1
0
mirror of https://github.com/hashcat/hashcat.git synced 2024-11-26 09:58:16 +00:00

Cleanup -m 9710 kernels to latest standard

This commit is contained in:
Jens Steube 2016-04-17 10:18:11 +02:00
parent 0063533fde
commit e7a205cee6
3 changed files with 352 additions and 457 deletions

View File

@ -5,6 +5,9 @@
#define _OLDOFFICE01_
//too much register pressure
//#define NEW_SIMD_CODE
#include "include/constants.h"
#include "include/kernel_vendor.h"
@ -18,9 +21,7 @@
#include "OpenCL/common.c"
#include "include/rp_kernel.h"
#include "OpenCL/rp.c"
#define COMPARE_S "OpenCL/check_single_comp4.c"
#define COMPARE_M "OpenCL/check_multi_comp4.c"
#include "OpenCL/simd.c"
typedef struct
{
@ -247,10 +248,6 @@ __kernel void m09710_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
const u32 lid = get_local_id (0);
__local RC4_KEY rc4_keys[64];
__local RC4_KEY *rc4_key = &rc4_keys[lid];
/**
* base
*/
@ -260,14 +257,12 @@ __kernel void m09710_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
if (gid >= gid_max) return;
u32 pw_buf0[4];
u32 pw_buf1[4];
pw_buf0[0] = pws[gid].i[ 0];
pw_buf0[1] = pws[gid].i[ 1];
pw_buf0[2] = pws[gid].i[ 2];
pw_buf0[3] = pws[gid].i[ 3];
u32 pw_buf1[4];
pw_buf1[0] = pws[gid].i[ 4];
pw_buf1[1] = pws[gid].i[ 5];
pw_buf1[2] = pws[gid].i[ 6];
@ -275,6 +270,14 @@ __kernel void m09710_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
const u32 pw_len = pws[gid].pw_len;
/**
* shared
*/
__local RC4_KEY rc4_keys[64];
__local RC4_KEY *rc4_key = &rc4_keys[lid];
/**
* esalt
*/
@ -292,62 +295,36 @@ __kernel void m09710_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
* loop
*/
for (u32 il_pos = 0; il_pos < il_cnt; il_pos++)
for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
{
u32 w0[4];
u32x w0[4] = { 0 };
u32x w1[4] = { 0 };
u32x w2[4] = { 0 };
u32x w3[4] = { 0 };
w0[0] = pw_buf0[0];
w0[1] = pw_buf0[1];
w0[2] = pw_buf0[2];
w0[3] = pw_buf0[3];
apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
u32 w1[4];
w1[0] = pw_buf1[0];
w1[1] = pw_buf1[1];
w1[2] = pw_buf1[2];
w1[3] = pw_buf1[3];
u32 w2[4];
/**
* md5
*/
w0[0] = w0[0];
w0[1] = w0[1] & 0xff;
w0[2] = 0x8000;
w0[3] = 0;
w1[0] = 0;
w1[1] = 0;
w1[2] = 0;
w1[3] = 0;
w2[0] = 0;
w2[1] = 0;
w2[2] = 0;
w2[3] = 0;
u32 w3[4];
w3[0] = 0;
w3[1] = 0;
w3[2] = 0;
w3[2] = 9 * 8;
w3[3] = 0;
apply_rules (rules_buf[il_pos].cmds, w0, w1, pw_len);
// first md5 to generate RC4 128 bit key
u32 w0_t[4];
u32 w1_t[4];
u32 w2_t[4];
u32 w3_t[4];
w0_t[0] = w0[0];
w0_t[1] = w0[1] & 0xff;
w0_t[2] = 0x8000;
w0_t[3] = 0;
w1_t[0] = 0;
w1_t[1] = 0;
w1_t[2] = 0;
w1_t[3] = 0;
w2_t[0] = 0;
w2_t[1] = 0;
w2_t[2] = 0;
w2_t[3] = 0;
w3_t[0] = 0;
w3_t[1] = 0;
w3_t[2] = 9 * 8;
w3_t[3] = 0;
u32 digest[4];
digest[0] = MD5M_A;
@ -355,55 +332,43 @@ __kernel void m09710_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
digest[2] = MD5M_C;
digest[3] = MD5M_D;
md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
md5_transform (w0, w1, w2, w3, digest);
// now the RC4 part
u32 key[4];
key[0] = digest[0];
key[1] = digest[1];
key[2] = digest[2];
key[3] = digest[3];
rc4_init_16 (rc4_key, key);
rc4_init_16 (rc4_key, digest);
u32 out[4];
u8 j = rc4_next_16 (rc4_key, 0, 0, encryptedVerifier, out);
w0_t[0] = out[0];
w0_t[1] = out[1];
w0_t[2] = out[2];
w0_t[3] = out[3];
w1_t[0] = 0x80;
w1_t[1] = 0;
w1_t[2] = 0;
w1_t[3] = 0;
w2_t[0] = 0;
w2_t[1] = 0;
w2_t[2] = 0;
w2_t[3] = 0;
w3_t[0] = 0;
w3_t[1] = 0;
w3_t[2] = 16 * 8;
w3_t[3] = 0;
w0[0] = out[0];
w0[1] = out[1];
w0[2] = out[2];
w0[3] = out[3];
w1[0] = 0x80;
w1[1] = 0;
w1[2] = 0;
w1[3] = 0;
w2[0] = 0;
w2[1] = 0;
w2[2] = 0;
w2[3] = 0;
w3[0] = 0;
w3[1] = 0;
w3[2] = 16 * 8;
w3[3] = 0;
digest[0] = MD5M_A;
digest[1] = MD5M_B;
digest[2] = MD5M_C;
digest[3] = MD5M_D;
md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
md5_transform (w0, w1, w2, w3, digest);
rc4_next_16 (rc4_key, 16, j, digest, out);
const u32 r0 = out[0];
const u32 r1 = out[1];
const u32 r2 = out[2];
const u32 r3 = out[3];
#include COMPARE_M
COMPARE_M_SIMD (out[0], out[1], out[2], out[3]);
}
}
@ -423,10 +388,6 @@ __kernel void m09710_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
const u32 lid = get_local_id (0);
__local RC4_KEY rc4_keys[64];
__local RC4_KEY *rc4_key = &rc4_keys[lid];
/**
* base
*/
@ -436,14 +397,12 @@ __kernel void m09710_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
if (gid >= gid_max) return;
u32 pw_buf0[4];
u32 pw_buf1[4];
pw_buf0[0] = pws[gid].i[ 0];
pw_buf0[1] = pws[gid].i[ 1];
pw_buf0[2] = pws[gid].i[ 2];
pw_buf0[3] = pws[gid].i[ 3];
u32 pw_buf1[4];
pw_buf1[0] = pws[gid].i[ 4];
pw_buf1[1] = pws[gid].i[ 5];
pw_buf1[2] = pws[gid].i[ 6];
@ -452,16 +411,12 @@ __kernel void m09710_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
const u32 pw_len = pws[gid].pw_len;
/**
* digest
* shared
*/
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]
};
__local RC4_KEY rc4_keys[64];
__local RC4_KEY *rc4_key = &rc4_keys[lid];
/**
* esalt
@ -476,66 +431,52 @@ __kernel void m09710_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
encryptedVerifier[2] = oldoffice01_bufs[salt_pos].encryptedVerifier[2];
encryptedVerifier[3] = oldoffice01_bufs[salt_pos].encryptedVerifier[3];
/**
* 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]
};
/**
* loop
*/
for (u32 il_pos = 0; il_pos < il_cnt; il_pos++)
for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
{
u32 w0[4];
u32x w0[4] = { 0 };
u32x w1[4] = { 0 };
u32x w2[4] = { 0 };
u32x w3[4] = { 0 };
w0[0] = pw_buf0[0];
w0[1] = pw_buf0[1];
w0[2] = pw_buf0[2];
w0[3] = pw_buf0[3];
apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
u32 w1[4];
w1[0] = pw_buf1[0];
w1[1] = pw_buf1[1];
w1[2] = pw_buf1[2];
w1[3] = pw_buf1[3];
u32 w2[4];
/**
* md5
*/
w0[0] = w0[0];
w0[1] = w0[1] & 0xff;
w0[2] = 0x8000;
w0[3] = 0;
w1[0] = 0;
w1[1] = 0;
w1[2] = 0;
w1[3] = 0;
w2[0] = 0;
w2[1] = 0;
w2[2] = 0;
w2[3] = 0;
u32 w3[4];
w3[0] = 0;
w3[1] = 0;
w3[2] = 0;
w3[2] = 9 * 8;
w3[3] = 0;
apply_rules (rules_buf[il_pos].cmds, w0, w1, pw_len);
// first md5 to generate RC4 128 bit key
u32 w0_t[4];
u32 w1_t[4];
u32 w2_t[4];
u32 w3_t[4];
w0_t[0] = w0[0];
w0_t[1] = w0[1] & 0xff;
w0_t[2] = 0x8000;
w0_t[3] = 0;
w1_t[0] = 0;
w1_t[1] = 0;
w1_t[2] = 0;
w1_t[3] = 0;
w2_t[0] = 0;
w2_t[1] = 0;
w2_t[2] = 0;
w2_t[3] = 0;
w3_t[0] = 0;
w3_t[1] = 0;
w3_t[2] = 9 * 8;
w3_t[3] = 0;
u32 digest[4];
digest[0] = MD5M_A;
@ -543,55 +484,43 @@ __kernel void m09710_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
digest[2] = MD5M_C;
digest[3] = MD5M_D;
md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
md5_transform (w0, w1, w2, w3, digest);
// now the RC4 part
u32 key[4];
key[0] = digest[0];
key[1] = digest[1];
key[2] = digest[2];
key[3] = digest[3];
rc4_init_16 (rc4_key, key);
rc4_init_16 (rc4_key, digest);
u32 out[4];
u8 j = rc4_next_16 (rc4_key, 0, 0, encryptedVerifier, out);
w0_t[0] = out[0];
w0_t[1] = out[1];
w0_t[2] = out[2];
w0_t[3] = out[3];
w1_t[0] = 0x80;
w1_t[1] = 0;
w1_t[2] = 0;
w1_t[3] = 0;
w2_t[0] = 0;
w2_t[1] = 0;
w2_t[2] = 0;
w2_t[3] = 0;
w3_t[0] = 0;
w3_t[1] = 0;
w3_t[2] = 16 * 8;
w3_t[3] = 0;
w0[0] = out[0];
w0[1] = out[1];
w0[2] = out[2];
w0[3] = out[3];
w1[0] = 0x80;
w1[1] = 0;
w1[2] = 0;
w1[3] = 0;
w2[0] = 0;
w2[1] = 0;
w2[2] = 0;
w2[3] = 0;
w3[0] = 0;
w3[1] = 0;
w3[2] = 16 * 8;
w3[3] = 0;
digest[0] = MD5M_A;
digest[1] = MD5M_B;
digest[2] = MD5M_C;
digest[3] = MD5M_D;
md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
md5_transform (w0, w1, w2, w3, digest);
rc4_next_16 (rc4_key, 16, j, digest, out);
const u32 r0 = out[0];
const u32 r1 = out[1];
const u32 r2 = out[2];
const u32 r3 = out[3];
#include COMPARE_S
COMPARE_S_SIMD (out[0], out[1], out[2], out[3]);
}
}

View File

@ -5,6 +5,9 @@
#define _OLDOFFICE01_
//too much register pressure
//#define NEW_SIMD_CODE
#include "include/constants.h"
#include "include/kernel_vendor.h"
@ -16,9 +19,7 @@
#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"
typedef struct
{
@ -245,10 +246,6 @@ __kernel void m09710_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
const u32 lid = get_local_id (0);
__local RC4_KEY rc4_keys[64];
__local RC4_KEY *rc4_key = &rc4_keys[lid];
/**
* base
*/
@ -257,40 +254,27 @@ __kernel void m09710_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
if (gid >= gid_max) return;
u32 wordl0[4];
u32 pw_buf0[4];
u32 pw_buf1[4];
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;
pw_buf0[0] = pws[gid].i[0];
pw_buf0[1] = pws[gid].i[1];
pw_buf0[2] = pws[gid].i[2];
pw_buf0[3] = pws[gid].i[3];
pw_buf1[0] = pws[gid].i[4];
pw_buf1[1] = pws[gid].i[5];
pw_buf1[2] = pws[gid].i[6];
pw_buf1[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);
}
/**
* shared
*/
__local RC4_KEY rc4_keys[64];
__local RC4_KEY *rc4_key = &rc4_keys[lid];
/**
* esalt
@ -309,68 +293,81 @@ __kernel void m09710_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
* loop
*/
for (u32 il_pos = 0; il_pos < il_cnt; il_pos++)
for (u32 il_pos = 0; il_pos < il_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] = pw_buf0[0];
wordl0[1] = pw_buf0[1];
wordl0[2] = pw_buf0[2];
wordl0[3] = pw_buf0[3];
wordl1[0] = pw_buf1[0];
wordl1[1] = pw_buf1[1];
wordl1[2] = pw_buf1[2];
wordl1[3] = pw_buf1[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);
}
// first md5 to generate RC4 128 bit key
u32x w0[4];
u32x w1[4];
u32x w2[4];
u32x w3[4];
u32 w0_t[4];
u32 w1_t[4];
u32 w2_t[4];
u32 w3_t[4];
w0[0] = wordl0[0] | wordr0[0];
w0[1] = wordl0[1] | wordr0[1];
w0_t[0] = wordl0[0] | wordr0[0];
w0_t[1] = (wordl0[1] | wordr0[1]) & 0xff;
w0_t[2] = 0x8000;
w0_t[3] = 0;
w1_t[0] = 0;
w1_t[1] = 0;
w1_t[2] = 0;
w1_t[3] = 0;
w2_t[0] = 0;
w2_t[1] = 0;
w2_t[2] = 0;
w2_t[3] = 0;
w3_t[0] = 0;
w3_t[1] = 0;
w3_t[2] = 9 * 8;
w3_t[3] = 0;
/**
* md5
*/
w0[0] = w0[0];
w0[1] = w0[1] & 0xff;
w0[2] = 0x8000;
w0[3] = 0;
w1[0] = 0;
w1[1] = 0;
w1[2] = 0;
w1[3] = 0;
w2[0] = 0;
w2[1] = 0;
w2[2] = 0;
w2[3] = 0;
w3[0] = 0;
w3[1] = 0;
w3[2] = 9 * 8;
w3[3] = 0;
u32 digest[4];
@ -379,55 +376,43 @@ __kernel void m09710_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
digest[2] = MD5M_C;
digest[3] = MD5M_D;
md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
md5_transform (w0, w1, w2, w3, digest);
// now the RC4 part
u32 key[4];
key[0] = digest[0];
key[1] = digest[1];
key[2] = digest[2];
key[3] = digest[3];
rc4_init_16 (rc4_key, key);
rc4_init_16 (rc4_key, digest);
u32 out[4];
u8 j = rc4_next_16 (rc4_key, 0, 0, encryptedVerifier, out);
w0_t[0] = out[0];
w0_t[1] = out[1];
w0_t[2] = out[2];
w0_t[3] = out[3];
w1_t[0] = 0x80;
w1_t[1] = 0;
w1_t[2] = 0;
w1_t[3] = 0;
w2_t[0] = 0;
w2_t[1] = 0;
w2_t[2] = 0;
w2_t[3] = 0;
w3_t[0] = 0;
w3_t[1] = 0;
w3_t[2] = 16 * 8;
w3_t[3] = 0;
w0[0] = out[0];
w0[1] = out[1];
w0[2] = out[2];
w0[3] = out[3];
w1[0] = 0x80;
w1[1] = 0;
w1[2] = 0;
w1[3] = 0;
w2[0] = 0;
w2[1] = 0;
w2[2] = 0;
w2[3] = 0;
w3[0] = 0;
w3[1] = 0;
w3[2] = 16 * 8;
w3[3] = 0;
digest[0] = MD5M_A;
digest[1] = MD5M_B;
digest[2] = MD5M_C;
digest[3] = MD5M_D;
md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
md5_transform (w0, w1, w2, w3, digest);
rc4_next_16 (rc4_key, 16, j, digest, out);
const u32 r0 = out[0];
const u32 r1 = out[1];
const u32 r2 = out[2];
const u32 r3 = out[3];
#include COMPARE_M
COMPARE_M_SIMD (out[0], out[1], out[2], out[3]);
}
}
@ -447,10 +432,6 @@ __kernel void m09710_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
const u32 lid = get_local_id (0);
__local RC4_KEY rc4_keys[64];
__local RC4_KEY *rc4_key = &rc4_keys[lid];
/**
* base
*/
@ -459,52 +440,27 @@ __kernel void m09710_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
if (gid >= gid_max) return;
u32 wordl0[4];
u32 pw_buf0[4];
u32 pw_buf1[4];
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;
pw_buf0[0] = pws[gid].i[0];
pw_buf0[1] = pws[gid].i[1];
pw_buf0[2] = pws[gid].i[2];
pw_buf0[3] = pws[gid].i[3];
pw_buf1[0] = pws[gid].i[4];
pw_buf1[1] = pws[gid].i[5];
pw_buf1[2] = pws[gid].i[6];
pw_buf1[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);
}
/**
* digest
* shared
*/
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]
};
__local RC4_KEY rc4_keys[64];
__local RC4_KEY *rc4_key = &rc4_keys[lid];
/**
* esalt
@ -519,72 +475,97 @@ __kernel void m09710_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
encryptedVerifier[2] = oldoffice01_bufs[salt_pos].encryptedVerifier[2];
encryptedVerifier[3] = oldoffice01_bufs[salt_pos].encryptedVerifier[3];
/**
* 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]
};
/**
* loop
*/
for (u32 il_pos = 0; il_pos < il_cnt; il_pos++)
for (u32 il_pos = 0; il_pos < il_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] = pw_buf0[0];
wordl0[1] = pw_buf0[1];
wordl0[2] = pw_buf0[2];
wordl0[3] = pw_buf0[3];
wordl1[0] = pw_buf1[0];
wordl1[1] = pw_buf1[1];
wordl1[2] = pw_buf1[2];
wordl1[3] = pw_buf1[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);
}
// first md5 to generate RC4 128 bit key
u32x w0[4];
u32x w1[4];
u32x w2[4];
u32x w3[4];
u32 w0_t[4];
u32 w1_t[4];
u32 w2_t[4];
u32 w3_t[4];
w0[0] = wordl0[0] | wordr0[0];
w0[1] = wordl0[1] | wordr0[1];
w0_t[0] = wordl0[0] | wordr0[0];
w0_t[1] = (wordl0[1] | wordr0[1]) & 0xff;
w0_t[2] = 0x8000;
w0_t[3] = 0;
w1_t[0] = 0;
w1_t[1] = 0;
w1_t[2] = 0;
w1_t[3] = 0;
w2_t[0] = 0;
w2_t[1] = 0;
w2_t[2] = 0;
w2_t[3] = 0;
w3_t[0] = 0;
w3_t[1] = 0;
w3_t[2] = 9 * 8;
w3_t[3] = 0;
/**
* md5
*/
w0[0] = w0[0];
w0[1] = w0[1] & 0xff;
w0[2] = 0x8000;
w0[3] = 0;
w1[0] = 0;
w1[1] = 0;
w1[2] = 0;
w1[3] = 0;
w2[0] = 0;
w2[1] = 0;
w2[2] = 0;
w2[3] = 0;
w3[0] = 0;
w3[1] = 0;
w3[2] = 9 * 8;
w3[3] = 0;
u32 digest[4];
@ -593,55 +574,43 @@ __kernel void m09710_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
digest[2] = MD5M_C;
digest[3] = MD5M_D;
md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
md5_transform (w0, w1, w2, w3, digest);
// now the RC4 part
u32 key[4];
key[0] = digest[0];
key[1] = digest[1];
key[2] = digest[2];
key[3] = digest[3];
rc4_init_16 (rc4_key, key);
rc4_init_16 (rc4_key, digest);
u32 out[4];
u8 j = rc4_next_16 (rc4_key, 0, 0, encryptedVerifier, out);
w0_t[0] = out[0];
w0_t[1] = out[1];
w0_t[2] = out[2];
w0_t[3] = out[3];
w1_t[0] = 0x80;
w1_t[1] = 0;
w1_t[2] = 0;
w1_t[3] = 0;
w2_t[0] = 0;
w2_t[1] = 0;
w2_t[2] = 0;
w2_t[3] = 0;
w3_t[0] = 0;
w3_t[1] = 0;
w3_t[2] = 16 * 8;
w3_t[3] = 0;
w0[0] = out[0];
w0[1] = out[1];
w0[2] = out[2];
w0[3] = out[3];
w1[0] = 0x80;
w1[1] = 0;
w1[2] = 0;
w1[3] = 0;
w2[0] = 0;
w2[1] = 0;
w2[2] = 0;
w2[3] = 0;
w3[0] = 0;
w3[1] = 0;
w3[2] = 16 * 8;
w3[3] = 0;
digest[0] = MD5M_A;
digest[1] = MD5M_B;
digest[2] = MD5M_C;
digest[3] = MD5M_D;
md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
md5_transform (w0, w1, w2, w3, digest);
rc4_next_16 (rc4_key, 16, j, digest, out);
const u32 r0 = out[0];
const u32 r1 = out[1];
const u32 r2 = out[2];
const u32 r3 = out[3];
#include COMPARE_S
COMPARE_S_SIMD (out[0], out[1], out[2], out[3]);
}
}

View File

@ -7,6 +7,9 @@
#define _OLDOFFICE01_
//too much register pressure
//#define NEW_SIMD_CODE
#include "include/constants.h"
#include "include/kernel_vendor.h"
@ -246,6 +249,10 @@ static void m09710m (__local RC4_KEY *rc4_keys, u32 w0[4], u32 w1[4], u32 w2[4],
const u32 gid = get_global_id (0);
const u32 lid = get_local_id (0);
/**
* shared
*/
__local RC4_KEY *rc4_key = &rc4_keys[lid];
/**
@ -308,14 +315,7 @@ static void m09710m (__local RC4_KEY *rc4_keys, u32 w0[4], u32 w1[4], u32 w2[4],
// now the RC4 part
u32 key[4];
key[0] = digest[0];
key[1] = digest[1];
key[2] = digest[2];
key[3] = digest[3];
rc4_init_16 (rc4_key, key);
rc4_init_16 (rc4_key, digest);
u32 out[4];
@ -360,19 +360,11 @@ static void m09710s (__local RC4_KEY *rc4_keys, u32 w0[4], u32 w1[4], u32 w2[4],
const u32 gid = get_global_id (0);
const u32 lid = get_local_id (0);
__local RC4_KEY *rc4_key = &rc4_keys[lid];
/**
* digest
* shared
*/
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]
};
__local RC4_KEY *rc4_key = &rc4_keys[lid];
/**
* esalt
@ -387,6 +379,18 @@ static void m09710s (__local RC4_KEY *rc4_keys, u32 w0[4], u32 w1[4], u32 w2[4],
encryptedVerifier[2] = oldoffice01_bufs[salt_pos].encryptedVerifier[2];
encryptedVerifier[3] = oldoffice01_bufs[salt_pos].encryptedVerifier[3];
/**
* 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]
};
/**
* loop
*/
@ -434,14 +438,7 @@ static void m09710s (__local RC4_KEY *rc4_keys, u32 w0[4], u32 w1[4], u32 w2[4],
// now the RC4 part
u32 key[4];
key[0] = digest[0];
key[1] = digest[1];
key[2] = digest[2];
key[3] = digest[3];
rc4_init_16 (rc4_key, key);
rc4_init_16 (rc4_key, digest);
u32 out[4];