mirror of
https://github.com/hashcat/hashcat.git
synced 2024-11-27 02:18:21 +00:00
563 lines
14 KiB
Common Lisp
563 lines
14 KiB
Common Lisp
/**
|
|
* Author......: See docs/credits.txt
|
|
* License.....: MIT
|
|
*/
|
|
|
|
//incompatible data-dependant code
|
|
//#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"
|
|
|
|
__constant u32a theMagicArray[64] =
|
|
{
|
|
0x91ac5114, 0x9f675443, 0x24e73be0, 0x28747bc2, 0x863313eb, 0x5a4fcb5c, 0x080a7337, 0x0e5d1c2f,
|
|
0x338fe6e5, 0xf89baedd, 0x16f24b8d, 0x2ce1d4dc, 0xb0cbdf9d, 0xd4706d17, 0xf94d423f, 0x9b1b1194,
|
|
0x9f5bc19b, 0x06059d03, 0x9d5e138a, 0x1e9a6ae8, 0xd97c1417, 0x58c72af6, 0xa199630a, 0xd7fd70c3,
|
|
0xf65e7413, 0x03c90b04, 0x2698f726, 0x8a929325, 0xb0a20d23, 0xed63796d, 0x1332fa3c, 0x35029aa3,
|
|
0xb3dd8e0a, 0x24bf51c3, 0x7ccd559f, 0x37af944c, 0x29085282, 0xb23b4e37, 0x9f170791, 0x113bfdcd,
|
|
0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000,
|
|
0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000,
|
|
0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000,
|
|
};
|
|
|
|
DECLSPEC u32 GETSHIFTEDINT_CONST (__constant u32 *a, const int n)
|
|
{
|
|
const int d = n / 4;
|
|
const int m = n & 3;
|
|
|
|
u64 tmp = hl32_to_64_S (a[d + 0], a[d + 1]);
|
|
|
|
tmp <<= m * 8;
|
|
|
|
return h32_from_64_S (tmp);
|
|
}
|
|
|
|
DECLSPEC void SETSHIFTEDINT (u32 *a, const int n, const u32 v)
|
|
{
|
|
const int d = n / 4;
|
|
const int m = n & 3;
|
|
|
|
u64 tmp = hl32_to_64_S (v, 0);
|
|
|
|
tmp >>= m * 8;
|
|
|
|
a[d + 0] |= h32_from_64_S (tmp);
|
|
a[d + 1] = l32_from_64_S (tmp);
|
|
}
|
|
|
|
__kernel void m07801_m04 (KERN_ATTR_RULES ())
|
|
{
|
|
/**
|
|
* 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 & 63;
|
|
|
|
/**
|
|
* salt
|
|
*/
|
|
|
|
u32 salt_buf[8];
|
|
|
|
salt_buf[0] = salt_bufs[salt_pos].salt_buf[0];
|
|
salt_buf[1] = salt_bufs[salt_pos].salt_buf[1];
|
|
salt_buf[2] = salt_bufs[salt_pos].salt_buf[2];
|
|
salt_buf[3] = salt_bufs[salt_pos].salt_buf[3];
|
|
salt_buf[4] = salt_bufs[salt_pos].salt_buf[4];
|
|
salt_buf[5] = salt_bufs[salt_pos].salt_buf[5];
|
|
salt_buf[6] = salt_bufs[salt_pos].salt_buf[6];
|
|
salt_buf[7] = salt_bufs[salt_pos].salt_buf[7];
|
|
|
|
const u32 salt_len = salt_bufs[salt_pos].salt_len;
|
|
|
|
/**
|
|
* 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 };
|
|
|
|
const u32x out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
|
|
|
|
/**
|
|
* SAP
|
|
*/
|
|
|
|
u32 s0[4];
|
|
u32 s1[4];
|
|
u32 s2[4];
|
|
u32 s3[4];
|
|
|
|
s0[0] = salt_buf[0];
|
|
s0[1] = salt_buf[1];
|
|
s0[2] = salt_buf[2];
|
|
s0[3] = salt_buf[3];
|
|
s1[0] = salt_buf[4];
|
|
s1[1] = salt_buf[5];
|
|
s1[2] = salt_buf[6];
|
|
s1[3] = salt_buf[7];
|
|
s2[0] = 0;
|
|
s2[1] = 0;
|
|
s2[2] = 0;
|
|
s2[3] = 0;
|
|
s3[0] = 0;
|
|
s3[1] = 0;
|
|
s3[2] = 0;
|
|
s3[3] = 0;
|
|
|
|
switch_buffer_by_offset_le_VV (s0, s1, s2, s3, out_len);
|
|
|
|
const u32x pw_salt_len = out_len + salt_len;
|
|
|
|
/**
|
|
* sha1
|
|
*/
|
|
|
|
u32 final[32];
|
|
|
|
final[ 0] = swap32_S (w0[0] | s0[0]);
|
|
final[ 1] = swap32_S (w0[1] | s0[1]);
|
|
final[ 2] = swap32_S (w0[2] | s0[2]);
|
|
final[ 3] = swap32_S (w0[3] | s0[3]);
|
|
final[ 4] = swap32_S (w1[0] | s1[0]);
|
|
final[ 5] = swap32_S (w1[1] | s1[1]);
|
|
final[ 6] = swap32_S (w1[2] | s1[2]);
|
|
final[ 7] = swap32_S (w1[3] | s1[3]);
|
|
final[ 8] = swap32_S (w2[0] | s2[0]);
|
|
final[ 9] = swap32_S (w2[1] | s2[1]);
|
|
final[10] = swap32_S (w2[2] | s2[2]);
|
|
final[11] = swap32_S (w2[3] | s2[3]);
|
|
final[12] = swap32_S (w3[0] | s3[0]);
|
|
final[13] = swap32_S (w3[1] | s3[1]);
|
|
final[14] = 0;
|
|
final[15] = pw_salt_len * 8;
|
|
final[16] = 0;
|
|
final[17] = 0;
|
|
final[18] = 0;
|
|
final[19] = 0;
|
|
final[20] = 0;
|
|
final[21] = 0;
|
|
final[22] = 0;
|
|
final[23] = 0;
|
|
final[24] = 0;
|
|
final[25] = 0;
|
|
final[26] = 0;
|
|
final[27] = 0;
|
|
final[28] = 0;
|
|
final[29] = 0;
|
|
final[30] = 0;
|
|
final[31] = 0;
|
|
|
|
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 (&final[0], &final[4], &final[8], &final[12], digest);
|
|
|
|
// prepare magic array range
|
|
|
|
u32 lengthMagicArray = 0x20;
|
|
u32 offsetMagicArray = 0;
|
|
|
|
lengthMagicArray += ((digest[0] >> 24) & 0xff) % 6;
|
|
lengthMagicArray += ((digest[0] >> 16) & 0xff) % 6;
|
|
lengthMagicArray += ((digest[0] >> 8) & 0xff) % 6;
|
|
lengthMagicArray += ((digest[0] >> 0) & 0xff) % 6;
|
|
lengthMagicArray += ((digest[1] >> 24) & 0xff) % 6;
|
|
lengthMagicArray += ((digest[1] >> 16) & 0xff) % 6;
|
|
lengthMagicArray += ((digest[1] >> 8) & 0xff) % 6;
|
|
lengthMagicArray += ((digest[1] >> 0) & 0xff) % 6;
|
|
lengthMagicArray += ((digest[2] >> 24) & 0xff) % 6;
|
|
lengthMagicArray += ((digest[2] >> 16) & 0xff) % 6;
|
|
offsetMagicArray += ((digest[2] >> 8) & 0xff) % 8;
|
|
offsetMagicArray += ((digest[2] >> 0) & 0xff) % 8;
|
|
offsetMagicArray += ((digest[3] >> 24) & 0xff) % 8;
|
|
offsetMagicArray += ((digest[3] >> 16) & 0xff) % 8;
|
|
offsetMagicArray += ((digest[3] >> 8) & 0xff) % 8;
|
|
offsetMagicArray += ((digest[3] >> 0) & 0xff) % 8;
|
|
offsetMagicArray += ((digest[4] >> 24) & 0xff) % 8;
|
|
offsetMagicArray += ((digest[4] >> 16) & 0xff) % 8;
|
|
offsetMagicArray += ((digest[4] >> 8) & 0xff) % 8;
|
|
offsetMagicArray += ((digest[4] >> 0) & 0xff) % 8;
|
|
|
|
// final
|
|
|
|
digest[0] = SHA1M_A;
|
|
digest[1] = SHA1M_B;
|
|
digest[2] = SHA1M_C;
|
|
digest[3] = SHA1M_D;
|
|
digest[4] = SHA1M_E;
|
|
|
|
final[ 0] = swap32_S (w0[0]);
|
|
final[ 1] = swap32_S (w0[1]);
|
|
final[ 2] = swap32_S (w0[2]);
|
|
final[ 3] = swap32_S (w0[3]);
|
|
final[ 4] = swap32_S (w1[0]);
|
|
final[ 5] = swap32_S (w1[1]);
|
|
final[ 6] = swap32_S (w1[2]);
|
|
final[ 7] = swap32_S (w1[3]);
|
|
final[ 8] = 0;
|
|
final[ 9] = 0;
|
|
final[10] = 0;
|
|
final[11] = 0;
|
|
final[12] = 0;
|
|
final[13] = 0;
|
|
final[14] = 0;
|
|
final[15] = 0;
|
|
|
|
u32 final_len = pw_len;
|
|
|
|
u32 i;
|
|
|
|
// append MagicArray
|
|
|
|
for (i = 0; i < lengthMagicArray - 4; i += 4)
|
|
{
|
|
const u32 tmp = GETSHIFTEDINT_CONST (theMagicArray, offsetMagicArray + i);
|
|
|
|
SETSHIFTEDINT (final, final_len + i, tmp);
|
|
}
|
|
|
|
const u32 mask = 0xffffffff << (((4 - (lengthMagicArray - i)) & 3) * 8);
|
|
|
|
const u32 tmp = GETSHIFTEDINT_CONST (theMagicArray, offsetMagicArray + i) & mask;
|
|
|
|
SETSHIFTEDINT (final, final_len + i, tmp);
|
|
|
|
final_len += lengthMagicArray;
|
|
|
|
// append Salt
|
|
|
|
for (i = 0; i < salt_len + 1; i += 4) // +1 for the 0x80
|
|
{
|
|
const u32 tmp = swap32_S (salt_buf[i / 4]); // attention, int[] not char[]
|
|
|
|
SETSHIFTEDINT (final, final_len + i, tmp);
|
|
}
|
|
|
|
final_len += salt_len;
|
|
|
|
// calculate
|
|
|
|
int left;
|
|
int off;
|
|
|
|
for (left = final_len, off = 0; left >= 56; left -= 64, off += 16)
|
|
{
|
|
sha1_transform (&final[off + 0], &final[off + 4], &final[off + 8], &final[off + 12], digest);
|
|
}
|
|
|
|
final[off + 15] = final_len * 8;
|
|
|
|
sha1_transform (&final[off + 0], &final[off + 4], &final[off + 8], &final[off + 12], digest);
|
|
|
|
COMPARE_M_SIMD (0, 0, digest[2] & 0xffff0000, digest[1]);
|
|
}
|
|
}
|
|
|
|
__kernel void m07801_m08 (KERN_ATTR_RULES ())
|
|
{
|
|
}
|
|
|
|
__kernel void m07801_m16 (KERN_ATTR_RULES ())
|
|
{
|
|
}
|
|
|
|
__kernel void m07801_s04 (KERN_ATTR_RULES ())
|
|
{
|
|
/**
|
|
* 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 & 63;
|
|
|
|
/**
|
|
* salt
|
|
*/
|
|
|
|
u32 salt_buf[8];
|
|
|
|
salt_buf[0] = salt_bufs[salt_pos].salt_buf[0];
|
|
salt_buf[1] = salt_bufs[salt_pos].salt_buf[1];
|
|
salt_buf[2] = salt_bufs[salt_pos].salt_buf[2];
|
|
salt_buf[3] = salt_bufs[salt_pos].salt_buf[3];
|
|
salt_buf[4] = salt_bufs[salt_pos].salt_buf[4];
|
|
salt_buf[5] = salt_bufs[salt_pos].salt_buf[5];
|
|
salt_buf[6] = salt_bufs[salt_pos].salt_buf[6];
|
|
salt_buf[7] = salt_bufs[salt_pos].salt_buf[7];
|
|
|
|
const u32 salt_len = salt_bufs[salt_pos].salt_len;
|
|
|
|
/**
|
|
* 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 };
|
|
|
|
const u32x out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
|
|
|
|
/**
|
|
* SAP
|
|
*/
|
|
|
|
u32 s0[4];
|
|
u32 s1[4];
|
|
u32 s2[4];
|
|
u32 s3[4];
|
|
|
|
s0[0] = salt_buf[0];
|
|
s0[1] = salt_buf[1];
|
|
s0[2] = salt_buf[2];
|
|
s0[3] = salt_buf[3];
|
|
s1[0] = salt_buf[4];
|
|
s1[1] = salt_buf[5];
|
|
s1[2] = salt_buf[6];
|
|
s1[3] = salt_buf[7];
|
|
s2[0] = 0;
|
|
s2[1] = 0;
|
|
s2[2] = 0;
|
|
s2[3] = 0;
|
|
s3[0] = 0;
|
|
s3[1] = 0;
|
|
s3[2] = 0;
|
|
s3[3] = 0;
|
|
|
|
switch_buffer_by_offset_le_VV (s0, s1, s2, s3, out_len);
|
|
|
|
const u32x pw_salt_len = out_len + salt_len;
|
|
|
|
/**
|
|
* sha1
|
|
*/
|
|
|
|
u32 final[32];
|
|
|
|
final[ 0] = swap32_S (w0[0] | s0[0]);
|
|
final[ 1] = swap32_S (w0[1] | s0[1]);
|
|
final[ 2] = swap32_S (w0[2] | s0[2]);
|
|
final[ 3] = swap32_S (w0[3] | s0[3]);
|
|
final[ 4] = swap32_S (w1[0] | s1[0]);
|
|
final[ 5] = swap32_S (w1[1] | s1[1]);
|
|
final[ 6] = swap32_S (w1[2] | s1[2]);
|
|
final[ 7] = swap32_S (w1[3] | s1[3]);
|
|
final[ 8] = swap32_S (w2[0] | s2[0]);
|
|
final[ 9] = swap32_S (w2[1] | s2[1]);
|
|
final[10] = swap32_S (w2[2] | s2[2]);
|
|
final[11] = swap32_S (w2[3] | s2[3]);
|
|
final[12] = swap32_S (w3[0] | s3[0]);
|
|
final[13] = swap32_S (w3[1] | s3[1]);
|
|
final[14] = 0;
|
|
final[15] = pw_salt_len * 8;
|
|
final[16] = 0;
|
|
final[17] = 0;
|
|
final[18] = 0;
|
|
final[19] = 0;
|
|
final[20] = 0;
|
|
final[21] = 0;
|
|
final[22] = 0;
|
|
final[23] = 0;
|
|
final[24] = 0;
|
|
final[25] = 0;
|
|
final[26] = 0;
|
|
final[27] = 0;
|
|
final[28] = 0;
|
|
final[29] = 0;
|
|
final[30] = 0;
|
|
final[31] = 0;
|
|
|
|
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 (&final[0], &final[4], &final[8], &final[12], digest);
|
|
|
|
// prepare magic array range
|
|
|
|
u32 lengthMagicArray = 0x20;
|
|
u32 offsetMagicArray = 0;
|
|
|
|
lengthMagicArray += ((digest[0] >> 24) & 0xff) % 6;
|
|
lengthMagicArray += ((digest[0] >> 16) & 0xff) % 6;
|
|
lengthMagicArray += ((digest[0] >> 8) & 0xff) % 6;
|
|
lengthMagicArray += ((digest[0] >> 0) & 0xff) % 6;
|
|
lengthMagicArray += ((digest[1] >> 24) & 0xff) % 6;
|
|
lengthMagicArray += ((digest[1] >> 16) & 0xff) % 6;
|
|
lengthMagicArray += ((digest[1] >> 8) & 0xff) % 6;
|
|
lengthMagicArray += ((digest[1] >> 0) & 0xff) % 6;
|
|
lengthMagicArray += ((digest[2] >> 24) & 0xff) % 6;
|
|
lengthMagicArray += ((digest[2] >> 16) & 0xff) % 6;
|
|
offsetMagicArray += ((digest[2] >> 8) & 0xff) % 8;
|
|
offsetMagicArray += ((digest[2] >> 0) & 0xff) % 8;
|
|
offsetMagicArray += ((digest[3] >> 24) & 0xff) % 8;
|
|
offsetMagicArray += ((digest[3] >> 16) & 0xff) % 8;
|
|
offsetMagicArray += ((digest[3] >> 8) & 0xff) % 8;
|
|
offsetMagicArray += ((digest[3] >> 0) & 0xff) % 8;
|
|
offsetMagicArray += ((digest[4] >> 24) & 0xff) % 8;
|
|
offsetMagicArray += ((digest[4] >> 16) & 0xff) % 8;
|
|
offsetMagicArray += ((digest[4] >> 8) & 0xff) % 8;
|
|
offsetMagicArray += ((digest[4] >> 0) & 0xff) % 8;
|
|
|
|
// final
|
|
|
|
digest[0] = SHA1M_A;
|
|
digest[1] = SHA1M_B;
|
|
digest[2] = SHA1M_C;
|
|
digest[3] = SHA1M_D;
|
|
digest[4] = SHA1M_E;
|
|
|
|
final[ 0] = swap32_S (w0[0]);
|
|
final[ 1] = swap32_S (w0[1]);
|
|
final[ 2] = swap32_S (w0[2]);
|
|
final[ 3] = swap32_S (w0[3]);
|
|
final[ 4] = swap32_S (w1[0]);
|
|
final[ 5] = swap32_S (w1[1]);
|
|
final[ 6] = swap32_S (w1[2]);
|
|
final[ 7] = swap32_S (w1[3]);
|
|
final[ 8] = 0;
|
|
final[ 9] = 0;
|
|
final[10] = 0;
|
|
final[11] = 0;
|
|
final[12] = 0;
|
|
final[13] = 0;
|
|
final[14] = 0;
|
|
final[15] = 0;
|
|
|
|
u32 final_len = pw_len;
|
|
|
|
u32 i;
|
|
|
|
// append MagicArray
|
|
|
|
for (i = 0; i < lengthMagicArray - 4; i += 4)
|
|
{
|
|
const u32 tmp = GETSHIFTEDINT_CONST (theMagicArray, offsetMagicArray + i);
|
|
|
|
SETSHIFTEDINT (final, final_len + i, tmp);
|
|
}
|
|
|
|
const u32 mask = 0xffffffff << (((4 - (lengthMagicArray - i)) & 3) * 8);
|
|
|
|
const u32 tmp = GETSHIFTEDINT_CONST (theMagicArray, offsetMagicArray + i) & mask;
|
|
|
|
SETSHIFTEDINT (final, final_len + i, tmp);
|
|
|
|
final_len += lengthMagicArray;
|
|
|
|
// append Salt
|
|
|
|
for (i = 0; i < salt_len + 1; i += 4) // +1 for the 0x80
|
|
{
|
|
const u32 tmp = swap32_S (salt_buf[i / 4]); // attention, int[] not char[]
|
|
|
|
SETSHIFTEDINT (final, final_len + i, tmp);
|
|
}
|
|
|
|
final_len += salt_len;
|
|
|
|
// calculate
|
|
|
|
int left;
|
|
int off;
|
|
|
|
for (left = final_len, off = 0; left >= 56; left -= 64, off += 16)
|
|
{
|
|
sha1_transform (&final[off + 0], &final[off + 4], &final[off + 8], &final[off + 12], digest);
|
|
}
|
|
|
|
final[off + 15] = final_len * 8;
|
|
|
|
sha1_transform (&final[off + 0], &final[off + 4], &final[off + 8], &final[off + 12], digest);
|
|
|
|
COMPARE_S_SIMD (0, 0, digest[2] & 0xffff0000, digest[1]);
|
|
}
|
|
}
|
|
|
|
__kernel void m07801_s08 (KERN_ATTR_RULES ())
|
|
{
|
|
}
|
|
|
|
__kernel void m07801_s16 (KERN_ATTR_RULES ())
|
|
{
|
|
}
|