph_chroot/lib
mount --bind /lib pph_chroot/lib

# if you have 64-bit binaries
mkdir lib64
mount --bind /lib64 pph_chroot/lib64

mkdir -p pph_chroot/dev/

# create a fake urandom file...
dd if=/dev/urandom of=pph_chroot/dev/urandom count=2048

chroot pph_chroot
# ... do the testing.

You might need to bind some other locations, depending on the libraries
you are using to run your test vectors.

I hope this is good enough to allow testing, but if not, please let me
know and I can help you put something together.

-Santiago.