arno/hashcat
1
0
Fork 0
mirror of https://github.com/hashcat/hashcat.git synced 2024-11-13 19:28:56 +00:00
Code Issues Releases Wiki Activity

New SIMD code for -a 1 -m 10

Browse Source
This commit is contained in:
Jens Steube 2016-02-21 18:40:06 +01:00
parent 21c66ea301
commit 0b29be3f86
1 changed files with 194 additions and 248 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

442
OpenCL/m00010_a1.cl
View File

@ -5,6 +5,8 @@
#define _MD5_
#define NEW_SIMD_CODE
#include "include/constants.h"
#include "include/kernel_vendor.h"
@ -16,9 +18,7 @@
#include "include/kernel_functions.c"
#include "OpenCL/types_ocl.c"
#include "OpenCL/common.c"
#define COMPARE_S "OpenCL/check_single_comp4.c"
#define COMPARE_M "OpenCL/check_multi_comp4.c"
#include "OpenCL/simd.c"
__kernel void m00010_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global void *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
{
@ -36,54 +36,31 @@ __kernel void m00010_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
if (gid >= gid_max) return;
u32 wordl0[4];
u32 pws0[4] = { 0 };
u32 pws1[4] = { 0 };
wordl0[0] = pws[gid].i[ 0];
wordl0[1] = pws[gid].i[ 1];
wordl0[2] = pws[gid].i[ 2];
wordl0[3] = pws[gid].i[ 3];
u32 wordl1[4];
wordl1[0] = pws[gid].i[ 4];
wordl1[1] = pws[gid].i[ 5];
wordl1[2] = pws[gid].i[ 6];
wordl1[3] = pws[gid].i[ 7];
u32 wordl2[4];
wordl2[0] = 0;
wordl2[1] = 0;
wordl2[2] = 0;
wordl2[3] = 0;
u32 wordl3[4];
wordl3[0] = 0;
wordl3[1] = 0;
wordl3[2] = 0;
wordl3[3] = 0;
pws0[0] = pws[gid].i[0];
pws0[1] = pws[gid].i[1];
pws0[2] = pws[gid].i[2];
pws0[3] = pws[gid].i[3];
pws1[0] = pws[gid].i[4];
pws1[1] = pws[gid].i[5];
pws1[2] = pws[gid].i[6];
pws1[3] = pws[gid].i[7];
const u32 pw_l_len = pws[gid].pw_len;
if (combs_mode == COMBINATOR_MODE_BASE_RIGHT)
{
switch_buffer_by_offset_le (wordl0, wordl1, wordl2, wordl3, combs_buf[0].pw_len);
}
/**
* salt
*/
u32 salt_buf0[4];
u32 salt_buf1[4];
salt_buf0[0] = salt_bufs[salt_pos].salt_buf[0];
salt_buf0[1] = salt_bufs[salt_pos].salt_buf[1];
salt_buf0[2] = salt_bufs[salt_pos].salt_buf[2];
salt_buf0[3] = salt_bufs[salt_pos].salt_buf[3];
u32 salt_buf1[4];
salt_buf1[0] = salt_bufs[salt_pos].salt_buf[4];
salt_buf1[1] = salt_bufs[salt_pos].salt_buf[5];
salt_buf1[2] = salt_bufs[salt_pos].salt_buf[6];
@ -95,119 +72,122 @@ __kernel void m00010_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
* loop
*/
for (u32 il_pos = 0; il_pos < combs_cnt; il_pos++)
for (u32 il_pos = 0; il_pos < combs_cnt; il_pos += VECT_SIZE)
{
const u32 pw_r_len = combs_buf[il_pos].pw_len;
const u32x pw_r_len = pwlenx_create_combt (combs_buf, il_pos);
const u32 pw_len = pw_l_len + pw_r_len;
const u32x pw_len = pw_l_len + pw_r_len;
u32 wordr0[4];
/**
* concat password candidate
*/
wordr0[0] = combs_buf[il_pos].i[0];
wordr0[1] = combs_buf[il_pos].i[1];
wordr0[2] = combs_buf[il_pos].i[2];
wordr0[3] = combs_buf[il_pos].i[3];
u32x wordl0[4] = { 0 };
u32x wordl1[4] = { 0 };
u32x wordl2[4] = { 0 };
u32x wordl3[4] = { 0 };
u32 wordr1[4];
wordl0[0] = pws0[0];
wordl0[1] = pws0[1];
wordl0[2] = pws0[2];
wordl0[3] = pws0[3];
wordl1[0] = pws1[0];
wordl1[1] = pws1[1];
wordl1[2] = pws1[2];
wordl1[3] = pws1[3];
wordr1[0] = combs_buf[il_pos].i[4];
wordr1[1] = combs_buf[il_pos].i[5];
wordr1[2] = combs_buf[il_pos].i[6];
wordr1[3] = combs_buf[il_pos].i[7];
u32x wordr0[4] = { 0 };
u32x wordr1[4] = { 0 };
u32x wordr2[4] = { 0 };
u32x wordr3[4] = { 0 };
u32 wordr2[4];
wordr2[0] = 0;
wordr2[1] = 0;
wordr2[2] = 0;
wordr2[3] = 0;
u32 wordr3[4];
wordr3[0] = 0;
wordr3[1] = 0;
wordr3[2] = 0;
wordr3[3] = 0;
wordr0[0] = ix_create_combt (combs_buf, il_pos, 0);
wordr0[1] = ix_create_combt (combs_buf, il_pos, 1);
wordr0[2] = ix_create_combt (combs_buf, il_pos, 2);
wordr0[3] = ix_create_combt (combs_buf, il_pos, 3);
wordr1[0] = ix_create_combt (combs_buf, il_pos, 4);
wordr1[1] = ix_create_combt (combs_buf, il_pos, 5);
wordr1[2] = ix_create_combt (combs_buf, il_pos, 6);
wordr1[3] = ix_create_combt (combs_buf, il_pos, 7);
if (combs_mode == COMBINATOR_MODE_BASE_LEFT)
{
switch_buffer_by_offset_le (wordr0, wordr1, wordr2, wordr3, pw_l_len);
switch_buffer_by_offset_le_VV (wordr0, wordr1, wordr2, wordr3, pw_l_len);
}
else
{
switch_buffer_by_offset_le_VV (wordl0, wordl1, wordl2, wordl3, pw_r_len);
}
u32x w0[4];
u32x w1[4];
u32x w2[4];
u32x w3[4];
w0[0] = wordl0[0] | wordr0[0];
w0[1] = wordl0[1] | wordr0[1];
w0[2] = wordl0[2] | wordr0[2];
w0[3] = wordl0[3] | wordr0[3];
w1[0] = wordl1[0] | wordr1[0];
w1[1] = wordl1[1] | wordr1[1];
w1[2] = wordl1[2] | wordr1[2];
w1[3] = wordl1[3] | wordr1[3];
w2[0] = wordl2[0] | wordr2[0];
w2[1] = wordl2[1] | wordr2[1];
w2[2] = wordl2[2] | wordr2[2];
w2[3] = wordl2[3] | wordr2[3];
w3[0] = wordl3[0] | wordr3[0];
w3[1] = wordl3[1] | wordr3[1];
w3[2] = 0;
w3[3] = 0;
/**
* append salt
*/
u32 s0[4];
u32x s0[4] = { 0 };
u32x s1[4] = { 0 };
u32x s2[4] = { 0 };
u32x s3[4] = { 0 };
s0[0] = salt_buf0[0];
s0[1] = salt_buf0[1];
s0[2] = salt_buf0[2];
s0[3] = salt_buf0[3];
u32 s1[4];
s1[0] = salt_buf1[0];
s1[1] = salt_buf1[1];
s1[2] = salt_buf1[2];
s1[3] = salt_buf1[3];
u32 s2[4];
switch_buffer_by_offset_le_VV (s0, s1, s2, s3, pw_len);
s2[0] = 0;
s2[1] = 0;
s2[2] = 0;
s2[3] = 0;
const u32x pw_salt_len = pw_len + salt_len;
u32 s3[4];
s3[0] = 0;
s3[1] = 0;
s3[2] = 0;
s3[3] = 0;
switch_buffer_by_offset_le (s0, s1, s2, s3, pw_len);
const u32 pw_salt_len = pw_len + salt_len;
u32 w0[4];
w0[0] = wordl0[0] | wordr0[0] | s0[0];
w0[1] = wordl0[1] | wordr0[1] | s0[1];
w0[2] = wordl0[2] | wordr0[2] | s0[2];
w0[3] = wordl0[3] | wordr0[3] | s0[3];
u32 w1[4];
w1[0] = wordl1[0] | wordr1[0] | s1[0];
w1[1] = wordl1[1] | wordr1[1] | s1[1];
w1[2] = wordl1[2] | wordr1[2] | s1[2];
w1[3] = wordl1[3] | wordr1[3] | s1[3];
u32 w2[4];
w2[0] = wordl2[0] | wordr2[0] | s2[0];
w2[1] = wordl2[1] | wordr2[1] | s2[1];
w2[2] = wordl2[2] | wordr2[2] | s2[2];
w2[3] = wordl2[3] | wordr2[3] | s2[3];
u32 w3[4];
w3[0] = wordl3[0] | wordr3[0] | s3[0];
w3[1] = wordl3[1] | wordr3[1] | s3[1];
w3[2] = pw_salt_len * 8;
w3[3] = 0;
append_0x80_4x4 (w0, w1, w2, w3, pw_salt_len);
w0[0] |= s0[0];
w0[1] |= s0[1];
w0[2] |= s0[2];
w0[3] |= s0[3];
w1[0] |= s1[0];
w1[1] |= s1[1];
w1[2] |= s1[2];
w1[3] |= s1[3];
w2[0] |= s2[0];
w2[1] |= s2[1];
w2[2] |= s2[2];
w2[3] |= s2[3];
w3[0] |= s3[0];
w3[1] |= s3[1];
w3[2] = pw_salt_len * 8;
w3[3] = 0;
/**
* md5
*/
u32 a = MD5M_A;
u32 b = MD5M_B;
u32 c = MD5M_C;
u32 d = MD5M_D;
u32x a = MD5M_A;
u32x b = MD5M_B;
u32x c = MD5M_C;
u32x d = MD5M_D;
MD5_STEP (MD5_Fo, a, b, c, d, w0[0], MD5C00, MD5S00);
MD5_STEP (MD5_Fo, d, a, b, c, w0[1], MD5C01, MD5S01);
@ -277,13 +257,7 @@ __kernel void m00010_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
MD5_STEP (MD5_I , c, d, a, b, w0[2], MD5C3e, MD5S32);
MD5_STEP (MD5_I , b, c, d, a, w2[1], MD5C3f, MD5S33);
const u32 r0 = a;
const u32 r1 = d;
const u32 r2 = c;
const u32 r3 = b;
#include COMPARE_M
COMPARE_M_SIMD (a, d, c, b);
}
}
@ -311,54 +285,31 @@ __kernel void m00010_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
if (gid >= gid_max) return;
u32 wordl0[4];
u32 pws0[4] = { 0 };
u32 pws1[4] = { 0 };
wordl0[0] = pws[gid].i[ 0];
wordl0[1] = pws[gid].i[ 1];
wordl0[2] = pws[gid].i[ 2];
wordl0[3] = pws[gid].i[ 3];
u32 wordl1[4];
wordl1[0] = pws[gid].i[ 4];
wordl1[1] = pws[gid].i[ 5];
wordl1[2] = pws[gid].i[ 6];
wordl1[3] = pws[gid].i[ 7];
u32 wordl2[4];
wordl2[0] = 0;
wordl2[1] = 0;
wordl2[2] = 0;
wordl2[3] = 0;
u32 wordl3[4];
wordl3[0] = 0;
wordl3[1] = 0;
wordl3[2] = 0;
wordl3[3] = 0;
pws0[0] = pws[gid].i[0];
pws0[1] = pws[gid].i[1];
pws0[2] = pws[gid].i[2];
pws0[3] = pws[gid].i[3];
pws1[0] = pws[gid].i[4];
pws1[1] = pws[gid].i[5];
pws1[2] = pws[gid].i[6];
pws1[3] = pws[gid].i[7];
const u32 pw_l_len = pws[gid].pw_len;
if (combs_mode == COMBINATOR_MODE_BASE_RIGHT)
{
switch_buffer_by_offset_le (wordl0, wordl1, wordl2, wordl3, combs_buf[0].pw_len);
}
/**
* salt
*/
u32 salt_buf0[4];
u32 salt_buf1[4];
salt_buf0[0] = salt_bufs[salt_pos].salt_buf[0];
salt_buf0[1] = salt_bufs[salt_pos].salt_buf[1];
salt_buf0[2] = salt_bufs[salt_pos].salt_buf[2];
salt_buf0[3] = salt_bufs[salt_pos].salt_buf[3];
u32 salt_buf1[4];
salt_buf1[0] = salt_bufs[salt_pos].salt_buf[4];
salt_buf1[1] = salt_bufs[salt_pos].salt_buf[5];
salt_buf1[2] = salt_bufs[salt_pos].salt_buf[6];
@ -382,119 +333,122 @@ __kernel void m00010_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
* loop
*/
for (u32 il_pos = 0; il_pos < combs_cnt; il_pos++)
for (u32 il_pos = 0; il_pos < combs_cnt; il_pos += VECT_SIZE)
{
const u32 pw_r_len = combs_buf[il_pos].pw_len;
const u32x pw_r_len = pwlenx_create_combt (combs_buf, il_pos);
const u32 pw_len = pw_l_len + pw_r_len;
const u32x pw_len = pw_l_len + pw_r_len;
u32 wordr0[4];
/**
* concat password candidate
*/
wordr0[0] = combs_buf[il_pos].i[0];
wordr0[1] = combs_buf[il_pos].i[1];
wordr0[2] = combs_buf[il_pos].i[2];
wordr0[3] = combs_buf[il_pos].i[3];
u32x wordl0[4] = { 0 };
u32x wordl1[4] = { 0 };
u32x wordl2[4] = { 0 };
u32x wordl3[4] = { 0 };
u32 wordr1[4];
wordl0[0] = pws0[0];
wordl0[1] = pws0[1];
wordl0[2] = pws0[2];
wordl0[3] = pws0[3];
wordl1[0] = pws1[0];
wordl1[1] = pws1[1];
wordl1[2] = pws1[2];
wordl1[3] = pws1[3];
wordr1[0] = combs_buf[il_pos].i[4];
wordr1[1] = combs_buf[il_pos].i[5];
wordr1[2] = combs_buf[il_pos].i[6];
wordr1[3] = combs_buf[il_pos].i[7];
u32x wordr0[4] = { 0 };
u32x wordr1[4] = { 0 };
u32x wordr2[4] = { 0 };
u32x wordr3[4] = { 0 };
u32 wordr2[4];
wordr2[0] = 0;
wordr2[1] = 0;
wordr2[2] = 0;
wordr2[3] = 0;
u32 wordr3[4];
wordr3[0] = 0;
wordr3[1] = 0;
wordr3[2] = 0;
wordr3[3] = 0;
wordr0[0] = ix_create_combt (combs_buf, il_pos, 0);
wordr0[1] = ix_create_combt (combs_buf, il_pos, 1);
wordr0[2] = ix_create_combt (combs_buf, il_pos, 2);
wordr0[3] = ix_create_combt (combs_buf, il_pos, 3);
wordr1[0] = ix_create_combt (combs_buf, il_pos, 4);
wordr1[1] = ix_create_combt (combs_buf, il_pos, 5);
wordr1[2] = ix_create_combt (combs_buf, il_pos, 6);
wordr1[3] = ix_create_combt (combs_buf, il_pos, 7);
if (combs_mode == COMBINATOR_MODE_BASE_LEFT)
{
switch_buffer_by_offset_le (wordr0, wordr1, wordr2, wordr3, pw_l_len);
switch_buffer_by_offset_le_VV (wordr0, wordr1, wordr2, wordr3, pw_l_len);
}
else
{
switch_buffer_by_offset_le_VV (wordl0, wordl1, wordl2, wordl3, pw_r_len);
}
u32x w0[4];
u32x w1[4];
u32x w2[4];
u32x w3[4];
w0[0] = wordl0[0] | wordr0[0];
w0[1] = wordl0[1] | wordr0[1];
w0[2] = wordl0[2] | wordr0[2];
w0[3] = wordl0[3] | wordr0[3];
w1[0] = wordl1[0] | wordr1[0];
w1[1] = wordl1[1] | wordr1[1];
w1[2] = wordl1[2] | wordr1[2];
w1[3] = wordl1[3] | wordr1[3];
w2[0] = wordl2[0] | wordr2[0];
w2[1] = wordl2[1] | wordr2[1];
w2[2] = wordl2[2] | wordr2[2];
w2[3] = wordl2[3] | wordr2[3];
w3[0] = wordl3[0] | wordr3[0];
w3[1] = wordl3[1] | wordr3[1];
w3[2] = 0;
w3[3] = 0;
/**
* append salt
*/
u32 s0[4];
u32x s0[4] = { 0 };
u32x s1[4] = { 0 };
u32x s2[4] = { 0 };
u32x s3[4] = { 0 };
s0[0] = salt_buf0[0];
s0[1] = salt_buf0[1];
s0[2] = salt_buf0[2];
s0[3] = salt_buf0[3];
u32 s1[4];
s1[0] = salt_buf1[0];
s1[1] = salt_buf1[1];
s1[2] = salt_buf1[2];
s1[3] = salt_buf1[3];
u32 s2[4];
switch_buffer_by_offset_le_VV (s0, s1, s2, s3, pw_len);
s2[0] = 0;
s2[1] = 0;
s2[2] = 0;
s2[3] = 0;
const u32x pw_salt_len = pw_len + salt_len;
u32 s3[4];
s3[0] = 0;
s3[1] = 0;
s3[2] = 0;
s3[3] = 0;
switch_buffer_by_offset_le (s0, s1, s2, s3, pw_len);
const u32 pw_salt_len = pw_len + salt_len;
u32 w0[4];
w0[0] = wordl0[0] | wordr0[0] | s0[0];
w0[1] = wordl0[1] | wordr0[1] | s0[1];
w0[2] = wordl0[2] | wordr0[2] | s0[2];
w0[3] = wordl0[3] | wordr0[3] | s0[3];
u32 w1[4];
w1[0] = wordl1[0] | wordr1[0] | s1[0];
w1[1] = wordl1[1] | wordr1[1] | s1[1];
w1[2] = wordl1[2] | wordr1[2] | s1[2];
w1[3] = wordl1[3] | wordr1[3] | s1[3];
u32 w2[4];
w2[0] = wordl2[0] | wordr2[0] | s2[0];
w2[1] = wordl2[1] | wordr2[1] | s2[1];
w2[2] = wordl2[2] | wordr2[2] | s2[2];
w2[3] = wordl2[3] | wordr2[3] | s2[3];
u32 w3[4];
w3[0] = wordl3[0] | wordr3[0] | s3[0];
w3[1] = wordl3[1] | wordr3[1] | s3[1];
w3[2] = pw_salt_len * 8;
w3[3] = 0;
append_0x80_4x4 (w0, w1, w2, w3, pw_salt_len);
w0[0] |= s0[0];
w0[1] |= s0[1];
w0[2] |= s0[2];
w0[3] |= s0[3];
w1[0] |= s1[0];
w1[1] |= s1[1];
w1[2] |= s1[2];
w1[3] |= s1[3];
w2[0] |= s2[0];
w2[1] |= s2[1];
w2[2] |= s2[2];
w2[3] |= s2[3];
w3[0] |= s3[0];
w3[1] |= s3[1];
w3[2] = pw_salt_len * 8;
w3[3] = 0;
/**
* md5
*/
u32 a = MD5M_A;
u32 b = MD5M_B;
u32 c = MD5M_C;
u32 d = MD5M_D;
u32x a = MD5M_A;
u32x b = MD5M_B;
u32x c = MD5M_C;
u32x d = MD5M_D;
MD5_STEP (MD5_Fo, a, b, c, d, w0[0], MD5C00, MD5S00);
MD5_STEP (MD5_Fo, d, a, b, c, w0[1], MD5C01, MD5S01);
@ -561,21 +515,13 @@ __kernel void m00010_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
MD5_STEP (MD5_I , b, c, d, a, w3[1], MD5C3b, MD5S33);
MD5_STEP (MD5_I , a, b, c, d, w1[0], MD5C3c, MD5S30);
bool q_cond = allx (search[0] != a);
if (q_cond) continue;
if (MATCHES_NONE_VS (a, search[0])) continue;
MD5_STEP (MD5_I , d, a, b, c, w2[3], MD5C3d, MD5S31);
MD5_STEP (MD5_I , c, d, a, b, w0[2], MD5C3e, MD5S32);
MD5_STEP (MD5_I , b, c, d, a, w2[1], MD5C3f, MD5S33);
const u32 r0 = a;
const u32 r1 = d;
const u32 r2 = c;
const u32 r3 = b;
#include COMPARE_S
COMPARE_S_SIMD (a, d, c, b);
}
}