phc-discussions - Re: [PHC] Gambit code

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: Mon, 14 Apr 2014 12:36:49 -0400
From: Bill Cox <waywardgeek@...il.com>
To: discussions@...sword-hashing.net
Subject: Re: [PHC] Gambit code

On Mon, Apr 14, 2014 at 11:26 AM, Steve Thomas <steve@...tu.com> wrote:

> I noticed you reference https://cryptocoding.net/index.php/Coding_rules in
> your paper and tried to followed "use separate types for different types of
> information". The reason you are suppose to do that is so you get a
> compiler error. You are just typedef-ing them as uint8_t of some length.
> That means they are all the same type with a fancy name on them. This is
> why file gambit.cpp, line 19 compiles:
> (uint8_t*)salt
> instead of:
> (salt*)salt
>
> Just noticed a bug file gambit.cpp, line 115:
> memset(mem, 0, sizeof(uint8_t)*cost_m);
> should be:
> memset(mem, 0, sizeof(uint64_t)*cost_m);
>
>
> Huh I was going to say you missed the part on cryptocoding.net about
> bounds
> checking, but there is no part about that.
>

The optimizer will probably remove this call to memset anyway :-)

I thought about reporting this as a bug in the various entries that try to
clear a password or memory using memset, but it still conveys the intent of
the author.  When we get down to just a few entries left standing, I think
the authors should tweak their entries to do a better job clearing memory.

Bill

Content of type "text/html" skipped