2 * Author......: Jens Steube <jens.steube@gmail.com>
8 #include "include/constants.h"
9 #include "include/kernel_vendor.h"
24 #include "include/kernel_functions.c"
26 #include "common_nv.c"
27 #include "include/rp_gpu.h"
31 #define VECT_COMPARE_S "check_single_vect1_comp4.c"
32 #define VECT_COMPARE_M "check_multi_vect1_comp4.c"
36 #define VECT_COMPARE_S "check_single_vect2_comp4.c"
37 #define VECT_COMPARE_M "check_multi_vect2_comp4.c"
40 __device__ __constant__ u32 lotus_magic_table[256] =
42 0xbd, 0x56, 0xea, 0xf2, 0xa2, 0xf1, 0xac, 0x2a,
43 0xb0, 0x93, 0xd1, 0x9c, 0x1b, 0x33, 0xfd, 0xd0,
44 0x30, 0x04, 0xb6, 0xdc, 0x7d, 0xdf, 0x32, 0x4b,
45 0xf7, 0xcb, 0x45, 0x9b, 0x31, 0xbb, 0x21, 0x5a,
46 0x41, 0x9f, 0xe1, 0xd9, 0x4a, 0x4d, 0x9e, 0xda,
47 0xa0, 0x68, 0x2c, 0xc3, 0x27, 0x5f, 0x80, 0x36,
48 0x3e, 0xee, 0xfb, 0x95, 0x1a, 0xfe, 0xce, 0xa8,
49 0x34, 0xa9, 0x13, 0xf0, 0xa6, 0x3f, 0xd8, 0x0c,
50 0x78, 0x24, 0xaf, 0x23, 0x52, 0xc1, 0x67, 0x17,
51 0xf5, 0x66, 0x90, 0xe7, 0xe8, 0x07, 0xb8, 0x60,
52 0x48, 0xe6, 0x1e, 0x53, 0xf3, 0x92, 0xa4, 0x72,
53 0x8c, 0x08, 0x15, 0x6e, 0x86, 0x00, 0x84, 0xfa,
54 0xf4, 0x7f, 0x8a, 0x42, 0x19, 0xf6, 0xdb, 0xcd,
55 0x14, 0x8d, 0x50, 0x12, 0xba, 0x3c, 0x06, 0x4e,
56 0xec, 0xb3, 0x35, 0x11, 0xa1, 0x88, 0x8e, 0x2b,
57 0x94, 0x99, 0xb7, 0x71, 0x74, 0xd3, 0xe4, 0xbf,
58 0x3a, 0xde, 0x96, 0x0e, 0xbc, 0x0a, 0xed, 0x77,
59 0xfc, 0x37, 0x6b, 0x03, 0x79, 0x89, 0x62, 0xc6,
60 0xd7, 0xc0, 0xd2, 0x7c, 0x6a, 0x8b, 0x22, 0xa3,
61 0x5b, 0x05, 0x5d, 0x02, 0x75, 0xd5, 0x61, 0xe3,
62 0x18, 0x8f, 0x55, 0x51, 0xad, 0x1f, 0x0b, 0x5e,
63 0x85, 0xe5, 0xc2, 0x57, 0x63, 0xca, 0x3d, 0x6c,
64 0xb4, 0xc5, 0xcc, 0x70, 0xb2, 0x91, 0x59, 0x0d,
65 0x47, 0x20, 0xc8, 0x4f, 0x58, 0xe0, 0x01, 0xe2,
66 0x16, 0x38, 0xc4, 0x6f, 0x3b, 0x0f, 0x65, 0x46,
67 0xbe, 0x7e, 0x2d, 0x7b, 0x82, 0xf9, 0x40, 0xb5,
68 0x1d, 0x73, 0xf8, 0xeb, 0x26, 0xc7, 0x87, 0x97,
69 0x25, 0x54, 0xb1, 0x28, 0xaa, 0x98, 0x9d, 0xa5,
70 0x64, 0x6d, 0x7a, 0xd4, 0x10, 0x81, 0x44, 0xef,
71 0x49, 0xd6, 0xae, 0x2e, 0xdd, 0x76, 0x5c, 0x2f,
72 0xa7, 0x1c, 0xc9, 0x09, 0x69, 0x9a, 0x83, 0xcf,
73 0x29, 0x39, 0xb9, 0xe9, 0x4c, 0xff, 0x43, 0xab,
77 #define BOX(S,i) u32x ((S)[(i)])
81 #define BOX(S,i) u32x ((S)[(i).x], (S)[(i).y])
85 #define uint_to_hex_upper8(i) l_bin2asc[(i)]
89 #define uint_to_hex_upper8(i) u32x (l_bin2asc[(i).x], l_bin2asc[(i).y])
92 __device__ __constant__ char c_bin2asc[16] = { '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'A', 'B', 'C', 'D', 'E', 'F' };
94 __device__ __shared__ short l_bin2asc[256];
96 __device__ static void lotus_mix (u32x *in, u32 s_lotus_magic_table[256])
100 for (int i = 0; i < 18; i++)
105 for (int j = 0; j < 12; j++)
110 p = (p + s--) & 0xff; p = ((tmp_in >> 0) & 0xff) ^ BOX (s_lotus_magic_table, p); tmp_out |= p << 0;
111 p = (p + s--) & 0xff; p = ((tmp_in >> 8) & 0xff) ^ BOX (s_lotus_magic_table, p); tmp_out |= p << 8;
112 p = (p + s--) & 0xff; p = ((tmp_in >> 16) & 0xff) ^ BOX (s_lotus_magic_table, p); tmp_out |= p << 16;
113 p = (p + s--) & 0xff; p = ((tmp_in >> 24) & 0xff) ^ BOX (s_lotus_magic_table, p); tmp_out |= p << 24;
120 __device__ static void lotus_transform_password (u32x *in, u32x *out, u32 s_lotus_magic_table[256])
122 u32x t = out[3] >> 24;
126 for (int i = 0; i < 4; i++)
128 t ^= (in[i] >> 0) & 0xff; c = BOX (s_lotus_magic_table, t); out[i] ^= c << 0; t = ((out[i] >> 0) & 0xff);
129 t ^= (in[i] >> 8) & 0xff; c = BOX (s_lotus_magic_table, t); out[i] ^= c << 8; t = ((out[i] >> 8) & 0xff);
130 t ^= (in[i] >> 16) & 0xff; c = BOX (s_lotus_magic_table, t); out[i] ^= c << 16; t = ((out[i] >> 16) & 0xff);
131 t ^= (in[i] >> 24) & 0xff; c = BOX (s_lotus_magic_table, t); out[i] ^= c << 24; t = ((out[i] >> 24) & 0xff);
135 __device__ static void pad (u32x w[4], const u32 len)
137 const u32 val = 16 - len;
139 const u32 mask1 = val << 24;
141 const u32 mask2 = val << 16
144 const u32 mask3 = val << 8
148 const u32 mask4 = val << 0
155 case 0: w[0] = mask4;
160 case 1: w[0] |= mask3;
165 case 2: w[0] |= mask2;
170 case 3: w[0] |= mask1;
175 case 4: w[1] = mask4;
179 case 5: w[1] |= mask3;
183 case 6: w[1] |= mask2;
187 case 7: w[1] |= mask1;
191 case 8: w[2] = mask4;
194 case 9: w[2] |= mask3;
197 case 10: w[2] |= mask2;
200 case 11: w[2] |= mask1;
203 case 12: w[3] = mask4;
205 case 13: w[3] |= mask3;
207 case 14: w[3] |= mask2;
209 case 15: w[3] |= mask1;
214 __device__ static void mdtransform_norecalc (u32x state[4], u32x block[4], u32 s_lotus_magic_table[256])
226 x[ 8] = state[0] ^ block[0];
227 x[ 9] = state[1] ^ block[1];
228 x[10] = state[2] ^ block[2];
229 x[11] = state[3] ^ block[3];
231 lotus_mix (x, s_lotus_magic_table);
239 __device__ static void mdtransform (u32x state[4], u32x checksum[4], u32x block[4], u32 s_lotus_magic_table[256])
241 mdtransform_norecalc (state, block, s_lotus_magic_table);
243 lotus_transform_password (block, checksum, s_lotus_magic_table);
246 __device__ static void domino_big_md (const u32x saved_key[16], const u32 size, u32x state[4], u32 s_lotus_magic_table[256])
265 for (curpos = 0, idx = 0; curpos + 16 < size; curpos += 16, idx += 4)
267 block[0] = saved_key[idx + 0];
268 block[1] = saved_key[idx + 1];
269 block[2] = saved_key[idx + 2];
270 block[3] = saved_key[idx + 3];
272 mdtransform (state, checksum, block, s_lotus_magic_table);
275 u32 left = size - curpos;
277 block[0] = saved_key[idx + 0];
278 block[1] = saved_key[idx + 1];
279 block[2] = saved_key[idx + 2];
280 block[3] = saved_key[idx + 3];
282 mdtransform (state, checksum, block, s_lotus_magic_table);
284 mdtransform_norecalc (state, checksum, s_lotus_magic_table);
287 __device__ __constant__ gpu_rule_t c_rules[1024];
289 extern "C" __global__ void __launch_bounds__ (256, 1) m08700_m04 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const void *esalt_bufs, u32 *d_return_buf, 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)
295 const u32 lid = threadIdx.x;
301 const u32 gid = (blockIdx.x * blockDim.x) + threadIdx.x;
307 l_bin2asc[lid] = c_bin2asc[(lid >> 0) & 15] << 8
308 | c_bin2asc[(lid >> 4) & 15] << 0;
316 __shared__ u32 s_lotus_magic_table[256];
318 s_lotus_magic_table[lid] = lotus_magic_table[lid];
322 if (gid >= gid_max) return;
330 pw_buf0[0] = pws[gid].i[ 0];
331 pw_buf0[1] = pws[gid].i[ 1];
332 pw_buf0[2] = pws[gid].i[ 2];
333 pw_buf0[3] = pws[gid].i[ 3];
337 pw_buf1[0] = pws[gid].i[ 4];
338 pw_buf1[1] = pws[gid].i[ 5];
339 pw_buf1[2] = pws[gid].i[ 6];
340 pw_buf1[3] = pws[gid].i[ 7];
342 const u32 pw_len = pws[gid].pw_len;
348 const u32 salt0 = salt_bufs[salt_pos].salt_buf[0];
349 const u32 salt1 = salt_bufs[salt_pos].salt_buf[1] & 0xff | '(' << 8;
355 for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++)
385 const u32 out_len = apply_rules (c_rules[il_pos].cmds, w0, w1, pw_len);
419 pad (&w[ 0], pw_len & 0xf);
421 else if (pw_len < 32)
423 pad (&w[ 4], pw_len & 0xf);
425 else if (pw_len < 48)
427 pad (&w[ 8], pw_len & 0xf);
429 else if (pw_len < 64)
431 pad (&w[12], pw_len & 0xf);
434 domino_big_md (w, pw_len, state, s_lotus_magic_table);
436 const u32x w0_t = uint_to_hex_upper8 ((state[0] >> 0) & 255) << 0
437 | uint_to_hex_upper8 ((state[0] >> 8) & 255) << 16;
438 const u32x w1_t = uint_to_hex_upper8 ((state[0] >> 16) & 255) << 0
439 | uint_to_hex_upper8 ((state[0] >> 24) & 255) << 16;
440 const u32x w2_t = uint_to_hex_upper8 ((state[1] >> 0) & 255) << 0
441 | uint_to_hex_upper8 ((state[1] >> 8) & 255) << 16;
442 const u32x w3_t = uint_to_hex_upper8 ((state[1] >> 16) & 255) << 0
443 | uint_to_hex_upper8 ((state[1] >> 24) & 255) << 16;
444 const u32x w4_t = uint_to_hex_upper8 ((state[2] >> 0) & 255) << 0
445 | uint_to_hex_upper8 ((state[2] >> 8) & 255) << 16;
446 const u32x w5_t = uint_to_hex_upper8 ((state[2] >> 16) & 255) << 0
447 | uint_to_hex_upper8 ((state[2] >> 24) & 255) << 16;
448 const u32x w6_t = uint_to_hex_upper8 ((state[3] >> 0) & 255) << 0
449 | uint_to_hex_upper8 ((state[3] >> 8) & 255) << 16;
450 //const u32x w7_t = uint_to_hex_upper8 ((state[3] >> 16) & 255) << 0
451 // | uint_to_hex_upper8 ((state[3] >> 24) & 255) << 16;
453 const u32 pade = 0x0e0e0e0e;
456 w[ 1] = salt1 | w0_t << 16;
457 w[ 2] = w0_t >> 16 | w1_t << 16;
458 w[ 3] = w1_t >> 16 | w2_t << 16;
459 w[ 4] = w2_t >> 16 | w3_t << 16;
460 w[ 5] = w3_t >> 16 | w4_t << 16;
461 w[ 6] = w4_t >> 16 | w5_t << 16;
462 w[ 7] = w5_t >> 16 | w6_t << 16;
463 w[ 8] = w6_t >> 16 | pade << 16; // | w7_t << 8;
477 domino_big_md (w, 34, state, s_lotus_magic_table);
479 u32x a = state[0] & 0xffffffff;
480 u32x b = state[1] & 0xffffffff;
481 u32x c = state[2] & 0x000000ff;
482 u32x d = state[3] & 0x00000000;
489 #include VECT_COMPARE_M
493 extern "C" __global__ void __launch_bounds__ (256, 1) m08700_m08 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const void *esalt_bufs, u32 *d_return_buf, 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)
497 extern "C" __global__ void __launch_bounds__ (256, 1) m08700_m16 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const void *esalt_bufs, u32 *d_return_buf, 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)
501 extern "C" __global__ void __launch_bounds__ (256, 1) m08700_s04 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const void *esalt_bufs, u32 *d_return_buf, 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)
507 const u32 lid = threadIdx.x;
513 const u32 gid = (blockIdx.x * blockDim.x) + threadIdx.x;
519 l_bin2asc[lid] = c_bin2asc[(lid >> 0) & 15] << 8
520 | c_bin2asc[(lid >> 4) & 15] << 0;
528 __shared__ u32 s_lotus_magic_table[256];
530 s_lotus_magic_table[lid] = lotus_magic_table[lid];
534 if (gid >= gid_max) return;
542 pw_buf0[0] = pws[gid].i[ 0];
543 pw_buf0[1] = pws[gid].i[ 1];
544 pw_buf0[2] = pws[gid].i[ 2];
545 pw_buf0[3] = pws[gid].i[ 3];
549 pw_buf1[0] = pws[gid].i[ 4];
550 pw_buf1[1] = pws[gid].i[ 5];
551 pw_buf1[2] = pws[gid].i[ 6];
552 pw_buf1[3] = pws[gid].i[ 7];
554 const u32 pw_len = pws[gid].pw_len;
560 const u32 salt0 = salt_bufs[salt_pos].salt_buf[0];
561 const u32 salt1 = salt_bufs[salt_pos].salt_buf[1] & 0xff | '(' << 8;
567 const u32 search[4] =
569 digests_buf[digests_offset].digest_buf[DGST_R0],
570 digests_buf[digests_offset].digest_buf[DGST_R1],
571 digests_buf[digests_offset].digest_buf[DGST_R2],
572 digests_buf[digests_offset].digest_buf[DGST_R3]
579 for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++)
609 const u32 out_len = apply_rules (c_rules[il_pos].cmds, w0, w1, pw_len);
643 pad (&w[ 0], pw_len & 0xf);
645 else if (pw_len < 32)
647 pad (&w[ 4], pw_len & 0xf);
649 else if (pw_len < 48)
651 pad (&w[ 8], pw_len & 0xf);
653 else if (pw_len < 64)
655 pad (&w[12], pw_len & 0xf);
658 domino_big_md (w, pw_len, state, s_lotus_magic_table);
660 const u32x w0_t = uint_to_hex_upper8 ((state[0] >> 0) & 255) << 0
661 | uint_to_hex_upper8 ((state[0] >> 8) & 255) << 16;
662 const u32x w1_t = uint_to_hex_upper8 ((state[0] >> 16) & 255) << 0
663 | uint_to_hex_upper8 ((state[0] >> 24) & 255) << 16;
664 const u32x w2_t = uint_to_hex_upper8 ((state[1] >> 0) & 255) << 0
665 | uint_to_hex_upper8 ((state[1] >> 8) & 255) << 16;
666 const u32x w3_t = uint_to_hex_upper8 ((state[1] >> 16) & 255) << 0
667 | uint_to_hex_upper8 ((state[1] >> 24) & 255) << 16;
668 const u32x w4_t = uint_to_hex_upper8 ((state[2] >> 0) & 255) << 0
669 | uint_to_hex_upper8 ((state[2] >> 8) & 255) << 16;
670 const u32x w5_t = uint_to_hex_upper8 ((state[2] >> 16) & 255) << 0
671 | uint_to_hex_upper8 ((state[2] >> 24) & 255) << 16;
672 const u32x w6_t = uint_to_hex_upper8 ((state[3] >> 0) & 255) << 0
673 | uint_to_hex_upper8 ((state[3] >> 8) & 255) << 16;
674 //const u32x w7_t = uint_to_hex_upper8 ((state[3] >> 16) & 255) << 0
675 // | uint_to_hex_upper8 ((state[3] >> 24) & 255) << 16;
677 const u32 pade = 0x0e0e0e0e;
680 w[ 1] = salt1 | w0_t << 16;
681 w[ 2] = w0_t >> 16 | w1_t << 16;
682 w[ 3] = w1_t >> 16 | w2_t << 16;
683 w[ 4] = w2_t >> 16 | w3_t << 16;
684 w[ 5] = w3_t >> 16 | w4_t << 16;
685 w[ 6] = w4_t >> 16 | w5_t << 16;
686 w[ 7] = w5_t >> 16 | w6_t << 16;
687 w[ 8] = w6_t >> 16 | pade << 16; // | w7_t << 8;
701 domino_big_md (w, 34, state, s_lotus_magic_table);
703 u32x a = state[0] & 0xffffffff;
704 u32x b = state[1] & 0xffffffff;
705 u32x c = state[2] & 0x000000ff;
706 u32x d = state[3] & 0x00000000;
713 #include VECT_COMPARE_S
717 extern "C" __global__ void __launch_bounds__ (256, 1) m08700_s08 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const void *esalt_bufs, u32 *d_return_buf, 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)
721 extern "C" __global__ void __launch_bounds__ (256, 1) m08700_s16 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const void *esalt_bufs, u32 *d_return_buf, 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)