[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sat, 8 Jun 2013 17:16:59 +0200
From: Lukas Odzioba <lukas.odzioba@...il.com>
To: john-dev@...ts.openwall.com
Subject: Re: sha3-opencl
2013/6/7 Dániel Bali <balijanosdaniel@...il.com>:
> Host side:
> https://github.com/balidani/JohnTheRipper/blob/bleeding-jumbo/src/opencl_rawkeccak256_fmt.c
>
> Kernel:
> https://github.com/balidani/JohnTheRipper/blob/bleeding-jumbo/src/opencl/keccak256_kernel.cl
First my comments to your code:
1) current copyright:
/*
* Keccak-256 OpenCL version
* by Daniel Bali <balijanosdaniel at gmail.com>
* based on public domain code by Matt Mahoney
* based on rawKeccak256_fmt.c by Dhiru Kholia
*
* Usage: john --format:raw-keccak-256-opencl <hash file>
*
* This file is part of John the Ripper password cracker,
* Copyright (c) 2013 by Solar Designer
*
*/
According to this document:
http://openwall.info/wiki/john/licensing
I think that it would be better if it looks like this:
/*
* Keccak-f[1600] 256bit OpenCL
* This software is Copyright (2013) Daniel Bali <balijanosdaniel at gmail.com>,
* and it is hereby released to the general public under the following terms:
* Redistribution and use in source and binary forms, with or without
* modification, are permitted.
* Code is based on:
* - public domain code by Matt Mahoney //this for kernel only
* - rawKeccak256_fmt.c by Dhiru Kholia //this for fmt only
*/
Since Solar did not any changes to this file there is no need to add his name.
Usage is optional, especially in non plug formats I doubt that someone
will read that, you might leave it if you wish.
2) You should add a few more tests, for example an empty string and
string with lenght = PLAINTEXT_LENGTH
3) Please get rid of those warnings:
opencl_rawkeccak256_fmt.c: In function ‘crypt_all’:
opencl_rawkeccak256_fmt.c:331:9: warning: unused variable ‘i’
[-Wunused-variable]
4) crypt_all_benchmark and crypt_all functions looks identical for me.
5) #define BUFSIZE ((PLAINTEXT_LENGTH+3)/4*4) //This is
eqal to 3, are you sure this is a proper value?
6) I suppose that your problems with memory are related to mess in
size of the buffers:
One bug below:
buffer_out = clCreateBuffer(context[ocl_gpu_id], CL_MEM_WRITE_ONLY,
DIGEST_SIZE * kpc, NULL, &ret_code);
#define CIPHERTEXT_LENGTH 64
#define DIGEST_SIZE 32
Lukas
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
