Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 30 Jul 2013 03:56:04 +0400
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: Parallella: bcrypt

Katja,

On Tue, Jul 30, 2013 at 12:44:09AM +0200, Katja Malvoni wrote:
> Code is in https://github.com/kmalvoni/JohnTheRipper/tree/master

I took a look at the code.  Impressive! :-)

I think you've made the task slightly more complicated for yourself (and
for future contributors to this code, if any) by keeping this source
file in pure asm (with macros, though) rather than using the C
preprocessor along with asm (as I had suggested).  Note that if you
change the filename from parallella_bf_encrypt2.s to
parallella_bf_encrypt2.S (with uppercase "S") you'd be able to use cpp
macros, including e.g. to #define human-friendly names for registers
(and #undef and re- #define them when you reuse the registers).  "gcc"
and I expect "e-gcc" too automatically invokes the C preprocessor on .S
files and then passes the preprocessor output to the assembler.

BTW, I think a better filename would be epiphany_bcrypt.s or
parallella_e_bcrypt.s (or .S as appropriate, if you switch to using
cpp).  We may use more than just the Epiphany chip on Parallella boards,
whereas this file contains code specifically for Epiphany.  I guess the
"e" in parallella_e_bcrypt.c refers to Epiphany, right?  Perhaps these
two files - the .c and the asm (whether .s or .S) should have the same
base filename.

Alexander

Powered by blists - more mailing lists

Your e-mail address:

Powered by Openwall GNU/*/Linux - Powered by OpenVZ