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"
37 #define VECT_COMPARE_M "check_multi_vect2_comp4.c"
41 #define VECT_COMPARE_M "check_multi_vect4_comp4.c"
44 static void sha1_transform (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u32x digest[5])
72 SHA1_STEP (SHA1_F0o, A, B, C, D, E, w0_t);
73 SHA1_STEP (SHA1_F0o, E, A, B, C, D, w1_t);
74 SHA1_STEP (SHA1_F0o, D, E, A, B, C, w2_t);
75 SHA1_STEP (SHA1_F0o, C, D, E, A, B, w3_t);
76 SHA1_STEP (SHA1_F0o, B, C, D, E, A, w4_t);
77 SHA1_STEP (SHA1_F0o, A, B, C, D, E, w5_t);
78 SHA1_STEP (SHA1_F0o, E, A, B, C, D, w6_t);
79 SHA1_STEP (SHA1_F0o, D, E, A, B, C, w7_t);
80 SHA1_STEP (SHA1_F0o, C, D, E, A, B, w8_t);
81 SHA1_STEP (SHA1_F0o, B, C, D, E, A, w9_t);
82 SHA1_STEP (SHA1_F0o, A, B, C, D, E, wa_t);
83 SHA1_STEP (SHA1_F0o, E, A, B, C, D, wb_t);
84 SHA1_STEP (SHA1_F0o, D, E, A, B, C, wc_t);
85 SHA1_STEP (SHA1_F0o, C, D, E, A, B, wd_t);
86 SHA1_STEP (SHA1_F0o, B, C, D, E, A, we_t);
87 SHA1_STEP (SHA1_F0o, A, B, C, D, E, wf_t);
88 w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F0o, E, A, B, C, D, w0_t);
89 w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F0o, D, E, A, B, C, w1_t);
90 w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F0o, C, D, E, A, B, w2_t);
91 w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F0o, B, C, D, E, A, w3_t);
96 w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w4_t);
97 w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w5_t);
98 w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w6_t);
99 w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w7_t);
100 w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w8_t);
101 w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w9_t);
102 wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wa_t);
103 wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, wb_t);
104 wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, wc_t);
105 wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, wd_t);
106 we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, we_t);
107 wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wf_t);
108 w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w0_t);
109 w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w1_t);
110 w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w2_t);
111 w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w3_t);
112 w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w4_t);
113 w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w5_t);
114 w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w6_t);
115 w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w7_t);
120 w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, w8_t);
121 w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, w9_t);
122 wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, wa_t);
123 wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, wb_t);
124 wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, wc_t);
125 wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, wd_t);
126 we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, we_t);
127 wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, wf_t);
128 w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, w0_t);
129 w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, w1_t);
130 w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, w2_t);
131 w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, w3_t);
132 w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, w4_t);
133 w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, w5_t);
134 w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, w6_t);
135 w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, w7_t);
136 w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, w8_t);
137 w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, w9_t);
138 wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, wa_t);
139 wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, wb_t);
144 wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, wc_t);
145 wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wd_t);
146 we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, we_t);
147 wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, wf_t);
148 w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w0_t);
149 w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w1_t);
150 w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w2_t);
151 w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w3_t);
152 w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w4_t);
153 w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w5_t);
154 w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w6_t);
155 w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w7_t);
156 w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w8_t);
157 w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w9_t);
158 wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, wa_t);
159 wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, wb_t);
160 wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wc_t);
161 wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, wd_t);
162 we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, we_t);
163 wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, wf_t);
172 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10300_init (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global saph_sha1_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)
178 const u32 gid = get_global_id (0);
180 if (gid >= gid_max) return;
184 word_buf0[0] = pws[gid].i[0];
185 word_buf0[1] = pws[gid].i[1];
186 word_buf0[2] = pws[gid].i[2];
187 word_buf0[3] = pws[gid].i[3];
191 word_buf1[0] = pws[gid].i[4];
192 word_buf1[1] = pws[gid].i[5];
193 word_buf1[2] = pws[gid].i[6];
194 word_buf1[3] = pws[gid].i[7];
198 word_buf2[0] = pws[gid].i[8];
199 word_buf2[1] = pws[gid].i[9];
201 const u32 pw_len = pws[gid].pw_len;
207 u32 salt_len = salt_bufs[salt_pos].salt_len;
211 salt_buf[0] = salt_bufs[salt_pos].salt_buf[0];
212 salt_buf[1] = salt_bufs[salt_pos].salt_buf[1];
213 salt_buf[2] = salt_bufs[salt_pos].salt_buf[2];
214 salt_buf[3] = salt_bufs[salt_pos].salt_buf[3];
248 switch_buffer_by_offset (w0, w1, w2, w3, pw_len);
250 w0[0] |= word_buf0[0];
251 w0[1] |= word_buf0[1];
252 w0[2] |= word_buf0[2];
253 w0[3] |= word_buf0[3];
255 w1[0] |= word_buf1[0];
256 w1[1] |= word_buf1[1];
257 w1[2] |= word_buf1[2];
258 w1[3] |= word_buf1[3];
260 w2[0] |= word_buf2[0];
261 w2[1] |= word_buf2[1];
263 const u32 pw_salt_len = pw_len + salt_len;
265 append_0x80_4 (w0, w1, w2, w3, pw_salt_len);
269 w0[0] = swap_workaround (w0[0]);
270 w0[1] = swap_workaround (w0[1]);
271 w0[2] = swap_workaround (w0[2]);
272 w0[3] = swap_workaround (w0[3]);
274 w1[0] = swap_workaround (w1[0]);
275 w1[1] = swap_workaround (w1[1]);
276 w1[2] = swap_workaround (w1[2]);
277 w1[3] = swap_workaround (w1[3]);
279 w2[0] = swap_workaround (w2[0]);
280 w2[1] = swap_workaround (w2[1]);
281 w2[2] = swap_workaround (w2[2]);
282 w2[3] = swap_workaround (w2[3]);
284 w3[0] = swap_workaround (w3[0]);
285 w3[1] = swap_workaround (w3[1]);
286 w3[2] = swap_workaround (w3[2]);
287 w3[3] = pw_salt_len * 8;
297 sha1_transform (w0, w1, w2, w3, digest);
299 tmps[gid].digest_buf[0] = digest[0];
300 tmps[gid].digest_buf[1] = digest[1];
301 tmps[gid].digest_buf[2] = digest[2];
302 tmps[gid].digest_buf[3] = digest[3];
303 tmps[gid].digest_buf[4] = digest[4];
306 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10300_loop (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global saph_sha1_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)
312 const u32 gid = get_global_id (0);
316 word_buf0[0] = swap_workaround (pws[gid].i[0]);
317 word_buf0[1] = swap_workaround (pws[gid].i[1]);
318 word_buf0[2] = swap_workaround (pws[gid].i[2]);
319 word_buf0[3] = swap_workaround (pws[gid].i[3]);
323 word_buf1[0] = swap_workaround (pws[gid].i[4]);
324 word_buf1[1] = swap_workaround (pws[gid].i[5]);
325 word_buf1[2] = swap_workaround (pws[gid].i[6]);
326 word_buf1[3] = swap_workaround (pws[gid].i[7]);
330 word_buf2[0] = swap_workaround (pws[gid].i[8]);
331 word_buf2[1] = swap_workaround (pws[gid].i[9]);
333 const u32 pw_len = pws[gid].pw_len;
337 digest[0] = tmps[gid].digest_buf[0];
338 digest[1] = tmps[gid].digest_buf[1];
339 digest[2] = tmps[gid].digest_buf[2];
340 digest[3] = tmps[gid].digest_buf[3];
341 digest[4] = tmps[gid].digest_buf[4];
347 for (u32 i = 0; i < loop_cnt; i++)
371 switch_buffer_by_offset_be (w0, w1, w2, w3, pw_len);
373 w0[0] |= word_buf0[0];
374 w0[1] |= word_buf0[1];
375 w0[2] |= word_buf0[2];
376 w0[3] |= word_buf0[3];
377 w1[0] |= word_buf1[0];
378 w1[1] |= word_buf1[1];
379 w1[2] |= word_buf1[2];
380 w1[3] |= word_buf1[3];
381 w2[0] |= word_buf2[0];
382 w2[1] |= word_buf2[1];
387 w3[3] = (pw_len + 20) * 8;
395 sha1_transform (w0, w1, w2, w3, digest);
398 tmps[gid].digest_buf[0] = digest[0];
399 tmps[gid].digest_buf[1] = digest[1];
400 tmps[gid].digest_buf[2] = digest[2];
401 tmps[gid].digest_buf[3] = digest[3];
402 tmps[gid].digest_buf[4] = digest[4];
405 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10300_comp (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global saph_sha1_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)
411 const u32 gid = get_global_id (0);
413 if (gid >= gid_max) return;
415 const u32 lid = get_local_id (0);
421 const u32x r0 = tmps[gid].digest_buf[0];
422 const u32x r1 = tmps[gid].digest_buf[1];
423 const u32x r2 = tmps[gid].digest_buf[2];
424 const u32x r3 = tmps[gid].digest_buf[3];
428 #include VECT_COMPARE_M