phc-discussions - Re: [PHC] Doubts about AntCrypt floating-point use

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: <53207.1396387308@critter.freebsd.dk>
Date: Tue, 01 Apr 2014 21:21:48 +0000
From: "Poul-Henning Kamp" <phk@....freebsd.dk>
To: discussions@...sword-hashing.net, CodesInChaos <codesinchaos@...il.com>
Subject: Re: [PHC] Doubts about AntCrypt floating-point use

In message <CAK9dnSxe1WppG-dL09zDVymn-AnDrher2d8GXbjfczzRncKp7A@...l.gmail.com>
, CodesInChaos writes:

>I expect the floating point use to be very non-portable.

Were it just floating point, but it is actually trignometic functions,
sin(), cos() and tan().

Never mind that they are hard to implement in hardware, they
are almost as hard to implement i software.

Almost all trig functions uses approximations and therfore it is
very common, amost expected, to see differences in the last bit of
the mantissa between different implementations or even the same
implementation on different platforms, compilers etc.

Using trig functions is almost guaranteed to be a total killer
for password portability, including across operating and hardware
upgrades to the intended computer.

Poul-Henning

(Disclosure:  PHC panel member.)

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk@...eBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.