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"
33 #define VECT_COMPARE_M "check_multi_vect1_comp4.c"
36 __constant u64 k_sha512[80] =
38 SHA512C00, SHA512C01, SHA512C02, SHA512C03,
39 SHA512C04, SHA512C05, SHA512C06, SHA512C07,
40 SHA512C08, SHA512C09, SHA512C0a, SHA512C0b,
41 SHA512C0c, SHA512C0d, SHA512C0e, SHA512C0f,
42 SHA512C10, SHA512C11, SHA512C12, SHA512C13,
43 SHA512C14, SHA512C15, SHA512C16, SHA512C17,
44 SHA512C18, SHA512C19, SHA512C1a, SHA512C1b,
45 SHA512C1c, SHA512C1d, SHA512C1e, SHA512C1f,
46 SHA512C20, SHA512C21, SHA512C22, SHA512C23,
47 SHA512C24, SHA512C25, SHA512C26, SHA512C27,
48 SHA512C28, SHA512C29, SHA512C2a, SHA512C2b,
49 SHA512C2c, SHA512C2d, SHA512C2e, SHA512C2f,
50 SHA512C30, SHA512C31, SHA512C32, SHA512C33,
51 SHA512C34, SHA512C35, SHA512C36, SHA512C37,
52 SHA512C38, SHA512C39, SHA512C3a, SHA512C3b,
53 SHA512C3c, SHA512C3d, SHA512C3e, SHA512C3f,
54 SHA512C40, SHA512C41, SHA512C42, SHA512C43,
55 SHA512C44, SHA512C45, SHA512C46, SHA512C47,
56 SHA512C48, SHA512C49, SHA512C4a, SHA512C4b,
57 SHA512C4c, SHA512C4d, SHA512C4e, SHA512C4f,
60 static void sha512_transform (const u64 w[16], u64 dgst[8])
88 #define ROUND_EXPAND() \
90 w0_t = SHA512_EXPAND (we_t, w9_t, w1_t, w0_t); \
91 w1_t = SHA512_EXPAND (wf_t, wa_t, w2_t, w1_t); \
92 w2_t = SHA512_EXPAND (w0_t, wb_t, w3_t, w2_t); \
93 w3_t = SHA512_EXPAND (w1_t, wc_t, w4_t, w3_t); \
94 w4_t = SHA512_EXPAND (w2_t, wd_t, w5_t, w4_t); \
95 w5_t = SHA512_EXPAND (w3_t, we_t, w6_t, w5_t); \
96 w6_t = SHA512_EXPAND (w4_t, wf_t, w7_t, w6_t); \
97 w7_t = SHA512_EXPAND (w5_t, w0_t, w8_t, w7_t); \
98 w8_t = SHA512_EXPAND (w6_t, w1_t, w9_t, w8_t); \
99 w9_t = SHA512_EXPAND (w7_t, w2_t, wa_t, w9_t); \
100 wa_t = SHA512_EXPAND (w8_t, w3_t, wb_t, wa_t); \
101 wb_t = SHA512_EXPAND (w9_t, w4_t, wc_t, wb_t); \
102 wc_t = SHA512_EXPAND (wa_t, w5_t, wd_t, wc_t); \
103 wd_t = SHA512_EXPAND (wb_t, w6_t, we_t, wd_t); \
104 we_t = SHA512_EXPAND (wc_t, w7_t, wf_t, we_t); \
105 wf_t = SHA512_EXPAND (wd_t, w8_t, w0_t, wf_t); \
108 #define ROUND_STEP(i) \
110 SHA512_STEP (SHA512_F0o, SHA512_F1o, a, b, c, d, e, f, g, h, w0_t, k_sha512[i + 0]); \
111 SHA512_STEP (SHA512_F0o, SHA512_F1o, h, a, b, c, d, e, f, g, w1_t, k_sha512[i + 1]); \
112 SHA512_STEP (SHA512_F0o, SHA512_F1o, g, h, a, b, c, d, e, f, w2_t, k_sha512[i + 2]); \
113 SHA512_STEP (SHA512_F0o, SHA512_F1o, f, g, h, a, b, c, d, e, w3_t, k_sha512[i + 3]); \
114 SHA512_STEP (SHA512_F0o, SHA512_F1o, e, f, g, h, a, b, c, d, w4_t, k_sha512[i + 4]); \
115 SHA512_STEP (SHA512_F0o, SHA512_F1o, d, e, f, g, h, a, b, c, w5_t, k_sha512[i + 5]); \
116 SHA512_STEP (SHA512_F0o, SHA512_F1o, c, d, e, f, g, h, a, b, w6_t, k_sha512[i + 6]); \
117 SHA512_STEP (SHA512_F0o, SHA512_F1o, b, c, d, e, f, g, h, a, w7_t, k_sha512[i + 7]); \
118 SHA512_STEP (SHA512_F0o, SHA512_F1o, a, b, c, d, e, f, g, h, w8_t, k_sha512[i + 8]); \
119 SHA512_STEP (SHA512_F0o, SHA512_F1o, h, a, b, c, d, e, f, g, w9_t, k_sha512[i + 9]); \
120 SHA512_STEP (SHA512_F0o, SHA512_F1o, g, h, a, b, c, d, e, f, wa_t, k_sha512[i + 10]); \
121 SHA512_STEP (SHA512_F0o, SHA512_F1o, f, g, h, a, b, c, d, e, wb_t, k_sha512[i + 11]); \
122 SHA512_STEP (SHA512_F0o, SHA512_F1o, e, f, g, h, a, b, c, d, wc_t, k_sha512[i + 12]); \
123 SHA512_STEP (SHA512_F0o, SHA512_F1o, d, e, f, g, h, a, b, c, wd_t, k_sha512[i + 13]); \
124 SHA512_STEP (SHA512_F0o, SHA512_F1o, c, d, e, f, g, h, a, b, we_t, k_sha512[i + 14]); \
125 SHA512_STEP (SHA512_F0o, SHA512_F1o, b, c, d, e, f, g, h, a, wf_t, k_sha512[i + 15]); \
131 for (int i = 16; i < 80; i += 16)
133 ROUND_EXPAND (); ROUND_STEP (i);
146 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m07900_init (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global drupal7_tmp_t *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)
152 const u32 gid = get_global_id (0);
154 if (gid >= gid_max) return;
158 w0[0] = pws[gid].i[ 0];
159 w0[1] = pws[gid].i[ 1];
160 w0[2] = pws[gid].i[ 2];
161 w0[3] = pws[gid].i[ 3];
165 w1[0] = pws[gid].i[ 4];
166 w1[1] = pws[gid].i[ 5];
167 w1[2] = pws[gid].i[ 6];
168 w1[3] = pws[gid].i[ 7];
172 w2[0] = pws[gid].i[ 8];
173 w2[1] = pws[gid].i[ 9];
174 w2[2] = pws[gid].i[10];
175 w2[3] = pws[gid].i[11];
184 const u32 pw_len = pws[gid].pw_len;
186 append_0x80_4 (w0, w1, w2, w3, pw_len);
194 salt_buf[0] = salt_bufs[salt_pos].salt_buf[0];
195 salt_buf[1] = salt_bufs[salt_pos].salt_buf[1];
201 u32 block_len = 8 + pw_len;
205 w[ 0] = ((u64) swap_workaround (salt_buf[0])) << 32 | (u64) swap_workaround (salt_buf[1]);
206 w[ 1] = ((u64) swap_workaround (w0[0])) << 32 | (u64) swap_workaround (w0[1]);
207 w[ 2] = ((u64) swap_workaround (w0[2])) << 32 | (u64) swap_workaround (w0[3]);
208 w[ 3] = ((u64) swap_workaround (w1[0])) << 32 | (u64) swap_workaround (w1[1]);
209 w[ 4] = ((u64) swap_workaround (w1[2])) << 32 | (u64) swap_workaround (w1[3]);
210 w[ 5] = ((u64) swap_workaround (w2[0])) << 32 | (u64) swap_workaround (w2[1]);
211 w[ 6] = ((u64) swap_workaround (w2[2])) << 32 | (u64) swap_workaround (w2[3]);
212 w[ 7] = ((u64) swap_workaround (w3[0])) << 32 | (u64) swap_workaround (w3[1]);
213 w[ 8] = ((u64) swap_workaround (w3[2])) << 32 | (u64) swap_workaround (w3[3]);
220 w[15] = block_len * 8;
228 digest[0] = SHA512M_A;
229 digest[1] = SHA512M_B;
230 digest[2] = SHA512M_C;
231 digest[3] = SHA512M_D;
232 digest[4] = SHA512M_E;
233 digest[5] = SHA512M_F;
234 digest[6] = SHA512M_G;
235 digest[7] = SHA512M_H;
237 sha512_transform (w, digest);
239 tmps[gid].digest_buf[0] = digest[0];
240 tmps[gid].digest_buf[1] = digest[1];
241 tmps[gid].digest_buf[2] = digest[2];
242 tmps[gid].digest_buf[3] = digest[3];
243 tmps[gid].digest_buf[4] = digest[4];
244 tmps[gid].digest_buf[5] = digest[5];
245 tmps[gid].digest_buf[6] = digest[6];
246 tmps[gid].digest_buf[7] = digest[7];
249 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m07900_loop (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global drupal7_tmp_t *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)
255 const u32 gid = get_global_id (0);
257 if (gid >= gid_max) return;
261 w0[0] = pws[gid].i[ 0];
262 w0[1] = pws[gid].i[ 1];
263 w0[2] = pws[gid].i[ 2];
264 w0[3] = pws[gid].i[ 3];
268 w1[0] = pws[gid].i[ 4];
269 w1[1] = pws[gid].i[ 5];
270 w1[2] = pws[gid].i[ 6];
271 w1[3] = pws[gid].i[ 7];
275 w2[0] = pws[gid].i[ 8];
276 w2[1] = pws[gid].i[ 9];
277 w2[2] = pws[gid].i[10];
278 w2[3] = pws[gid].i[11];
287 const u32 pw_len = pws[gid].pw_len;
289 append_0x80_4 (w0, w1, w2, w3, pw_len);
297 digest[0] = tmps[gid].digest_buf[0];
298 digest[1] = tmps[gid].digest_buf[1];
299 digest[2] = tmps[gid].digest_buf[2];
300 digest[3] = tmps[gid].digest_buf[3];
301 digest[4] = tmps[gid].digest_buf[4];
302 digest[5] = tmps[gid].digest_buf[5];
303 digest[6] = tmps[gid].digest_buf[6];
304 digest[7] = tmps[gid].digest_buf[7];
310 u32 block_len = (64 + pw_len);
322 w[ 8] = ((u64) swap_workaround (w0[0])) << 32 | (u64) swap_workaround (w0[1]);
323 w[ 9] = ((u64) swap_workaround (w0[2])) << 32 | (u64) swap_workaround (w0[3]);
324 w[10] = ((u64) swap_workaround (w1[0])) << 32 | (u64) swap_workaround (w1[1]);
325 w[11] = ((u64) swap_workaround (w1[2])) << 32 | (u64) swap_workaround (w1[3]);
326 w[12] = ((u64) swap_workaround (w2[0])) << 32 | (u64) swap_workaround (w2[1]);
327 w[13] = ((u64) swap_workaround (w2[2])) << 32 | (u64) swap_workaround (w2[3]);
329 w[15] = block_len * 8;
335 for (u32 i = 0; i < loop_cnt; i++)
346 digest[0] = SHA512M_A;
347 digest[1] = SHA512M_B;
348 digest[2] = SHA512M_C;
349 digest[3] = SHA512M_D;
350 digest[4] = SHA512M_E;
351 digest[5] = SHA512M_F;
352 digest[6] = SHA512M_G;
353 digest[7] = SHA512M_H;
355 sha512_transform (w, digest);
358 tmps[gid].digest_buf[0] = digest[0];
359 tmps[gid].digest_buf[1] = digest[1];
360 tmps[gid].digest_buf[2] = digest[2];
361 tmps[gid].digest_buf[3] = digest[3];
362 tmps[gid].digest_buf[4] = digest[4];
363 tmps[gid].digest_buf[5] = digest[5];
364 tmps[gid].digest_buf[6] = digest[6];
365 tmps[gid].digest_buf[7] = digest[7];
368 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m07900_comp (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global drupal7_tmp_t *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)
374 const u32 gid = get_global_id (0);
376 if (gid >= gid_max) return;
378 const u32 lid = get_local_id (0);
384 const u32x r0 = l32_from_64 (tmps[gid].digest_buf[0]);
385 const u32x r1 = h32_from_64 (tmps[gid].digest_buf[0]);
386 const u32x r2 = l32_from_64 (tmps[gid].digest_buf[1]);
387 const u32x r3 = h32_from_64 (tmps[gid].digest_buf[1]);
391 #include VECT_COMPARE_M