2 * Author......: Jens Steube <jens.steube@gmail.com>
8 #include "include/constants.h"
9 #include "include/kernel_vendor.h"
28 #include "include/kernel_functions.c"
29 #include "types_amd.c"
30 #include "common_amd.c"
31 #include "include/rp_gpu.h"
35 #define VECT_COMPARE_S "check_single_vect1_comp4.c"
36 #define VECT_COMPARE_M "check_multi_vect1_comp4.c"
40 #define VECT_COMPARE_S "check_single_vect2_comp4.c"
41 #define VECT_COMPARE_M "check_multi_vect2_comp4.c"
45 #define VECT_COMPARE_S "check_single_vect4_comp4.c"
46 #define VECT_COMPARE_M "check_multi_vect4_comp4.c"
49 __constant u32 k_sha256[64] =
51 SHA256C00, SHA256C01, SHA256C02, SHA256C03,
52 SHA256C04, SHA256C05, SHA256C06, SHA256C07,
53 SHA256C08, SHA256C09, SHA256C0a, SHA256C0b,
54 SHA256C0c, SHA256C0d, SHA256C0e, SHA256C0f,
55 SHA256C10, SHA256C11, SHA256C12, SHA256C13,
56 SHA256C14, SHA256C15, SHA256C16, SHA256C17,
57 SHA256C18, SHA256C19, SHA256C1a, SHA256C1b,
58 SHA256C1c, SHA256C1d, SHA256C1e, SHA256C1f,
59 SHA256C20, SHA256C21, SHA256C22, SHA256C23,
60 SHA256C24, SHA256C25, SHA256C26, SHA256C27,
61 SHA256C28, SHA256C29, SHA256C2a, SHA256C2b,
62 SHA256C2c, SHA256C2d, SHA256C2e, SHA256C2f,
63 SHA256C30, SHA256C31, SHA256C32, SHA256C33,
64 SHA256C34, SHA256C35, SHA256C36, SHA256C37,
65 SHA256C38, SHA256C39, SHA256C3a, SHA256C3b,
66 SHA256C3c, SHA256C3d, SHA256C3e, SHA256C3f,
69 static void sha256_transform (u32x digest[8], const u32x w[16])
97 #define ROUND_EXPAND() \
99 w0_t = SHA256_EXPAND (we_t, w9_t, w1_t, w0_t); \
100 w1_t = SHA256_EXPAND (wf_t, wa_t, w2_t, w1_t); \
101 w2_t = SHA256_EXPAND (w0_t, wb_t, w3_t, w2_t); \
102 w3_t = SHA256_EXPAND (w1_t, wc_t, w4_t, w3_t); \
103 w4_t = SHA256_EXPAND (w2_t, wd_t, w5_t, w4_t); \
104 w5_t = SHA256_EXPAND (w3_t, we_t, w6_t, w5_t); \
105 w6_t = SHA256_EXPAND (w4_t, wf_t, w7_t, w6_t); \
106 w7_t = SHA256_EXPAND (w5_t, w0_t, w8_t, w7_t); \
107 w8_t = SHA256_EXPAND (w6_t, w1_t, w9_t, w8_t); \
108 w9_t = SHA256_EXPAND (w7_t, w2_t, wa_t, w9_t); \
109 wa_t = SHA256_EXPAND (w8_t, w3_t, wb_t, wa_t); \
110 wb_t = SHA256_EXPAND (w9_t, w4_t, wc_t, wb_t); \
111 wc_t = SHA256_EXPAND (wa_t, w5_t, wd_t, wc_t); \
112 wd_t = SHA256_EXPAND (wb_t, w6_t, we_t, wd_t); \
113 we_t = SHA256_EXPAND (wc_t, w7_t, wf_t, we_t); \
114 wf_t = SHA256_EXPAND (wd_t, w8_t, w0_t, wf_t); \
117 #define ROUND_STEP(i) \
119 SHA256_STEP (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w0_t, k_sha256[i + 0]); \
120 SHA256_STEP (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w1_t, k_sha256[i + 1]); \
121 SHA256_STEP (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, w2_t, k_sha256[i + 2]); \
122 SHA256_STEP (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, w3_t, k_sha256[i + 3]); \
123 SHA256_STEP (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, w4_t, k_sha256[i + 4]); \
124 SHA256_STEP (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, w5_t, k_sha256[i + 5]); \
125 SHA256_STEP (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, w6_t, k_sha256[i + 6]); \
126 SHA256_STEP (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, w7_t, k_sha256[i + 7]); \
127 SHA256_STEP (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w8_t, k_sha256[i + 8]); \
128 SHA256_STEP (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w9_t, k_sha256[i + 9]); \
129 SHA256_STEP (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, wa_t, k_sha256[i + 10]); \
130 SHA256_STEP (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, wb_t, k_sha256[i + 11]); \
131 SHA256_STEP (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, wc_t, k_sha256[i + 12]); \
132 SHA256_STEP (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, wd_t, k_sha256[i + 13]); \
133 SHA256_STEP (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, we_t, k_sha256[i + 14]); \
134 SHA256_STEP (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, wf_t, k_sha256[i + 15]); \
140 for (int i = 16; i < 64; i += 16)
142 ROUND_EXPAND (); ROUND_STEP (i);
155 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08000_m04 (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global void *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
161 const u32 lid = get_local_id (0);
167 const u32 gid = get_global_id (0);
169 if (gid >= gid_max) return;
173 pw_buf0[0] = pws[gid].i[ 0];
174 pw_buf0[1] = pws[gid].i[ 1];
175 pw_buf0[2] = pws[gid].i[ 2];
176 pw_buf0[3] = pws[gid].i[ 3];
180 pw_buf1[0] = pws[gid].i[ 4];
181 pw_buf1[1] = pws[gid].i[ 5];
182 pw_buf1[2] = pws[gid].i[ 6];
183 pw_buf1[3] = pws[gid].i[ 7];
185 const u32 pw_len = pws[gid].pw_len;
191 const u32 salt_buf0 = swap_workaround (salt_bufs[salt_pos].salt_buf[ 0]);
192 const u32 salt_buf1 = swap_workaround (salt_bufs[salt_pos].salt_buf[ 1]);
193 const u32 salt_buf2 = swap_workaround (salt_bufs[salt_pos].salt_buf[ 2]); // 0x80
199 for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++)
223 const u32 out_len = apply_rules (rules_buf[il_pos].cmds, w0, w1, pw_len);
230 make_unicode (w0, w0_t, w1_t);
232 make_unicode (w1, w2_t, w3_t);
236 w_t[ 0] = swap_workaround (w0_t[0]);
237 w_t[ 1] = swap_workaround (w0_t[1]);
238 w_t[ 2] = swap_workaround (w0_t[2]);
239 w_t[ 3] = swap_workaround (w0_t[3]);
240 w_t[ 4] = swap_workaround (w1_t[0]);
241 w_t[ 5] = swap_workaround (w1_t[1]);
242 w_t[ 6] = swap_workaround (w1_t[2]);
243 w_t[ 7] = swap_workaround (w1_t[3]);
244 w_t[ 8] = swap_workaround (w2_t[0]);
245 w_t[ 9] = swap_workaround (w2_t[1]);
246 w_t[10] = swap_workaround (w2_t[2]);
247 w_t[11] = swap_workaround (w2_t[3]);
248 w_t[12] = swap_workaround (w3_t[0]);
249 w_t[13] = swap_workaround (w3_t[1]);
250 w_t[14] = swap_workaround (w3_t[2]);
251 w_t[15] = swap_workaround (w3_t[3]);
253 w_t[ 0] = w_t[ 0] >> 8;
254 w_t[ 1] = w_t[ 1] >> 8;
255 w_t[ 2] = w_t[ 2] >> 8;
256 w_t[ 3] = w_t[ 3] >> 8;
257 w_t[ 4] = w_t[ 4] >> 8;
258 w_t[ 5] = w_t[ 5] >> 8;
259 w_t[ 6] = w_t[ 6] >> 8;
260 w_t[ 7] = w_t[ 7] >> 8;
261 w_t[ 8] = w_t[ 8] >> 8;
262 w_t[ 9] = w_t[ 9] >> 8;
263 w_t[10] = w_t[10] >> 8;
264 w_t[11] = w_t[11] >> 8;
265 w_t[12] = w_t[12] >> 8;
266 w_t[13] = w_t[13] >> 8;
267 w_t[14] = w_t[14] >> 8;
268 w_t[15] = w_t[15] >> 8;
272 digest[0] = SHA256M_A;
273 digest[1] = SHA256M_B;
274 digest[2] = SHA256M_C;
275 digest[3] = SHA256M_D;
276 digest[4] = SHA256M_E;
277 digest[5] = SHA256M_F;
278 digest[6] = SHA256M_G;
279 digest[7] = SHA256M_H;
281 sha256_transform (digest, w_t); // 0 - 64
300 sha256_transform (digest, w_t); // 64 - 128
301 sha256_transform (digest, w_t); // 128 - 192
302 sha256_transform (digest, w_t); // 192 - 256
303 sha256_transform (digest, w_t); // 256 - 320
304 sha256_transform (digest, w_t); // 320 - 384
305 sha256_transform (digest, w_t); // 384 - 448
307 w_t[15] = 0 | salt_buf0 >> 16;
309 sha256_transform (digest, w_t); // 448 - 512
311 w_t[ 0] = salt_buf0 << 16 | salt_buf1 >> 16;
312 w_t[ 1] = salt_buf1 << 16 | salt_buf2 >> 16;
313 w_t[ 2] = salt_buf2 << 16 | 0;
314 w_t[15] = (510 + 8) * 8;
316 sha256_transform (digest, w_t); // 512 - 576
318 const u32x r0 = digest[3];
319 const u32x r1 = digest[7];
320 const u32x r2 = digest[2];
321 const u32x r3 = digest[6];
323 #include VECT_COMPARE_M
327 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08000_m08 (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global void *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
331 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08000_m16 (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global void *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
335 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08000_s04 (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global void *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
341 const u32 lid = get_local_id (0);
347 const u32 gid = get_global_id (0);
349 if (gid >= gid_max) return;
353 pw_buf0[0] = pws[gid].i[ 0];
354 pw_buf0[1] = pws[gid].i[ 1];
355 pw_buf0[2] = pws[gid].i[ 2];
356 pw_buf0[3] = pws[gid].i[ 3];
360 pw_buf1[0] = pws[gid].i[ 4];
361 pw_buf1[1] = pws[gid].i[ 5];
362 pw_buf1[2] = pws[gid].i[ 6];
363 pw_buf1[3] = pws[gid].i[ 7];
365 const u32 pw_len = pws[gid].pw_len;
371 const u32 salt_buf0 = swap_workaround (salt_bufs[salt_pos].salt_buf[ 0]);
372 const u32 salt_buf1 = swap_workaround (salt_bufs[salt_pos].salt_buf[ 1]);
373 const u32 salt_buf2 = swap_workaround (salt_bufs[salt_pos].salt_buf[ 2]); // 0x80
379 const u32 search[4] =
381 digests_buf[digests_offset].digest_buf[DGST_R0],
382 digests_buf[digests_offset].digest_buf[DGST_R1],
383 digests_buf[digests_offset].digest_buf[DGST_R2],
384 digests_buf[digests_offset].digest_buf[DGST_R3]
391 for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++)
415 const u32 out_len = apply_rules (rules_buf[il_pos].cmds, w0, w1, pw_len);
422 make_unicode (w0, w0_t, w1_t);
424 make_unicode (w1, w2_t, w3_t);
428 w_t[ 0] = swap_workaround (w0_t[0]);
429 w_t[ 1] = swap_workaround (w0_t[1]);
430 w_t[ 2] = swap_workaround (w0_t[2]);
431 w_t[ 3] = swap_workaround (w0_t[3]);
432 w_t[ 4] = swap_workaround (w1_t[0]);
433 w_t[ 5] = swap_workaround (w1_t[1]);
434 w_t[ 6] = swap_workaround (w1_t[2]);
435 w_t[ 7] = swap_workaround (w1_t[3]);
436 w_t[ 8] = swap_workaround (w2_t[0]);
437 w_t[ 9] = swap_workaround (w2_t[1]);
438 w_t[10] = swap_workaround (w2_t[2]);
439 w_t[11] = swap_workaround (w2_t[3]);
440 w_t[12] = swap_workaround (w3_t[0]);
441 w_t[13] = swap_workaround (w3_t[1]);
442 w_t[14] = swap_workaround (w3_t[2]);
443 w_t[15] = swap_workaround (w3_t[3]);
445 w_t[ 0] = w_t[ 0] >> 8;
446 w_t[ 1] = w_t[ 1] >> 8;
447 w_t[ 2] = w_t[ 2] >> 8;
448 w_t[ 3] = w_t[ 3] >> 8;
449 w_t[ 4] = w_t[ 4] >> 8;
450 w_t[ 5] = w_t[ 5] >> 8;
451 w_t[ 6] = w_t[ 6] >> 8;
452 w_t[ 7] = w_t[ 7] >> 8;
453 w_t[ 8] = w_t[ 8] >> 8;
454 w_t[ 9] = w_t[ 9] >> 8;
455 w_t[10] = w_t[10] >> 8;
456 w_t[11] = w_t[11] >> 8;
457 w_t[12] = w_t[12] >> 8;
458 w_t[13] = w_t[13] >> 8;
459 w_t[14] = w_t[14] >> 8;
460 w_t[15] = w_t[15] >> 8;
464 digest[0] = SHA256M_A;
465 digest[1] = SHA256M_B;
466 digest[2] = SHA256M_C;
467 digest[3] = SHA256M_D;
468 digest[4] = SHA256M_E;
469 digest[5] = SHA256M_F;
470 digest[6] = SHA256M_G;
471 digest[7] = SHA256M_H;
473 sha256_transform (digest, w_t); // 0 - 64
492 sha256_transform (digest, w_t); // 64 - 128
493 sha256_transform (digest, w_t); // 128 - 192
494 sha256_transform (digest, w_t); // 192 - 256
495 sha256_transform (digest, w_t); // 256 - 320
496 sha256_transform (digest, w_t); // 320 - 384
497 sha256_transform (digest, w_t); // 384 - 448
499 w_t[15] = 0 | salt_buf0 >> 16;
501 sha256_transform (digest, w_t); // 448 - 512
503 w_t[ 0] = salt_buf0 << 16 | salt_buf1 >> 16;
504 w_t[ 1] = salt_buf1 << 16 | salt_buf2 >> 16;
505 w_t[ 2] = salt_buf2 << 16 | 0;
506 w_t[15] = (510 + 8) * 8;
508 sha256_transform (digest, w_t); // 512 - 576
510 const u32x r0 = digest[3];
511 const u32x r1 = digest[7];
512 const u32x r2 = digest[2];
513 const u32x r3 = digest[6];
515 #include VECT_COMPARE_S
519 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08000_s08 (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global void *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
523 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08000_s16 (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global void *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)