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 u64 k_sha512[80] =
51 SHA512C00, SHA512C01, SHA512C02, SHA512C03,
52 SHA512C04, SHA512C05, SHA512C06, SHA512C07,
53 SHA512C08, SHA512C09, SHA512C0a, SHA512C0b,
54 SHA512C0c, SHA512C0d, SHA512C0e, SHA512C0f,
55 SHA512C10, SHA512C11, SHA512C12, SHA512C13,
56 SHA512C14, SHA512C15, SHA512C16, SHA512C17,
57 SHA512C18, SHA512C19, SHA512C1a, SHA512C1b,
58 SHA512C1c, SHA512C1d, SHA512C1e, SHA512C1f,
59 SHA512C20, SHA512C21, SHA512C22, SHA512C23,
60 SHA512C24, SHA512C25, SHA512C26, SHA512C27,
61 SHA512C28, SHA512C29, SHA512C2a, SHA512C2b,
62 SHA512C2c, SHA512C2d, SHA512C2e, SHA512C2f,
63 SHA512C30, SHA512C31, SHA512C32, SHA512C33,
64 SHA512C34, SHA512C35, SHA512C36, SHA512C37,
65 SHA512C38, SHA512C39, SHA512C3a, SHA512C3b,
66 SHA512C3c, SHA512C3d, SHA512C3e, SHA512C3f,
67 SHA512C40, SHA512C41, SHA512C42, SHA512C43,
68 SHA512C44, SHA512C45, SHA512C46, SHA512C47,
69 SHA512C48, SHA512C49, SHA512C4a, SHA512C4b,
70 SHA512C4c, SHA512C4d, SHA512C4e, SHA512C4f,
73 static void sha512_transform (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u64x digest[8])
75 u64x w0_t = hl32_to_64 (w0[0], w0[1]);
76 u64x w1_t = hl32_to_64 (w0[2], w0[3]);
77 u64x w2_t = hl32_to_64 (w1[0], w1[1]);
78 u64x w3_t = hl32_to_64 (w1[2], w1[3]);
79 u64x w4_t = hl32_to_64 (w2[0], w2[1]);
80 u64x w5_t = hl32_to_64 (w2[2], w2[3]);
81 u64x w6_t = hl32_to_64 (w3[0], w3[1]);
90 u64x wf_t = hl32_to_64 (w3[2], w3[3]);
101 #define ROUND_EXPAND() \
103 w0_t = SHA512_EXPAND (we_t, w9_t, w1_t, w0_t); \
104 w1_t = SHA512_EXPAND (wf_t, wa_t, w2_t, w1_t); \
105 w2_t = SHA512_EXPAND (w0_t, wb_t, w3_t, w2_t); \
106 w3_t = SHA512_EXPAND (w1_t, wc_t, w4_t, w3_t); \
107 w4_t = SHA512_EXPAND (w2_t, wd_t, w5_t, w4_t); \
108 w5_t = SHA512_EXPAND (w3_t, we_t, w6_t, w5_t); \
109 w6_t = SHA512_EXPAND (w4_t, wf_t, w7_t, w6_t); \
110 w7_t = SHA512_EXPAND (w5_t, w0_t, w8_t, w7_t); \
111 w8_t = SHA512_EXPAND (w6_t, w1_t, w9_t, w8_t); \
112 w9_t = SHA512_EXPAND (w7_t, w2_t, wa_t, w9_t); \
113 wa_t = SHA512_EXPAND (w8_t, w3_t, wb_t, wa_t); \
114 wb_t = SHA512_EXPAND (w9_t, w4_t, wc_t, wb_t); \
115 wc_t = SHA512_EXPAND (wa_t, w5_t, wd_t, wc_t); \
116 wd_t = SHA512_EXPAND (wb_t, w6_t, we_t, wd_t); \
117 we_t = SHA512_EXPAND (wc_t, w7_t, wf_t, we_t); \
118 wf_t = SHA512_EXPAND (wd_t, w8_t, w0_t, wf_t); \
121 #define ROUND_STEP(i) \
123 SHA512_STEP (SHA512_F0o, SHA512_F1o, a, b, c, d, e, f, g, h, w0_t, k_sha512[i + 0]); \
124 SHA512_STEP (SHA512_F0o, SHA512_F1o, h, a, b, c, d, e, f, g, w1_t, k_sha512[i + 1]); \
125 SHA512_STEP (SHA512_F0o, SHA512_F1o, g, h, a, b, c, d, e, f, w2_t, k_sha512[i + 2]); \
126 SHA512_STEP (SHA512_F0o, SHA512_F1o, f, g, h, a, b, c, d, e, w3_t, k_sha512[i + 3]); \
127 SHA512_STEP (SHA512_F0o, SHA512_F1o, e, f, g, h, a, b, c, d, w4_t, k_sha512[i + 4]); \
128 SHA512_STEP (SHA512_F0o, SHA512_F1o, d, e, f, g, h, a, b, c, w5_t, k_sha512[i + 5]); \
129 SHA512_STEP (SHA512_F0o, SHA512_F1o, c, d, e, f, g, h, a, b, w6_t, k_sha512[i + 6]); \
130 SHA512_STEP (SHA512_F0o, SHA512_F1o, b, c, d, e, f, g, h, a, w7_t, k_sha512[i + 7]); \
131 SHA512_STEP (SHA512_F0o, SHA512_F1o, a, b, c, d, e, f, g, h, w8_t, k_sha512[i + 8]); \
132 SHA512_STEP (SHA512_F0o, SHA512_F1o, h, a, b, c, d, e, f, g, w9_t, k_sha512[i + 9]); \
133 SHA512_STEP (SHA512_F0o, SHA512_F1o, g, h, a, b, c, d, e, f, wa_t, k_sha512[i + 10]); \
134 SHA512_STEP (SHA512_F0o, SHA512_F1o, f, g, h, a, b, c, d, e, wb_t, k_sha512[i + 11]); \
135 SHA512_STEP (SHA512_F0o, SHA512_F1o, e, f, g, h, a, b, c, d, wc_t, k_sha512[i + 12]); \
136 SHA512_STEP (SHA512_F0o, SHA512_F1o, d, e, f, g, h, a, b, c, wd_t, k_sha512[i + 13]); \
137 SHA512_STEP (SHA512_F0o, SHA512_F1o, c, d, e, f, g, h, a, b, we_t, k_sha512[i + 14]); \
138 SHA512_STEP (SHA512_F0o, SHA512_F1o, b, c, d, e, f, g, h, a, wf_t, k_sha512[i + 15]); \
144 for (int i = 16; i < 80; i += 16)
146 ROUND_EXPAND (); ROUND_STEP (i);
170 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m01720_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)
176 const u32 lid = get_local_id (0);
182 const u32 gid = get_global_id (0);
184 if (gid >= gid_max) return;
188 pw_buf0[0] = pws[gid].i[ 0];
189 pw_buf0[1] = pws[gid].i[ 1];
190 pw_buf0[2] = pws[gid].i[ 2];
191 pw_buf0[3] = pws[gid].i[ 3];
195 pw_buf1[0] = pws[gid].i[ 4];
196 pw_buf1[1] = pws[gid].i[ 5];
197 pw_buf1[2] = pws[gid].i[ 6];
198 pw_buf1[3] = pws[gid].i[ 7];
200 const u32 pw_len = pws[gid].pw_len;
208 salt_buf0[0] = salt_bufs[salt_pos].salt_buf[0];
209 salt_buf0[1] = salt_bufs[salt_pos].salt_buf[1];
210 salt_buf0[2] = salt_bufs[salt_pos].salt_buf[2];
211 salt_buf0[3] = salt_bufs[salt_pos].salt_buf[3];
215 salt_buf1[0] = salt_bufs[salt_pos].salt_buf[4];
216 salt_buf1[1] = salt_bufs[salt_pos].salt_buf[5];
217 salt_buf1[2] = salt_bufs[salt_pos].salt_buf[6];
218 salt_buf1[3] = salt_bufs[salt_pos].salt_buf[7];
220 const u32 salt_len = salt_bufs[salt_pos].salt_len;
226 for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++)
256 const u32 out_len = apply_rules (rules_buf[il_pos].cmds, w0, w1, pw_len);
262 const u32 out_salt_len = out_len + salt_len;
264 switch_buffer_by_offset (w0, w1, w2, w3, salt_len);
266 w0[0] |= salt_buf0[0];
267 w0[1] |= salt_buf0[1];
268 w0[2] |= salt_buf0[2];
269 w0[3] |= salt_buf0[3];
270 w1[0] |= salt_buf1[0];
271 w1[1] |= salt_buf1[1];
272 w1[2] |= salt_buf1[2];
273 w1[3] |= salt_buf1[3];
275 append_0x80_4 (w0, w1, w2, w3, out_salt_len);
286 w0_t[0] = swap_workaround (w0[0]);
287 w0_t[1] = swap_workaround (w0[1]);
288 w0_t[2] = swap_workaround (w0[2]);
289 w0_t[3] = swap_workaround (w0[3]);
290 w1_t[0] = swap_workaround (w1[0]);
291 w1_t[1] = swap_workaround (w1[1]);
292 w1_t[2] = swap_workaround (w1[2]);
293 w1_t[3] = swap_workaround (w1[3]);
294 w2_t[0] = swap_workaround (w2[0]);
295 w2_t[1] = swap_workaround (w2[1]);
296 w2_t[2] = swap_workaround (w2[2]);
297 w2_t[3] = swap_workaround (w2[3]);
298 w3_t[0] = swap_workaround (w3[0]);
299 w3_t[1] = swap_workaround (w3[1]);
301 w3_t[3] = out_salt_len * 8;
305 digest[0] = SHA512M_A;
306 digest[1] = SHA512M_B;
307 digest[2] = SHA512M_C;
308 digest[3] = SHA512M_D;
309 digest[4] = SHA512M_E;
310 digest[5] = SHA512M_F;
311 digest[6] = SHA512M_G;
312 digest[7] = SHA512M_H;
314 sha512_transform (w0_t, w1_t, w2_t, w3_t, digest);
317 const u32x r0 = l32_from_64 (digest[7]);
318 const u32x r1 = h32_from_64 (digest[7]);
319 const u32x r2 = l32_from_64 (digest[3]);
320 const u32x r3 = h32_from_64 (digest[3]);
322 #include VECT_COMPARE_M
326 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m01720_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)
330 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m01720_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)
334 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m01720_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)
340 const u32 lid = get_local_id (0);
346 const u32 gid = get_global_id (0);
348 if (gid >= gid_max) return;
352 pw_buf0[0] = pws[gid].i[ 0];
353 pw_buf0[1] = pws[gid].i[ 1];
354 pw_buf0[2] = pws[gid].i[ 2];
355 pw_buf0[3] = pws[gid].i[ 3];
359 pw_buf1[0] = pws[gid].i[ 4];
360 pw_buf1[1] = pws[gid].i[ 5];
361 pw_buf1[2] = pws[gid].i[ 6];
362 pw_buf1[3] = pws[gid].i[ 7];
364 const u32 pw_len = pws[gid].pw_len;
372 salt_buf0[0] = salt_bufs[salt_pos].salt_buf[0];
373 salt_buf0[1] = salt_bufs[salt_pos].salt_buf[1];
374 salt_buf0[2] = salt_bufs[salt_pos].salt_buf[2];
375 salt_buf0[3] = salt_bufs[salt_pos].salt_buf[3];
379 salt_buf1[0] = salt_bufs[salt_pos].salt_buf[4];
380 salt_buf1[1] = salt_bufs[salt_pos].salt_buf[5];
381 salt_buf1[2] = salt_bufs[salt_pos].salt_buf[6];
382 salt_buf1[3] = salt_bufs[salt_pos].salt_buf[7];
384 const u32 salt_len = salt_bufs[salt_pos].salt_len;
390 const u32 search[4] =
392 digests_buf[digests_offset].digest_buf[DGST_R0],
393 digests_buf[digests_offset].digest_buf[DGST_R1],
394 digests_buf[digests_offset].digest_buf[DGST_R2],
395 digests_buf[digests_offset].digest_buf[DGST_R3]
402 for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++)
432 const u32 out_len = apply_rules (rules_buf[il_pos].cmds, w0, w1, pw_len);
438 const u32 out_salt_len = out_len + salt_len;
440 switch_buffer_by_offset (w0, w1, w2, w3, salt_len);
442 w0[0] |= salt_buf0[0];
443 w0[1] |= salt_buf0[1];
444 w0[2] |= salt_buf0[2];
445 w0[3] |= salt_buf0[3];
446 w1[0] |= salt_buf1[0];
447 w1[1] |= salt_buf1[1];
448 w1[2] |= salt_buf1[2];
449 w1[3] |= salt_buf1[3];
451 append_0x80_4 (w0, w1, w2, w3, out_salt_len);
462 w0_t[0] = swap_workaround (w0[0]);
463 w0_t[1] = swap_workaround (w0[1]);
464 w0_t[2] = swap_workaround (w0[2]);
465 w0_t[3] = swap_workaround (w0[3]);
466 w1_t[0] = swap_workaround (w1[0]);
467 w1_t[1] = swap_workaround (w1[1]);
468 w1_t[2] = swap_workaround (w1[2]);
469 w1_t[3] = swap_workaround (w1[3]);
470 w2_t[0] = swap_workaround (w2[0]);
471 w2_t[1] = swap_workaround (w2[1]);
472 w2_t[2] = swap_workaround (w2[2]);
473 w2_t[3] = swap_workaround (w2[3]);
474 w3_t[0] = swap_workaround (w3[0]);
475 w3_t[1] = swap_workaround (w3[1]);
477 w3_t[3] = out_salt_len * 8;
481 digest[0] = SHA512M_A;
482 digest[1] = SHA512M_B;
483 digest[2] = SHA512M_C;
484 digest[3] = SHA512M_D;
485 digest[4] = SHA512M_E;
486 digest[5] = SHA512M_F;
487 digest[6] = SHA512M_G;
488 digest[7] = SHA512M_H;
490 sha512_transform (w0_t, w1_t, w2_t, w3_t, digest);
493 const u32x r0 = l32_from_64 (digest[7]);
494 const u32x r1 = h32_from_64 (digest[7]);
495 const u32x r2 = l32_from_64 (digest[3]);
496 const u32x r3 = h32_from_64 (digest[3]);
498 #include VECT_COMPARE_S
502 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m01720_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)
506 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m01720_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)