1
0
mirror of https://github.com/hashcat/hashcat.git synced 2024-12-27 00:48:13 +00:00
hashcat/OpenCL/m12600_a0-pure.cl
Jens Steube 04d5e5a119 New Attack-Mode: Association Attack. Like JtR's single mode. Very early
stage. See hashcat Forum for detailed writeup.
2020-09-29 15:56:32 +02:00

315 lines
8.4 KiB
Common Lisp

/**
* Author......: See docs/credits.txt
* License.....: MIT
*/
//#define NEW_SIMD_CODE
#ifdef KERNEL_STATIC
#include "inc_vendor.h"
#include "inc_types.h"
#include "inc_platform.cl"
#include "inc_common.cl"
#include "inc_rp.h"
#include "inc_rp.cl"
#include "inc_scalar.cl"
#include "inc_hash_sha1.cl"
#include "inc_hash_sha256.cl"
#endif
#if VECT_SIZE == 1
#define uint_to_hex_lower8_le(i) make_u32x (l_bin2asc[(i)])
#elif VECT_SIZE == 2
#define uint_to_hex_lower8_le(i) make_u32x (l_bin2asc[(i).s0], l_bin2asc[(i).s1])
#elif VECT_SIZE == 4
#define uint_to_hex_lower8_le(i) make_u32x (l_bin2asc[(i).s0], l_bin2asc[(i).s1], l_bin2asc[(i).s2], l_bin2asc[(i).s3])
#elif VECT_SIZE == 8
#define uint_to_hex_lower8_le(i) make_u32x (l_bin2asc[(i).s0], l_bin2asc[(i).s1], l_bin2asc[(i).s2], l_bin2asc[(i).s3], l_bin2asc[(i).s4], l_bin2asc[(i).s5], l_bin2asc[(i).s6], l_bin2asc[(i).s7])
#elif VECT_SIZE == 16
#define uint_to_hex_lower8_le(i) make_u32x (l_bin2asc[(i).s0], l_bin2asc[(i).s1], l_bin2asc[(i).s2], l_bin2asc[(i).s3], l_bin2asc[(i).s4], l_bin2asc[(i).s5], l_bin2asc[(i).s6], l_bin2asc[(i).s7], l_bin2asc[(i).s8], l_bin2asc[(i).s9], l_bin2asc[(i).sa], l_bin2asc[(i).sb], l_bin2asc[(i).sc], l_bin2asc[(i).sd], l_bin2asc[(i).se], l_bin2asc[(i).sf])
#endif
KERNEL_FQ void m12600_mxx (KERN_ATTR_RULES ())
{
/**
* modifier
*/
const u64 gid = get_global_id (0);
const u64 lid = get_local_id (0);
const u64 lsz = get_local_size (0);
/**
* bin2asc table
*/
LOCAL_VK u32 l_bin2asc[256];
for (u32 i = lid; i < 256; i += lsz)
{
const u32 i0 = (i >> 0) & 15;
const u32 i1 = (i >> 4) & 15;
l_bin2asc[i] = ((i0 < 10) ? '0' + i0 : 'A' - 10 + i0) << 0
| ((i1 < 10) ? '0' + i1 : 'A' - 10 + i1) << 8;
}
SYNC_THREADS ();
if (gid >= gid_max) return;
/**
* salt
*/
u32 pc256[8];
pc256[0] = salt_bufs[SALT_POS].salt_buf_pc[0];
pc256[1] = salt_bufs[SALT_POS].salt_buf_pc[1];
pc256[2] = salt_bufs[SALT_POS].salt_buf_pc[2];
pc256[3] = salt_bufs[SALT_POS].salt_buf_pc[3];
pc256[4] = salt_bufs[SALT_POS].salt_buf_pc[4];
pc256[5] = salt_bufs[SALT_POS].salt_buf_pc[5];
pc256[6] = salt_bufs[SALT_POS].salt_buf_pc[6];
pc256[7] = salt_bufs[SALT_POS].salt_buf_pc[7];
/**
* base
*/
COPY_PW (pws[gid]);
/**
* loop
*/
for (u32 il_pos = 0; il_pos < il_cnt; il_pos++)
{
pw_t tmp = PASTE_PW;
tmp.pw_len = apply_rules (rules_buf[il_pos].cmds, tmp.i, tmp.pw_len);
sha1_ctx_t ctx0;
sha1_init (&ctx0);
sha1_update_swap (&ctx0, tmp.i, tmp.pw_len);
sha1_final (&ctx0);
const u32 a = ctx0.h[0];
const u32 b = ctx0.h[1];
const u32 c = ctx0.h[2];
const u32 d = ctx0.h[3];
const u32 e = ctx0.h[4];
sha256_ctx_t ctx;
ctx.h[0] = pc256[0];
ctx.h[1] = pc256[1];
ctx.h[2] = pc256[2];
ctx.h[3] = pc256[3];
ctx.h[4] = pc256[4];
ctx.h[5] = pc256[5];
ctx.h[6] = pc256[6];
ctx.h[7] = pc256[7];
ctx.len = 64;
ctx.w0[0] = uint_to_hex_lower8_le ((a >> 16) & 255) << 0
| uint_to_hex_lower8_le ((a >> 24) & 255) << 16;
ctx.w0[1] = uint_to_hex_lower8_le ((a >> 0) & 255) << 0
| uint_to_hex_lower8_le ((a >> 8) & 255) << 16;
ctx.w0[2] = uint_to_hex_lower8_le ((b >> 16) & 255) << 0
| uint_to_hex_lower8_le ((b >> 24) & 255) << 16;
ctx.w0[3] = uint_to_hex_lower8_le ((b >> 0) & 255) << 0
| uint_to_hex_lower8_le ((b >> 8) & 255) << 16;
ctx.w1[0] = uint_to_hex_lower8_le ((c >> 16) & 255) << 0
| uint_to_hex_lower8_le ((c >> 24) & 255) << 16;
ctx.w1[1] = uint_to_hex_lower8_le ((c >> 0) & 255) << 0
| uint_to_hex_lower8_le ((c >> 8) & 255) << 16;
ctx.w1[2] = uint_to_hex_lower8_le ((d >> 16) & 255) << 0
| uint_to_hex_lower8_le ((d >> 24) & 255) << 16;
ctx.w1[3] = uint_to_hex_lower8_le ((d >> 0) & 255) << 0
| uint_to_hex_lower8_le ((d >> 8) & 255) << 16;
ctx.w2[0] = uint_to_hex_lower8_le ((e >> 16) & 255) << 0
| uint_to_hex_lower8_le ((e >> 24) & 255) << 16;
ctx.w2[1] = uint_to_hex_lower8_le ((e >> 0) & 255) << 0
| uint_to_hex_lower8_le ((e >> 8) & 255) << 16;
ctx.w2[2] = 0;
ctx.w2[3] = 0;
ctx.w3[0] = 0;
ctx.w3[1] = 0;
ctx.w3[2] = 0;
ctx.w3[3] = 0;
ctx.len += 40;
sha256_final (&ctx);
ctx.h[0] -= pc256[0];
ctx.h[1] -= pc256[1];
ctx.h[2] -= pc256[2];
ctx.h[3] -= pc256[3];
ctx.h[4] -= pc256[4];
ctx.h[5] -= pc256[5];
ctx.h[6] -= pc256[6];
ctx.h[7] -= pc256[7];
const u32 r0 = ctx.h[DGST_R0];
const u32 r1 = ctx.h[DGST_R1];
const u32 r2 = ctx.h[DGST_R2];
const u32 r3 = ctx.h[DGST_R3];
COMPARE_M_SCALAR (r0, r1, r2, r3);
}
}
KERNEL_FQ void m12600_sxx (KERN_ATTR_RULES ())
{
/**
* modifier
*/
const u64 gid = get_global_id (0);
const u64 lid = get_local_id (0);
const u64 lsz = get_local_size (0);
/**
* bin2asc table
*/
LOCAL_VK u32 l_bin2asc[256];
for (u32 i = lid; i < 256; i += lsz)
{
const u32 i0 = (i >> 0) & 15;
const u32 i1 = (i >> 4) & 15;
l_bin2asc[i] = ((i0 < 10) ? '0' + i0 : 'A' - 10 + i0) << 0
| ((i1 < 10) ? '0' + i1 : 'A' - 10 + i1) << 8;
}
SYNC_THREADS ();
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]
};
/**
* salt
*/
u32 pc256[8];
pc256[0] = salt_bufs[SALT_POS].salt_buf_pc[0];
pc256[1] = salt_bufs[SALT_POS].salt_buf_pc[1];
pc256[2] = salt_bufs[SALT_POS].salt_buf_pc[2];
pc256[3] = salt_bufs[SALT_POS].salt_buf_pc[3];
pc256[4] = salt_bufs[SALT_POS].salt_buf_pc[4];
pc256[5] = salt_bufs[SALT_POS].salt_buf_pc[5];
pc256[6] = salt_bufs[SALT_POS].salt_buf_pc[6];
pc256[7] = salt_bufs[SALT_POS].salt_buf_pc[7];
/**
* base
*/
COPY_PW (pws[gid]);
/**
* loop
*/
for (u32 il_pos = 0; il_pos < il_cnt; il_pos++)
{
pw_t tmp = PASTE_PW;
tmp.pw_len = apply_rules (rules_buf[il_pos].cmds, tmp.i, tmp.pw_len);
sha1_ctx_t ctx0;
sha1_init (&ctx0);
sha1_update_swap (&ctx0, tmp.i, tmp.pw_len);
sha1_final (&ctx0);
const u32 a = ctx0.h[0];
const u32 b = ctx0.h[1];
const u32 c = ctx0.h[2];
const u32 d = ctx0.h[3];
const u32 e = ctx0.h[4];
sha256_ctx_t ctx;
ctx.h[0] = pc256[0];
ctx.h[1] = pc256[1];
ctx.h[2] = pc256[2];
ctx.h[3] = pc256[3];
ctx.h[4] = pc256[4];
ctx.h[5] = pc256[5];
ctx.h[6] = pc256[6];
ctx.h[7] = pc256[7];
ctx.len = 64;
ctx.w0[0] = uint_to_hex_lower8_le ((a >> 16) & 255) << 0
| uint_to_hex_lower8_le ((a >> 24) & 255) << 16;
ctx.w0[1] = uint_to_hex_lower8_le ((a >> 0) & 255) << 0
| uint_to_hex_lower8_le ((a >> 8) & 255) << 16;
ctx.w0[2] = uint_to_hex_lower8_le ((b >> 16) & 255) << 0
| uint_to_hex_lower8_le ((b >> 24) & 255) << 16;
ctx.w0[3] = uint_to_hex_lower8_le ((b >> 0) & 255) << 0
| uint_to_hex_lower8_le ((b >> 8) & 255) << 16;
ctx.w1[0] = uint_to_hex_lower8_le ((c >> 16) & 255) << 0
| uint_to_hex_lower8_le ((c >> 24) & 255) << 16;
ctx.w1[1] = uint_to_hex_lower8_le ((c >> 0) & 255) << 0
| uint_to_hex_lower8_le ((c >> 8) & 255) << 16;
ctx.w1[2] = uint_to_hex_lower8_le ((d >> 16) & 255) << 0
| uint_to_hex_lower8_le ((d >> 24) & 255) << 16;
ctx.w1[3] = uint_to_hex_lower8_le ((d >> 0) & 255) << 0
| uint_to_hex_lower8_le ((d >> 8) & 255) << 16;
ctx.w2[0] = uint_to_hex_lower8_le ((e >> 16) & 255) << 0
| uint_to_hex_lower8_le ((e >> 24) & 255) << 16;
ctx.w2[1] = uint_to_hex_lower8_le ((e >> 0) & 255) << 0
| uint_to_hex_lower8_le ((e >> 8) & 255) << 16;
ctx.w2[2] = 0;
ctx.w2[3] = 0;
ctx.w3[0] = 0;
ctx.w3[1] = 0;
ctx.w3[2] = 0;
ctx.w3[3] = 0;
ctx.len += 40;
sha256_final (&ctx);
ctx.h[0] -= pc256[0];
ctx.h[1] -= pc256[1];
ctx.h[2] -= pc256[2];
ctx.h[3] -= pc256[3];
ctx.h[4] -= pc256[4];
ctx.h[5] -= pc256[5];
ctx.h[6] -= pc256[6];
ctx.h[7] -= pc256[7];
const u32 r0 = ctx.h[DGST_R0];
const u32 r1 = ctx.h[DGST_R1];
const u32 r2 = ctx.h[DGST_R2];
const u32 r3 = ctx.h[DGST_R3];
COMPARE_S_SCALAR (r0, r1, r2, r3);
}
}