1
0
mirror of https://github.com/hashcat/hashcat.git synced 2024-12-27 17:08:12 +00:00
hashcat/OpenCL/m07500_a3-pure.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

456 lines
9.9 KiB
Common Lisp

/**
* Author......: See docs/credits.txt
* License.....: MIT
*/
//shared mem too small
//#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_hash_md4.cl"
#include "inc_hash_md5.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;
}
DECLSPEC int decrypt_and_check (__local RC4_KEY *rc4_key, u32 *data, u32 *timestamp_ct)
{
rc4_init_16 (rc4_key, data);
u32 out[4];
u8 j = 0;
j = rc4_next_16 (rc4_key, 0, j, timestamp_ct + 0, out);
if ((out[3] & 0xffff0000) != 0x30320000) return 0;
j = rc4_next_16 (rc4_key, 16, j, timestamp_ct + 4, out);
if (((out[0] & 0xff) < '0') || ((out[0] & 0xff) > '9')) return 0; out[0] >>= 8;
if (((out[0] & 0xff) < '0') || ((out[0] & 0xff) > '9')) return 0; out[0] >>= 8;
if (((out[0] & 0xff) < '0') || ((out[0] & 0xff) > '9')) return 0; out[0] >>= 8;
if (((out[0] & 0xff) < '0') || ((out[0] & 0xff) > '9')) return 0;
if (((out[1] & 0xff) < '0') || ((out[1] & 0xff) > '9')) return 0; out[1] >>= 8;
if (((out[1] & 0xff) < '0') || ((out[1] & 0xff) > '9')) return 0; out[1] >>= 8;
if (((out[1] & 0xff) < '0') || ((out[1] & 0xff) > '9')) return 0; out[1] >>= 8;
if (((out[1] & 0xff) < '0') || ((out[1] & 0xff) > '9')) return 0;
if (((out[2] & 0xff) < '0') || ((out[2] & 0xff) > '9')) return 0; out[2] >>= 8;
if (((out[2] & 0xff) < '0') || ((out[2] & 0xff) > '9')) return 0; out[2] >>= 8;
if (((out[2] & 0xff) < '0') || ((out[2] & 0xff) > '9')) return 0; out[2] >>= 8;
if (((out[2] & 0xff) < '0') || ((out[2] & 0xff) > '9')) return 0;
return 1;
}
DECLSPEC void kerb_prepare (const u32 *K, const u32 *checksum, u32 *digest)
{
// K1=MD5_HMAC(K,1); with 1 encoded as little indian on 4 bytes (01000000 in hexa);
u32 w0[4];
u32 w1[4];
u32 w2[4];
u32 w3[4];
w0[0] = K[0];
w0[1] = K[1];
w0[2] = K[2];
w0[3] = K[3];
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] = 0;
w3[3] = 0;
md5_hmac_ctx_t ctx1;
md5_hmac_init_64 (&ctx1, w0, w1, w2, w3);
w0[0] = 1;
w0[1] = 0;
w0[2] = 0;
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] = 0;
w3[3] = 0;
md5_hmac_update_64 (&ctx1, w0, w1, w2, w3, 4);
md5_hmac_final (&ctx1);
w0[0] = ctx1.opad.h[0];
w0[1] = ctx1.opad.h[1];
w0[2] = ctx1.opad.h[2];
w0[3] = ctx1.opad.h[3];
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] = 0;
w3[3] = 0;
md5_hmac_ctx_t ctx;
md5_hmac_init_64 (&ctx, w0, w1, w2, w3);
w0[0] = checksum[0];
w0[1] = checksum[1];
w0[2] = checksum[2];
w0[3] = checksum[3];
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] = 0;
w3[3] = 0;
md5_hmac_update_64 (&ctx, w0, w1, w2, w3, 16);
md5_hmac_final (&ctx);
digest[0] = ctx.opad.h[0];
digest[1] = ctx.opad.h[1];
digest[2] = ctx.opad.h[2];
digest[3] = ctx.opad.h[3];
}
__kernel void __attribute__((reqd_work_group_size(64, 1, 1))) m07500_mxx (KERN_ATTR_VECTOR_ESALT (krb5pa_t))
{
/**
* modifier
*/
const u64 lid = get_local_id (0);
const u64 gid = get_global_id (0);
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 (int i = 0, idx = 0; i < pw_len; i += 4, idx += 1)
{
w[idx] = pws[gid].i[idx];
}
__local RC4_KEY rc4_keys[64];
__local RC4_KEY *rc4_key = &rc4_keys[lid];
u32 checksum[4];
checksum[0] = esalt_bufs[digests_offset].checksum[0];
checksum[1] = esalt_bufs[digests_offset].checksum[1];
checksum[2] = esalt_bufs[digests_offset].checksum[2];
checksum[3] = esalt_bufs[digests_offset].checksum[3];
u32 timestamp_ct[8];
timestamp_ct[0] = esalt_bufs[digests_offset].timestamp[0];
timestamp_ct[1] = esalt_bufs[digests_offset].timestamp[1];
timestamp_ct[2] = esalt_bufs[digests_offset].timestamp[2];
timestamp_ct[3] = esalt_bufs[digests_offset].timestamp[3];
timestamp_ct[4] = esalt_bufs[digests_offset].timestamp[4];
timestamp_ct[5] = esalt_bufs[digests_offset].timestamp[5];
timestamp_ct[6] = esalt_bufs[digests_offset].timestamp[6];
timestamp_ct[7] = esalt_bufs[digests_offset].timestamp[7];
/**
* 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;
md4_ctx_t ctx;
md4_init (&ctx);
md4_update_utf16le (&ctx, w, pw_len);
md4_final (&ctx);
u32 digest[4];
kerb_prepare (ctx.h, checksum, digest);
if (decrypt_and_check (rc4_key, digest, timestamp_ct) == 1)
{
if (atomic_inc (&hashes_shown[digests_offset]) == 0)
{
mark_hash (plains_buf, d_return_buf, salt_pos, digests_cnt, 0, digests_offset + 0, gid, il_pos);
}
}
}
}
__kernel void __attribute__((reqd_work_group_size(64, 1, 1))) m07500_sxx (KERN_ATTR_VECTOR_ESALT (krb5pa_t))
{
/**
* modifier
*/
const u64 lid = get_local_id (0);
const u64 gid = get_global_id (0);
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 (int i = 0, idx = 0; i < pw_len; i += 4, idx += 1)
{
w[idx] = pws[gid].i[idx];
}
__local RC4_KEY rc4_keys[64];
__local RC4_KEY *rc4_key = &rc4_keys[lid];
u32 checksum[4];
checksum[0] = esalt_bufs[digests_offset].checksum[0];
checksum[1] = esalt_bufs[digests_offset].checksum[1];
checksum[2] = esalt_bufs[digests_offset].checksum[2];
checksum[3] = esalt_bufs[digests_offset].checksum[3];
u32 timestamp_ct[8];
timestamp_ct[0] = esalt_bufs[digests_offset].timestamp[0];
timestamp_ct[1] = esalt_bufs[digests_offset].timestamp[1];
timestamp_ct[2] = esalt_bufs[digests_offset].timestamp[2];
timestamp_ct[3] = esalt_bufs[digests_offset].timestamp[3];
timestamp_ct[4] = esalt_bufs[digests_offset].timestamp[4];
timestamp_ct[5] = esalt_bufs[digests_offset].timestamp[5];
timestamp_ct[6] = esalt_bufs[digests_offset].timestamp[6];
timestamp_ct[7] = esalt_bufs[digests_offset].timestamp[7];
/**
* 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;
md4_ctx_t ctx;
md4_init (&ctx);
md4_update_utf16le (&ctx, w, pw_len);
md4_final (&ctx);
u32 digest[4];
kerb_prepare (ctx.h, checksum, digest);
if (decrypt_and_check (rc4_key, digest, timestamp_ct) == 1)
{
if (atomic_inc (&hashes_shown[digests_offset]) == 0)
{
mark_hash (plains_buf, d_return_buf, salt_pos, digests_cnt, 0, digests_offset + 0, gid, il_pos);
}
}
}
}