2 * Author......: Jens Steube <jens.steube@gmail.com>
8 #include "include/constants.h"
9 #include "include/kernel_vendor.h"
16 #include "include/kernel_functions.c"
17 #include "OpenCL/types_ocl.c"
18 #include "OpenCL/common.c"
19 #include "include/rp_gpu.h"
20 #include "OpenCL/rp.c"
22 #define COMPARE_S "OpenCL/check_single_comp4.c"
23 #define COMPARE_M "OpenCL/check_multi_comp4.c"
25 __constant u64 k_sha512[80] =
27 SHA512C00, SHA512C01, SHA512C02, SHA512C03,
28 SHA512C04, SHA512C05, SHA512C06, SHA512C07,
29 SHA512C08, SHA512C09, SHA512C0a, SHA512C0b,
30 SHA512C0c, SHA512C0d, SHA512C0e, SHA512C0f,
31 SHA512C10, SHA512C11, SHA512C12, SHA512C13,
32 SHA512C14, SHA512C15, SHA512C16, SHA512C17,
33 SHA512C18, SHA512C19, SHA512C1a, SHA512C1b,
34 SHA512C1c, SHA512C1d, SHA512C1e, SHA512C1f,
35 SHA512C20, SHA512C21, SHA512C22, SHA512C23,
36 SHA512C24, SHA512C25, SHA512C26, SHA512C27,
37 SHA512C28, SHA512C29, SHA512C2a, SHA512C2b,
38 SHA512C2c, SHA512C2d, SHA512C2e, SHA512C2f,
39 SHA512C30, SHA512C31, SHA512C32, SHA512C33,
40 SHA512C34, SHA512C35, SHA512C36, SHA512C37,
41 SHA512C38, SHA512C39, SHA512C3a, SHA512C3b,
42 SHA512C3c, SHA512C3d, SHA512C3e, SHA512C3f,
43 SHA512C40, SHA512C41, SHA512C42, SHA512C43,
44 SHA512C44, SHA512C45, SHA512C46, SHA512C47,
45 SHA512C48, SHA512C49, SHA512C4a, SHA512C4b,
46 SHA512C4c, SHA512C4d, SHA512C4e, SHA512C4f,
49 static void sha512_transform (const u64 w0[4], const u64 w1[4], const u64 w2[4], const u64 w3[4], u64 digest[8])
77 #define ROUND_EXPAND() \
79 w0_t = SHA512_EXPAND (we_t, w9_t, w1_t, w0_t); \
80 w1_t = SHA512_EXPAND (wf_t, wa_t, w2_t, w1_t); \
81 w2_t = SHA512_EXPAND (w0_t, wb_t, w3_t, w2_t); \
82 w3_t = SHA512_EXPAND (w1_t, wc_t, w4_t, w3_t); \
83 w4_t = SHA512_EXPAND (w2_t, wd_t, w5_t, w4_t); \
84 w5_t = SHA512_EXPAND (w3_t, we_t, w6_t, w5_t); \
85 w6_t = SHA512_EXPAND (w4_t, wf_t, w7_t, w6_t); \
86 w7_t = SHA512_EXPAND (w5_t, w0_t, w8_t, w7_t); \
87 w8_t = SHA512_EXPAND (w6_t, w1_t, w9_t, w8_t); \
88 w9_t = SHA512_EXPAND (w7_t, w2_t, wa_t, w9_t); \
89 wa_t = SHA512_EXPAND (w8_t, w3_t, wb_t, wa_t); \
90 wb_t = SHA512_EXPAND (w9_t, w4_t, wc_t, wb_t); \
91 wc_t = SHA512_EXPAND (wa_t, w5_t, wd_t, wc_t); \
92 wd_t = SHA512_EXPAND (wb_t, w6_t, we_t, wd_t); \
93 we_t = SHA512_EXPAND (wc_t, w7_t, wf_t, we_t); \
94 wf_t = SHA512_EXPAND (wd_t, w8_t, w0_t, wf_t); \
97 #define ROUND_STEP(i) \
99 SHA512_STEP (SHA512_F0o, SHA512_F1o, a, b, c, d, e, f, g, h, w0_t, k_sha512[i + 0]); \
100 SHA512_STEP (SHA512_F0o, SHA512_F1o, h, a, b, c, d, e, f, g, w1_t, k_sha512[i + 1]); \
101 SHA512_STEP (SHA512_F0o, SHA512_F1o, g, h, a, b, c, d, e, f, w2_t, k_sha512[i + 2]); \
102 SHA512_STEP (SHA512_F0o, SHA512_F1o, f, g, h, a, b, c, d, e, w3_t, k_sha512[i + 3]); \
103 SHA512_STEP (SHA512_F0o, SHA512_F1o, e, f, g, h, a, b, c, d, w4_t, k_sha512[i + 4]); \
104 SHA512_STEP (SHA512_F0o, SHA512_F1o, d, e, f, g, h, a, b, c, w5_t, k_sha512[i + 5]); \
105 SHA512_STEP (SHA512_F0o, SHA512_F1o, c, d, e, f, g, h, a, b, w6_t, k_sha512[i + 6]); \
106 SHA512_STEP (SHA512_F0o, SHA512_F1o, b, c, d, e, f, g, h, a, w7_t, k_sha512[i + 7]); \
107 SHA512_STEP (SHA512_F0o, SHA512_F1o, a, b, c, d, e, f, g, h, w8_t, k_sha512[i + 8]); \
108 SHA512_STEP (SHA512_F0o, SHA512_F1o, h, a, b, c, d, e, f, g, w9_t, k_sha512[i + 9]); \
109 SHA512_STEP (SHA512_F0o, SHA512_F1o, g, h, a, b, c, d, e, f, wa_t, k_sha512[i + 10]); \
110 SHA512_STEP (SHA512_F0o, SHA512_F1o, f, g, h, a, b, c, d, e, wb_t, k_sha512[i + 11]); \
111 SHA512_STEP (SHA512_F0o, SHA512_F1o, e, f, g, h, a, b, c, d, wc_t, k_sha512[i + 12]); \
112 SHA512_STEP (SHA512_F0o, SHA512_F1o, d, e, f, g, h, a, b, c, wd_t, k_sha512[i + 13]); \
113 SHA512_STEP (SHA512_F0o, SHA512_F1o, c, d, e, f, g, h, a, b, we_t, k_sha512[i + 14]); \
114 SHA512_STEP (SHA512_F0o, SHA512_F1o, b, c, d, e, f, g, h, a, wf_t, k_sha512[i + 15]); \
120 for (int i = 16; i < 80; i += 16)
122 ROUND_EXPAND (); ROUND_STEP (i);
135 static void hmac_sha512_pad (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u64 ipad[8], u64 opad[8])
142 w0_t[0] = hl32_to_64 (w0[0], w0[1]) ^ 0x3636363636363636;
143 w0_t[1] = hl32_to_64 (w0[2], w0[3]) ^ 0x3636363636363636;
144 w0_t[2] = hl32_to_64 (w1[0], w1[1]) ^ 0x3636363636363636;
145 w0_t[3] = hl32_to_64 (w1[2], w1[3]) ^ 0x3636363636363636;
146 w1_t[0] = hl32_to_64 (w2[0], w2[1]) ^ 0x3636363636363636;
147 w1_t[1] = hl32_to_64 (w2[2], w2[3]) ^ 0x3636363636363636;
148 w1_t[2] = hl32_to_64 (w3[0], w3[1]) ^ 0x3636363636363636;
149 w1_t[3] = hl32_to_64 (w3[2], w3[3]) ^ 0x3636363636363636;
150 w2_t[0] = 0 ^ 0x3636363636363636;
151 w2_t[1] = 0 ^ 0x3636363636363636;
152 w2_t[2] = 0 ^ 0x3636363636363636;
153 w2_t[3] = 0 ^ 0x3636363636363636;
154 w3_t[0] = 0 ^ 0x3636363636363636;
155 w3_t[1] = 0 ^ 0x3636363636363636;
156 w3_t[2] = 0 ^ 0x3636363636363636;
157 w3_t[3] = 0 ^ 0x3636363636363636;
168 sha512_transform (w0_t, w1_t, w2_t, w3_t, ipad);
170 w0_t[0] = hl32_to_64 (w0[0], w0[1]) ^ 0x5c5c5c5c5c5c5c5c;
171 w0_t[1] = hl32_to_64 (w0[2], w0[3]) ^ 0x5c5c5c5c5c5c5c5c;
172 w0_t[2] = hl32_to_64 (w1[0], w1[1]) ^ 0x5c5c5c5c5c5c5c5c;
173 w0_t[3] = hl32_to_64 (w1[2], w1[3]) ^ 0x5c5c5c5c5c5c5c5c;
174 w1_t[0] = hl32_to_64 (w2[0], w2[1]) ^ 0x5c5c5c5c5c5c5c5c;
175 w1_t[1] = hl32_to_64 (w2[2], w2[3]) ^ 0x5c5c5c5c5c5c5c5c;
176 w1_t[2] = hl32_to_64 (w3[0], w3[1]) ^ 0x5c5c5c5c5c5c5c5c;
177 w1_t[3] = hl32_to_64 (w3[2], w3[3]) ^ 0x5c5c5c5c5c5c5c5c;
178 w2_t[0] = 0 ^ 0x5c5c5c5c5c5c5c5c;
179 w2_t[1] = 0 ^ 0x5c5c5c5c5c5c5c5c;
180 w2_t[2] = 0 ^ 0x5c5c5c5c5c5c5c5c;
181 w2_t[3] = 0 ^ 0x5c5c5c5c5c5c5c5c;
182 w3_t[0] = 0 ^ 0x5c5c5c5c5c5c5c5c;
183 w3_t[1] = 0 ^ 0x5c5c5c5c5c5c5c5c;
184 w3_t[2] = 0 ^ 0x5c5c5c5c5c5c5c5c;
185 w3_t[3] = 0 ^ 0x5c5c5c5c5c5c5c5c;
196 sha512_transform (w0_t, w1_t, w2_t, w3_t, opad);
199 static void hmac_sha512_run (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u64 ipad[8], u64 opad[8], u64 digest[8])
206 w0_t[0] = hl32_to_64 (w0[0], w0[1]);
207 w0_t[1] = hl32_to_64 (w0[2], w0[3]);
208 w0_t[2] = hl32_to_64 (w1[0], w1[1]);
209 w0_t[3] = hl32_to_64 (w1[2], w1[3]);
210 w1_t[0] = hl32_to_64 (w2[0], w2[1]);
211 w1_t[1] = hl32_to_64 (w2[2], w2[3]);
212 w1_t[2] = hl32_to_64 (w3[0], w3[1]);
221 w3_t[3] = hl32_to_64 (w3[2], w3[3]);
232 sha512_transform (w0_t, w1_t, w2_t, w3_t, digest);
242 w2_t[0] = 0x8000000000000000;
249 w3_t[3] = (128 + 64) * 8;
260 sha512_transform (w0_t, w1_t, w2_t, w3_t, digest);
263 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m01760_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)
269 const u32 lid = get_local_id (0);
275 const u32 gid = get_global_id (0);
277 if (gid >= gid_max) return;
281 pw_buf0[0] = pws[gid].i[ 0];
282 pw_buf0[1] = pws[gid].i[ 1];
283 pw_buf0[2] = pws[gid].i[ 2];
284 pw_buf0[3] = pws[gid].i[ 3];
288 pw_buf1[0] = pws[gid].i[ 4];
289 pw_buf1[1] = pws[gid].i[ 5];
290 pw_buf1[2] = pws[gid].i[ 6];
291 pw_buf1[3] = pws[gid].i[ 7];
293 const u32 pw_len = pws[gid].pw_len;
301 salt_buf0[0] = salt_bufs[salt_pos].salt_buf[ 0];
302 salt_buf0[1] = salt_bufs[salt_pos].salt_buf[ 1];
303 salt_buf0[2] = salt_bufs[salt_pos].salt_buf[ 2];
304 salt_buf0[3] = salt_bufs[salt_pos].salt_buf[ 3];
308 salt_buf1[0] = salt_bufs[salt_pos].salt_buf[ 4];
309 salt_buf1[1] = salt_bufs[salt_pos].salt_buf[ 5];
310 salt_buf1[2] = salt_bufs[salt_pos].salt_buf[ 6];
311 salt_buf1[3] = salt_bufs[salt_pos].salt_buf[ 7];
319 w0_t[0] = swap32 (salt_buf0[0]);
320 w0_t[1] = swap32 (salt_buf0[1]);
321 w0_t[2] = swap32 (salt_buf0[2]);
322 w0_t[3] = swap32 (salt_buf0[3]);
326 w1_t[0] = swap32 (salt_buf1[0]);
327 w1_t[1] = swap32 (salt_buf1[1]);
328 w1_t[2] = swap32 (salt_buf1[2]);
329 w1_t[3] = swap32 (salt_buf1[3]);
348 hmac_sha512_pad (w0_t, w1_t, w2_t, w3_t, ipad, opad);
354 for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++)
384 const u32 out_len = apply_rules (rules_buf[il_pos].cmds, w0, w1, pw_len);
386 append_0x80_2x4 (w0, w1, out_len);
388 w0_t[0] = swap32 (w0[0]);
389 w0_t[1] = swap32 (w0[1]);
390 w0_t[2] = swap32 (w0[2]);
391 w0_t[3] = swap32 (w0[3]);
392 w1_t[0] = swap32 (w1[0]);
393 w1_t[1] = swap32 (w1[1]);
394 w1_t[2] = swap32 (w1[2]);
395 w1_t[3] = swap32 (w1[3]);
403 w3_t[3] = (128 + out_len) * 8;
407 hmac_sha512_run (w0_t, w1_t, w2_t, w3_t, ipad, opad, digest);
410 const u32 r0 = l32_from_64 (digest[7]);
411 const u32 r1 = h32_from_64 (digest[7]);
412 const u32 r2 = l32_from_64 (digest[3]);
413 const u32 r3 = h32_from_64 (digest[3]);
419 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m01760_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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
423 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m01760_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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
427 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m01760_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)
433 const u32 lid = get_local_id (0);
439 const u32 gid = get_global_id (0);
441 if (gid >= gid_max) return;
445 pw_buf0[0] = pws[gid].i[ 0];
446 pw_buf0[1] = pws[gid].i[ 1];
447 pw_buf0[2] = pws[gid].i[ 2];
448 pw_buf0[3] = pws[gid].i[ 3];
452 pw_buf1[0] = pws[gid].i[ 4];
453 pw_buf1[1] = pws[gid].i[ 5];
454 pw_buf1[2] = pws[gid].i[ 6];
455 pw_buf1[3] = pws[gid].i[ 7];
457 const u32 pw_len = pws[gid].pw_len;
465 salt_buf0[0] = salt_bufs[salt_pos].salt_buf[ 0];
466 salt_buf0[1] = salt_bufs[salt_pos].salt_buf[ 1];
467 salt_buf0[2] = salt_bufs[salt_pos].salt_buf[ 2];
468 salt_buf0[3] = salt_bufs[salt_pos].salt_buf[ 3];
472 salt_buf1[0] = salt_bufs[salt_pos].salt_buf[ 4];
473 salt_buf1[1] = salt_bufs[salt_pos].salt_buf[ 5];
474 salt_buf1[2] = salt_bufs[salt_pos].salt_buf[ 6];
475 salt_buf1[3] = salt_bufs[salt_pos].salt_buf[ 7];
483 w0_t[0] = swap32 (salt_buf0[0]);
484 w0_t[1] = swap32 (salt_buf0[1]);
485 w0_t[2] = swap32 (salt_buf0[2]);
486 w0_t[3] = swap32 (salt_buf0[3]);
490 w1_t[0] = swap32 (salt_buf1[0]);
491 w1_t[1] = swap32 (salt_buf1[1]);
492 w1_t[2] = swap32 (salt_buf1[2]);
493 w1_t[3] = swap32 (salt_buf1[3]);
512 hmac_sha512_pad (w0_t, w1_t, w2_t, w3_t, ipad, opad);
518 const u32 search[4] =
520 digests_buf[digests_offset].digest_buf[DGST_R0],
521 digests_buf[digests_offset].digest_buf[DGST_R1],
522 digests_buf[digests_offset].digest_buf[DGST_R2],
523 digests_buf[digests_offset].digest_buf[DGST_R3]
530 for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++)
560 const u32 out_len = apply_rules (rules_buf[il_pos].cmds, w0, w1, pw_len);
562 append_0x80_2x4 (w0, w1, out_len);
564 w0_t[0] = swap32 (w0[0]);
565 w0_t[1] = swap32 (w0[1]);
566 w0_t[2] = swap32 (w0[2]);
567 w0_t[3] = swap32 (w0[3]);
568 w1_t[0] = swap32 (w1[0]);
569 w1_t[1] = swap32 (w1[1]);
570 w1_t[2] = swap32 (w1[2]);
571 w1_t[3] = swap32 (w1[3]);
579 w3_t[3] = (128 + out_len) * 8;
583 hmac_sha512_run (w0_t, w1_t, w2_t, w3_t, ipad, opad, digest);
586 const u32 r0 = l32_from_64 (digest[7]);
587 const u32 r1 = h32_from_64 (digest[7]);
588 const u32 r2 = l32_from_64 (digest[3]);
589 const u32 r3 = h32_from_64 (digest[3]);
595 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m01760_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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
599 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m01760_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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)