1
0
mirror of https://github.com/hashcat/hashcat.git synced 2024-11-05 06:58:56 +00:00
hashcat/OpenCL/m06100_a3-pure.cl
Jens Steube 04d5e5a119 New Attack-Mode: Association Attack. Like JtR's single mode. Very early
stage. See hashcat Forum for detailed writeup.
2020-09-29 15:56:32 +02:00

224 lines
4.1 KiB
Common Lisp

/**
* Author......: See docs/credits.txt
* License.....: MIT
*/
#define NEW_SIMD_CODE
#ifdef KERNEL_STATIC
#include "inc_vendor.h"
#include "inc_types.h"
#include "inc_platform.cl"
#include "inc_common.cl"
#include "inc_simd.cl"
#include "inc_hash_whirlpool.cl"
#endif
KERNEL_FQ void m06100_mxx (KERN_ATTR_VECTOR ())
{
/**
* modifier
*/
const u64 gid = get_global_id (0);
const u64 lid = get_local_id (0);
const u64 lsz = get_local_size (0);
/**
* Whirlpool shared
*/
#ifdef REAL_SHM
LOCAL_VK u64 s_MT0[256];
LOCAL_VK u64 s_MT1[256];
LOCAL_VK u64 s_MT2[256];
LOCAL_VK u64 s_MT3[256];
LOCAL_VK u64 s_MT4[256];
LOCAL_VK u64 s_MT5[256];
LOCAL_VK u64 s_MT6[256];
LOCAL_VK u64 s_MT7[256];
for (u32 i = lid; i < 256; i += lsz)
{
s_MT0[i] = MT0[i];
s_MT1[i] = MT1[i];
s_MT2[i] = MT2[i];
s_MT3[i] = MT3[i];
s_MT4[i] = MT4[i];
s_MT5[i] = MT5[i];
s_MT6[i] = MT6[i];
s_MT7[i] = MT7[i];
}
SYNC_THREADS ();
#else
CONSTANT_AS u64a *s_MT0 = MT0;
CONSTANT_AS u64a *s_MT1 = MT1;
CONSTANT_AS u64a *s_MT2 = MT2;
CONSTANT_AS u64a *s_MT3 = MT3;
CONSTANT_AS u64a *s_MT4 = MT4;
CONSTANT_AS u64a *s_MT5 = MT5;
CONSTANT_AS u64a *s_MT6 = MT6;
CONSTANT_AS u64a *s_MT7 = MT7;
#endif
if (gid >= gid_max) return;
/**
* base
*/
const u32 pw_len = pws[gid].pw_len;
u32x w[64] = { 0 };
for (u32 i = 0, idx = 0; i < pw_len; i += 4, idx += 1)
{
w[idx] = pws[gid].i[idx];
}
/**
* loop
*/
u32x w0l = w[0];
for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
{
const u32x w0r = words_buf_r[il_pos / VECT_SIZE];
const u32x w0 = w0l | w0r;
w[0] = w0;
whirlpool_ctx_vector_t ctx;
whirlpool_init_vector (&ctx, s_MT0, s_MT1, s_MT2, s_MT3, s_MT4, s_MT5, s_MT6, s_MT7);
whirlpool_update_vector (&ctx, w, pw_len);
whirlpool_final_vector (&ctx);
const u32x r0 = ctx.h[DGST_R0];
const u32x r1 = ctx.h[DGST_R1];
const u32x r2 = ctx.h[DGST_R2];
const u32x r3 = ctx.h[DGST_R3];
COMPARE_M_SIMD (r0, r1, r2, r3);
}
}
KERNEL_FQ void m06100_sxx (KERN_ATTR_VECTOR ())
{
/**
* modifier
*/
const u64 gid = get_global_id (0);
const u64 lid = get_local_id (0);
const u64 lsz = get_local_size (0);
/**
* Whirlpool shared
*/
#ifdef REAL_SHM
LOCAL_VK u64 s_MT0[256];
LOCAL_VK u64 s_MT1[256];
LOCAL_VK u64 s_MT2[256];
LOCAL_VK u64 s_MT3[256];
LOCAL_VK u64 s_MT4[256];
LOCAL_VK u64 s_MT5[256];
LOCAL_VK u64 s_MT6[256];
LOCAL_VK u64 s_MT7[256];
for (u32 i = lid; i < 256; i += lsz)
{
s_MT0[i] = MT0[i];
s_MT1[i] = MT1[i];
s_MT2[i] = MT2[i];
s_MT3[i] = MT3[i];
s_MT4[i] = MT4[i];
s_MT5[i] = MT5[i];
s_MT6[i] = MT6[i];
s_MT7[i] = MT7[i];
}
SYNC_THREADS ();
#else
CONSTANT_AS u64a *s_MT0 = MT0;
CONSTANT_AS u64a *s_MT1 = MT1;
CONSTANT_AS u64a *s_MT2 = MT2;
CONSTANT_AS u64a *s_MT3 = MT3;
CONSTANT_AS u64a *s_MT4 = MT4;
CONSTANT_AS u64a *s_MT5 = MT5;
CONSTANT_AS u64a *s_MT6 = MT6;
CONSTANT_AS u64a *s_MT7 = MT7;
#endif
if (gid >= gid_max) return;
/**
* digest
*/
const u32 search[4] =
{
digests_buf[DIGESTS_OFFSET].digest_buf[DGST_R0],
digests_buf[DIGESTS_OFFSET].digest_buf[DGST_R1],
digests_buf[DIGESTS_OFFSET].digest_buf[DGST_R2],
digests_buf[DIGESTS_OFFSET].digest_buf[DGST_R3]
};
/**
* base
*/
const u32 pw_len = pws[gid].pw_len;
u32x w[64] = { 0 };
for (u32 i = 0, idx = 0; i < pw_len; i += 4, idx += 1)
{
w[idx] = pws[gid].i[idx];
}
/**
* loop
*/
u32x w0l = w[0];
for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
{
const u32x w0r = words_buf_r[il_pos / VECT_SIZE];
const u32x w0 = w0l | w0r;
w[0] = w0;
whirlpool_ctx_vector_t ctx;
whirlpool_init_vector (&ctx, s_MT0, s_MT1, s_MT2, s_MT3, s_MT4, s_MT5, s_MT6, s_MT7);
whirlpool_update_vector (&ctx, w, pw_len);
whirlpool_final_vector (&ctx);
const u32x r0 = ctx.h[DGST_R0];
const u32x r1 = ctx.h[DGST_R1];
const u32x r2 = ctx.h[DGST_R2];
const u32x r3 = ctx.h[DGST_R3];
COMPARE_S_SIMD (r0, r1, r2, r3);
}
}