1
0
mirror of https://github.com/hashcat/hashcat.git synced 2024-11-01 04:58:57 +00:00
hashcat/OpenCL/m09810_a0-optimized.cl
R. Yushaev 5de004103a Replace kernel parameter lists with macros
Substitute long parameter lists in ~2900 kernel function declarations
with macros. This cleans up the code, reduces probability of copy-paste
errors and highlights the differences between kernel functions. Also
reduces the size of the OpenCL folder by ~3 MB.
2018-11-16 11:44:33 +01:00

407 lines
8.3 KiB
Common Lisp

/**
* Author......: See docs/credits.txt
* License.....: MIT
*/
//too much register pressure
//#define NEW_SIMD_CODE
#include "inc_vendor.cl"
#include "inc_hash_constants.h"
#include "inc_hash_functions.cl"
#include "inc_types.cl"
#include "inc_common.cl"
#include "inc_rp_optimized.h"
#include "inc_rp_optimized.cl"
#include "inc_simd.cl"
#include "inc_hash_sha1.cl"
typedef struct
{
u8 S[256];
u32 wtf_its_faster;
} RC4_KEY;
DECLSPEC void swap (__local RC4_KEY *rc4_key, const u8 i, const u8 j)
{
u8 tmp;
tmp = rc4_key->S[i];
rc4_key->S[i] = rc4_key->S[j];
rc4_key->S[j] = tmp;
}
DECLSPEC void rc4_init_16 (__local RC4_KEY *rc4_key, const u32 *data)
{
u32 v = 0x03020100;
u32 a = 0x04040404;
__local u32 *ptr = (__local u32 *) rc4_key->S;
#ifdef _unroll
#pragma unroll
#endif
for (u32 i = 0; i < 64; i++)
{
*ptr++ = v; v += a;
}
u32 j = 0;
for (u32 i = 0; i < 16; i++)
{
u32 idx = i * 16;
u32 v;
v = data[0];
j += rc4_key->S[idx] + (v >> 0); swap (rc4_key, idx, j); idx++;
j += rc4_key->S[idx] + (v >> 8); swap (rc4_key, idx, j); idx++;
j += rc4_key->S[idx] + (v >> 16); swap (rc4_key, idx, j); idx++;
j += rc4_key->S[idx] + (v >> 24); swap (rc4_key, idx, j); idx++;
v = data[1];
j += rc4_key->S[idx] + (v >> 0); swap (rc4_key, idx, j); idx++;
j += rc4_key->S[idx] + (v >> 8); swap (rc4_key, idx, j); idx++;
j += rc4_key->S[idx] + (v >> 16); swap (rc4_key, idx, j); idx++;
j += rc4_key->S[idx] + (v >> 24); swap (rc4_key, idx, j); idx++;
v = data[2];
j += rc4_key->S[idx] + (v >> 0); swap (rc4_key, idx, j); idx++;
j += rc4_key->S[idx] + (v >> 8); swap (rc4_key, idx, j); idx++;
j += rc4_key->S[idx] + (v >> 16); swap (rc4_key, idx, j); idx++;
j += rc4_key->S[idx] + (v >> 24); swap (rc4_key, idx, j); idx++;
v = data[3];
j += rc4_key->S[idx] + (v >> 0); swap (rc4_key, idx, j); idx++;
j += rc4_key->S[idx] + (v >> 8); swap (rc4_key, idx, j); idx++;
j += rc4_key->S[idx] + (v >> 16); swap (rc4_key, idx, j); idx++;
j += rc4_key->S[idx] + (v >> 24); swap (rc4_key, idx, j); idx++;
}
}
DECLSPEC u8 rc4_next_16 (__local RC4_KEY *rc4_key, u8 i, u8 j, const u32 *in, u32 *out)
{
#ifdef _unroll
#pragma unroll
#endif
for (u32 k = 0; k < 4; k++)
{
u32 xor4 = 0;
u8 idx;
i += 1;
j += rc4_key->S[i];
swap (rc4_key, i, j);
idx = rc4_key->S[i] + rc4_key->S[j];
xor4 |= rc4_key->S[idx] << 0;
i += 1;
j += rc4_key->S[i];
swap (rc4_key, i, j);
idx = rc4_key->S[i] + rc4_key->S[j];
xor4 |= rc4_key->S[idx] << 8;
i += 1;
j += rc4_key->S[i];
swap (rc4_key, i, j);
idx = rc4_key->S[i] + rc4_key->S[j];
xor4 |= rc4_key->S[idx] << 16;
i += 1;
j += rc4_key->S[i];
swap (rc4_key, i, j);
idx = rc4_key->S[i] + rc4_key->S[j];
xor4 |= rc4_key->S[idx] << 24;
out[k] = in[k] ^ xor4;
}
return j;
}
__kernel void __attribute__((reqd_work_group_size(64, 1, 1))) m09810_m04 (KERN_ATTR_RULES_ESALT (oldoffice34_t))
{
/**
* modifier
*/
const u64 lid = get_local_id (0);
/**
* base
*/
const u64 gid = get_global_id (0);
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];
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_len = pws[gid].pw_len;
/**
* shared
*/
__local RC4_KEY rc4_keys[64];
__local RC4_KEY *rc4_key = &rc4_keys[lid];
/**
* esalt
*/
const u32 version = esalt_bufs[digests_offset].version;
u32 encryptedVerifier[4];
encryptedVerifier[0] = esalt_bufs[digests_offset].encryptedVerifier[0];
encryptedVerifier[1] = esalt_bufs[digests_offset].encryptedVerifier[1];
encryptedVerifier[2] = esalt_bufs[digests_offset].encryptedVerifier[2];
encryptedVerifier[3] = esalt_bufs[digests_offset].encryptedVerifier[3];
/**
* loop
*/
for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
{
u32x w0[4] = { 0 };
u32x w1[4] = { 0 };
u32x w2[4] = { 0 };
u32x w3[4] = { 0 };
apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
/**
* RC4 + SHA1
*/
u32 key[4];
key[0] = w0[0];
key[1] = w0[1] & 0xff;
key[2] = 0;
key[3] = 0;
rc4_init_16 (rc4_key, key);
u32 out[4];
u8 j = rc4_next_16 (rc4_key, 0, 0, encryptedVerifier, out);
w0[0] = swap32 (out[0]);
w0[1] = swap32 (out[1]);
w0[2] = swap32 (out[2]);
w0[3] = swap32 (out[3]);
w1[0] = 0x80000000;
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] = 0;
w3[3] = 16 * 8;
u32 digest[5];
digest[0] = SHA1M_A;
digest[1] = SHA1M_B;
digest[2] = SHA1M_C;
digest[3] = SHA1M_D;
digest[4] = SHA1M_E;
sha1_transform (w0, w1, w2, w3, digest);
digest[0] = swap32_S (digest[0]);
digest[1] = swap32_S (digest[1]);
digest[2] = swap32_S (digest[2]);
digest[3] = swap32_S (digest[3]);
rc4_next_16 (rc4_key, 16, j, digest, out);
COMPARE_M_SIMD (out[0], out[1], out[2], out[3]);
}
}
__kernel void __attribute__((reqd_work_group_size(64, 1, 1))) m09810_m08 (KERN_ATTR_RULES_ESALT (oldoffice34_t))
{
}
__kernel void __attribute__((reqd_work_group_size(64, 1, 1))) m09810_m16 (KERN_ATTR_RULES_ESALT (oldoffice34_t))
{
}
__kernel void __attribute__((reqd_work_group_size(64, 1, 1))) m09810_s04 (KERN_ATTR_RULES_ESALT (oldoffice34_t))
{
/**
* modifier
*/
const u64 lid = get_local_id (0);
/**
* base
*/
const u64 gid = get_global_id (0);
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];
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_len = pws[gid].pw_len;
/**
* shared
*/
__local RC4_KEY rc4_keys[64];
__local RC4_KEY *rc4_key = &rc4_keys[lid];
/**
* esalt
*/
const u32 version = esalt_bufs[digests_offset].version;
u32 encryptedVerifier[4];
encryptedVerifier[0] = esalt_bufs[digests_offset].encryptedVerifier[0];
encryptedVerifier[1] = esalt_bufs[digests_offset].encryptedVerifier[1];
encryptedVerifier[2] = esalt_bufs[digests_offset].encryptedVerifier[2];
encryptedVerifier[3] = esalt_bufs[digests_offset].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 += VECT_SIZE)
{
u32x w0[4] = { 0 };
u32x w1[4] = { 0 };
u32x w2[4] = { 0 };
u32x w3[4] = { 0 };
apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
/**
* RC4 + SHA1
*/
u32 key[4];
key[0] = w0[0];
key[1] = w0[1] & 0xff;
key[2] = 0;
key[3] = 0;
rc4_init_16 (rc4_key, key);
u32 out[4];
u8 j = rc4_next_16 (rc4_key, 0, 0, encryptedVerifier, out);
w0[0] = swap32 (out[0]);
w0[1] = swap32 (out[1]);
w0[2] = swap32 (out[2]);
w0[3] = swap32 (out[3]);
w1[0] = 0x80000000;
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] = 0;
w3[3] = 16 * 8;
u32 digest[5];
digest[0] = SHA1M_A;
digest[1] = SHA1M_B;
digest[2] = SHA1M_C;
digest[3] = SHA1M_D;
digest[4] = SHA1M_E;
sha1_transform (w0, w1, w2, w3, digest);
digest[0] = swap32_S (digest[0]);
digest[1] = swap32_S (digest[1]);
digest[2] = swap32_S (digest[2]);
digest[3] = swap32_S (digest[3]);
rc4_next_16 (rc4_key, 16, j, digest, out);
COMPARE_S_SIMD (out[0], out[1], out[2], out[3]);
}
}
__kernel void __attribute__((reqd_work_group_size(64, 1, 1))) m09810_s08 (KERN_ATTR_RULES_ESALT (oldoffice34_t))
{
}
__kernel void __attribute__((reqd_work_group_size(64, 1, 1))) m09810_s16 (KERN_ATTR_RULES_ESALT (oldoffice34_t))
{
}