2 * Author......: Jens Steube <jens.steube@gmail.com>
10 #include "inc_vendor.cl"
11 #include "inc_hash_constants.h"
12 #include "inc_hash_functions.cl"
13 #include "inc_types.cl"
14 #include "inc_common.cl"
17 #include "inc_simd.cl"
19 __constant u32 k_sha256[64] =
21 SHA256C00, SHA256C01, SHA256C02, SHA256C03,
22 SHA256C04, SHA256C05, SHA256C06, SHA256C07,
23 SHA256C08, SHA256C09, SHA256C0a, SHA256C0b,
24 SHA256C0c, SHA256C0d, SHA256C0e, SHA256C0f,
25 SHA256C10, SHA256C11, SHA256C12, SHA256C13,
26 SHA256C14, SHA256C15, SHA256C16, SHA256C17,
27 SHA256C18, SHA256C19, SHA256C1a, SHA256C1b,
28 SHA256C1c, SHA256C1d, SHA256C1e, SHA256C1f,
29 SHA256C20, SHA256C21, SHA256C22, SHA256C23,
30 SHA256C24, SHA256C25, SHA256C26, SHA256C27,
31 SHA256C28, SHA256C29, SHA256C2a, SHA256C2b,
32 SHA256C2c, SHA256C2d, SHA256C2e, SHA256C2f,
33 SHA256C30, SHA256C31, SHA256C32, SHA256C33,
34 SHA256C34, SHA256C35, SHA256C36, SHA256C37,
35 SHA256C38, SHA256C39, SHA256C3a, SHA256C3b,
36 SHA256C3c, SHA256C3d, SHA256C3e, SHA256C3f,
39 void sha256_transform (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u32x digest[8])
67 #define ROUND_EXPAND() \
69 w0_t = SHA256_EXPAND (we_t, w9_t, w1_t, w0_t); \
70 w1_t = SHA256_EXPAND (wf_t, wa_t, w2_t, w1_t); \
71 w2_t = SHA256_EXPAND (w0_t, wb_t, w3_t, w2_t); \
72 w3_t = SHA256_EXPAND (w1_t, wc_t, w4_t, w3_t); \
73 w4_t = SHA256_EXPAND (w2_t, wd_t, w5_t, w4_t); \
74 w5_t = SHA256_EXPAND (w3_t, we_t, w6_t, w5_t); \
75 w6_t = SHA256_EXPAND (w4_t, wf_t, w7_t, w6_t); \
76 w7_t = SHA256_EXPAND (w5_t, w0_t, w8_t, w7_t); \
77 w8_t = SHA256_EXPAND (w6_t, w1_t, w9_t, w8_t); \
78 w9_t = SHA256_EXPAND (w7_t, w2_t, wa_t, w9_t); \
79 wa_t = SHA256_EXPAND (w8_t, w3_t, wb_t, wa_t); \
80 wb_t = SHA256_EXPAND (w9_t, w4_t, wc_t, wb_t); \
81 wc_t = SHA256_EXPAND (wa_t, w5_t, wd_t, wc_t); \
82 wd_t = SHA256_EXPAND (wb_t, w6_t, we_t, wd_t); \
83 we_t = SHA256_EXPAND (wc_t, w7_t, wf_t, we_t); \
84 wf_t = SHA256_EXPAND (wd_t, w8_t, w0_t, wf_t); \
87 #define ROUND_STEP(i) \
89 SHA256_STEP (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w0_t, k_sha256[i + 0]); \
90 SHA256_STEP (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w1_t, k_sha256[i + 1]); \
91 SHA256_STEP (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, w2_t, k_sha256[i + 2]); \
92 SHA256_STEP (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, w3_t, k_sha256[i + 3]); \
93 SHA256_STEP (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, w4_t, k_sha256[i + 4]); \
94 SHA256_STEP (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, w5_t, k_sha256[i + 5]); \
95 SHA256_STEP (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, w6_t, k_sha256[i + 6]); \
96 SHA256_STEP (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, w7_t, k_sha256[i + 7]); \
97 SHA256_STEP (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w8_t, k_sha256[i + 8]); \
98 SHA256_STEP (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w9_t, k_sha256[i + 9]); \
99 SHA256_STEP (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, wa_t, k_sha256[i + 10]); \
100 SHA256_STEP (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, wb_t, k_sha256[i + 11]); \
101 SHA256_STEP (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, wc_t, k_sha256[i + 12]); \
102 SHA256_STEP (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, wd_t, k_sha256[i + 13]); \
103 SHA256_STEP (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, we_t, k_sha256[i + 14]); \
104 SHA256_STEP (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, wf_t, k_sha256[i + 15]); \
112 for (int i = 16; i < 64; i += 16)
114 ROUND_EXPAND (); ROUND_STEP (i);
127 void hmac_sha256_pad (u32x w0[4], u32x w1[4], u32x w2[4], u32x w3[4], u32x ipad[8], u32x opad[8])
129 w0[0] = w0[0] ^ 0x36363636;
130 w0[1] = w0[1] ^ 0x36363636;
131 w0[2] = w0[2] ^ 0x36363636;
132 w0[3] = w0[3] ^ 0x36363636;
133 w1[0] = w1[0] ^ 0x36363636;
134 w1[1] = w1[1] ^ 0x36363636;
135 w1[2] = w1[2] ^ 0x36363636;
136 w1[3] = w1[3] ^ 0x36363636;
137 w2[0] = w2[0] ^ 0x36363636;
138 w2[1] = w2[1] ^ 0x36363636;
139 w2[2] = w2[2] ^ 0x36363636;
140 w2[3] = w2[3] ^ 0x36363636;
141 w3[0] = w3[0] ^ 0x36363636;
142 w3[1] = w3[1] ^ 0x36363636;
143 w3[2] = w3[2] ^ 0x36363636;
144 w3[3] = w3[3] ^ 0x36363636;
155 sha256_transform (w0, w1, w2, w3, ipad);
157 w0[0] = w0[0] ^ 0x6a6a6a6a;
158 w0[1] = w0[1] ^ 0x6a6a6a6a;
159 w0[2] = w0[2] ^ 0x6a6a6a6a;
160 w0[3] = w0[3] ^ 0x6a6a6a6a;
161 w1[0] = w1[0] ^ 0x6a6a6a6a;
162 w1[1] = w1[1] ^ 0x6a6a6a6a;
163 w1[2] = w1[2] ^ 0x6a6a6a6a;
164 w1[3] = w1[3] ^ 0x6a6a6a6a;
165 w2[0] = w2[0] ^ 0x6a6a6a6a;
166 w2[1] = w2[1] ^ 0x6a6a6a6a;
167 w2[2] = w2[2] ^ 0x6a6a6a6a;
168 w2[3] = w2[3] ^ 0x6a6a6a6a;
169 w3[0] = w3[0] ^ 0x6a6a6a6a;
170 w3[1] = w3[1] ^ 0x6a6a6a6a;
171 w3[2] = w3[2] ^ 0x6a6a6a6a;
172 w3[3] = w3[3] ^ 0x6a6a6a6a;
183 sha256_transform (w0, w1, w2, w3, opad);
186 void hmac_sha256_run (u32x w0[4], u32x w1[4], u32x w2[4], u32x w3[4], u32x ipad[8], u32x opad[8], u32x digest[8])
197 sha256_transform (w0, w1, w2, w3, digest);
214 w3[3] = (64 + 32) * 8;
225 sha256_transform (w0, w1, w2, w3, digest);
228 __kernel void m01460_m04 (__global pw_t *pws, __global kernel_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_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_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 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
234 const u32 lid = get_local_id (0);
240 const u32 gid = get_global_id (0);
242 if (gid >= gid_max) return;
247 pw_buf0[0] = pws[gid].i[0];
248 pw_buf0[1] = pws[gid].i[1];
249 pw_buf0[2] = pws[gid].i[2];
250 pw_buf0[3] = pws[gid].i[3];
251 pw_buf1[0] = pws[gid].i[4];
252 pw_buf1[1] = pws[gid].i[5];
253 pw_buf1[2] = pws[gid].i[6];
254 pw_buf1[3] = pws[gid].i[7];
256 const u32 pw_len = pws[gid].pw_len;
267 salt_buf0[0] = swap32_S (salt_bufs[salt_pos].salt_buf[ 0]);
268 salt_buf0[1] = swap32_S (salt_bufs[salt_pos].salt_buf[ 1]);
269 salt_buf0[2] = swap32_S (salt_bufs[salt_pos].salt_buf[ 2]);
270 salt_buf0[3] = swap32_S (salt_bufs[salt_pos].salt_buf[ 3]);
271 salt_buf1[0] = swap32_S (salt_bufs[salt_pos].salt_buf[ 4]);
272 salt_buf1[1] = swap32_S (salt_bufs[salt_pos].salt_buf[ 5]);
273 salt_buf1[2] = swap32_S (salt_bufs[salt_pos].salt_buf[ 6]);
274 salt_buf1[3] = swap32_S (salt_bufs[salt_pos].salt_buf[ 7]);
275 salt_buf2[0] = swap32_S (salt_bufs[salt_pos].salt_buf[ 8]);
276 salt_buf2[1] = swap32_S (salt_bufs[salt_pos].salt_buf[ 9]);
277 salt_buf2[2] = swap32_S (salt_bufs[salt_pos].salt_buf[10]);
278 salt_buf2[3] = swap32_S (salt_bufs[salt_pos].salt_buf[11]);
279 salt_buf3[0] = swap32_S (salt_bufs[salt_pos].salt_buf[12]);
280 salt_buf3[1] = swap32_S (salt_bufs[salt_pos].salt_buf[13]);
281 salt_buf3[2] = swap32_S (salt_bufs[salt_pos].salt_buf[14]);
282 salt_buf3[3] = swap32_S (salt_bufs[salt_pos].salt_buf[15]);
284 const u32 salt_len = salt_bufs[salt_pos].salt_len;
295 w0_t[0] = salt_buf0[0];
296 w0_t[1] = salt_buf0[1];
297 w0_t[2] = salt_buf0[2];
298 w0_t[3] = salt_buf0[3];
299 w1_t[0] = salt_buf1[0];
300 w1_t[1] = salt_buf1[1];
301 w1_t[2] = salt_buf1[2];
302 w1_t[3] = salt_buf1[3];
303 w2_t[0] = salt_buf2[0];
304 w2_t[1] = salt_buf2[1];
305 w2_t[2] = salt_buf2[2];
306 w2_t[3] = salt_buf2[3];
307 w3_t[0] = salt_buf3[0];
308 w3_t[1] = salt_buf3[1];
309 w3_t[2] = salt_buf3[2];
310 w3_t[3] = salt_buf3[3];
315 hmac_sha256_pad (w0_t, w1_t, w2_t, w3_t, ipad, opad);
321 for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
328 const u32x out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
330 append_0x80_2x4_VV (w0, w1, out_len);
332 w0_t[0] = swap32 (w0[0]);
333 w0_t[1] = swap32 (w0[1]);
334 w0_t[2] = swap32 (w0[2]);
335 w0_t[3] = swap32 (w0[3]);
336 w1_t[0] = swap32 (w1[0]);
337 w1_t[1] = swap32 (w1[1]);
338 w1_t[2] = swap32 (w1[2]);
339 w1_t[3] = swap32 (w1[3]);
347 w3_t[3] = (64 + out_len) * 8;
351 hmac_sha256_run (w0_t, w1_t, w2_t, w3_t, ipad, opad, digest);
353 COMPARE_M_SIMD (digest[3], digest[7], digest[2], digest[6]);
357 __kernel void m01460_m08 (__global pw_t *pws, __global kernel_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_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_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 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
361 __kernel void m01460_m16 (__global pw_t *pws, __global kernel_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_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_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 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
365 __kernel void m01460_s04 (__global pw_t *pws, __global kernel_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_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_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 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
371 const u32 lid = get_local_id (0);
377 const u32 gid = get_global_id (0);
379 if (gid >= gid_max) return;
384 pw_buf0[0] = pws[gid].i[0];
385 pw_buf0[1] = pws[gid].i[1];
386 pw_buf0[2] = pws[gid].i[2];
387 pw_buf0[3] = pws[gid].i[3];
388 pw_buf1[0] = pws[gid].i[4];
389 pw_buf1[1] = pws[gid].i[5];
390 pw_buf1[2] = pws[gid].i[6];
391 pw_buf1[3] = pws[gid].i[7];
393 const u32 pw_len = pws[gid].pw_len;
404 salt_buf0[0] = swap32_S (salt_bufs[salt_pos].salt_buf[ 0]);
405 salt_buf0[1] = swap32_S (salt_bufs[salt_pos].salt_buf[ 1]);
406 salt_buf0[2] = swap32_S (salt_bufs[salt_pos].salt_buf[ 2]);
407 salt_buf0[3] = swap32_S (salt_bufs[salt_pos].salt_buf[ 3]);
408 salt_buf1[0] = swap32_S (salt_bufs[salt_pos].salt_buf[ 4]);
409 salt_buf1[1] = swap32_S (salt_bufs[salt_pos].salt_buf[ 5]);
410 salt_buf1[2] = swap32_S (salt_bufs[salt_pos].salt_buf[ 6]);
411 salt_buf1[3] = swap32_S (salt_bufs[salt_pos].salt_buf[ 7]);
412 salt_buf2[0] = swap32_S (salt_bufs[salt_pos].salt_buf[ 8]);
413 salt_buf2[1] = swap32_S (salt_bufs[salt_pos].salt_buf[ 9]);
414 salt_buf2[2] = swap32_S (salt_bufs[salt_pos].salt_buf[10]);
415 salt_buf2[3] = swap32_S (salt_bufs[salt_pos].salt_buf[11]);
416 salt_buf3[0] = swap32_S (salt_bufs[salt_pos].salt_buf[12]);
417 salt_buf3[1] = swap32_S (salt_bufs[salt_pos].salt_buf[13]);
418 salt_buf3[2] = swap32_S (salt_bufs[salt_pos].salt_buf[14]);
419 salt_buf3[3] = swap32_S (salt_bufs[salt_pos].salt_buf[15]);
421 const u32 salt_len = salt_bufs[salt_pos].salt_len;
432 w0_t[0] = salt_buf0[0];
433 w0_t[1] = salt_buf0[1];
434 w0_t[2] = salt_buf0[2];
435 w0_t[3] = salt_buf0[3];
436 w1_t[0] = salt_buf1[0];
437 w1_t[1] = salt_buf1[1];
438 w1_t[2] = salt_buf1[2];
439 w1_t[3] = salt_buf1[3];
440 w2_t[0] = salt_buf2[0];
441 w2_t[1] = salt_buf2[1];
442 w2_t[2] = salt_buf2[2];
443 w2_t[3] = salt_buf2[3];
444 w3_t[0] = salt_buf3[0];
445 w3_t[1] = salt_buf3[1];
446 w3_t[2] = salt_buf3[2];
447 w3_t[3] = salt_buf3[3];
452 hmac_sha256_pad (w0_t, w1_t, w2_t, w3_t, ipad, opad);
458 const u32 search[4] =
460 digests_buf[digests_offset].digest_buf[DGST_R0],
461 digests_buf[digests_offset].digest_buf[DGST_R1],
462 digests_buf[digests_offset].digest_buf[DGST_R2],
463 digests_buf[digests_offset].digest_buf[DGST_R3]
470 for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
477 const u32x out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
479 append_0x80_2x4_VV (w0, w1, out_len);
481 w0_t[0] = swap32 (w0[0]);
482 w0_t[1] = swap32 (w0[1]);
483 w0_t[2] = swap32 (w0[2]);
484 w0_t[3] = swap32 (w0[3]);
485 w1_t[0] = swap32 (w1[0]);
486 w1_t[1] = swap32 (w1[1]);
487 w1_t[2] = swap32 (w1[2]);
488 w1_t[3] = swap32 (w1[3]);
496 w3_t[3] = (64 + out_len) * 8;
500 hmac_sha256_run (w0_t, w1_t, w2_t, w3_t, ipad, opad, digest);
502 COMPARE_S_SIMD (digest[3], digest[7], digest[2], digest[6]);
506 __kernel void m01460_s08 (__global pw_t *pws, __global kernel_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_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_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 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
510 __kernel void m01460_s16 (__global pw_t *pws, __global kernel_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_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_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 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)