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"
29 #define VECT_COMPARE_M "check_multi_vect1_comp4.c"
33 #define VECT_COMPARE_M "check_multi_vect2_comp4.c"
36 __device__ static void sha1_transform (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u32x digest[5])
64 SHA1_STEP (SHA1_F0o, A, B, C, D, E, w0_t);
65 SHA1_STEP (SHA1_F0o, E, A, B, C, D, w1_t);
66 SHA1_STEP (SHA1_F0o, D, E, A, B, C, w2_t);
67 SHA1_STEP (SHA1_F0o, C, D, E, A, B, w3_t);
68 SHA1_STEP (SHA1_F0o, B, C, D, E, A, w4_t);
69 SHA1_STEP (SHA1_F0o, A, B, C, D, E, w5_t);
70 SHA1_STEP (SHA1_F0o, E, A, B, C, D, w6_t);
71 SHA1_STEP (SHA1_F0o, D, E, A, B, C, w7_t);
72 SHA1_STEP (SHA1_F0o, C, D, E, A, B, w8_t);
73 SHA1_STEP (SHA1_F0o, B, C, D, E, A, w9_t);
74 SHA1_STEP (SHA1_F0o, A, B, C, D, E, wa_t);
75 SHA1_STEP (SHA1_F0o, E, A, B, C, D, wb_t);
76 SHA1_STEP (SHA1_F0o, D, E, A, B, C, wc_t);
77 SHA1_STEP (SHA1_F0o, C, D, E, A, B, wd_t);
78 SHA1_STEP (SHA1_F0o, B, C, D, E, A, we_t);
79 SHA1_STEP (SHA1_F0o, A, B, C, D, E, wf_t);
80 w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F0o, E, A, B, C, D, w0_t);
81 w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F0o, D, E, A, B, C, w1_t);
82 w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F0o, C, D, E, A, B, w2_t);
83 w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F0o, B, C, D, E, A, w3_t);
88 w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w4_t);
89 w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w5_t);
90 w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w6_t);
91 w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w7_t);
92 w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w8_t);
93 w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w9_t);
94 wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wa_t);
95 wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, wb_t);
96 wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, wc_t);
97 wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, wd_t);
98 we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, we_t);
99 wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wf_t);
100 w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w0_t);
101 w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w1_t);
102 w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w2_t);
103 w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w3_t);
104 w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w4_t);
105 w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w5_t);
106 w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w6_t);
107 w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w7_t);
112 w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, w8_t);
113 w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, w9_t);
114 wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, wa_t);
115 wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, wb_t);
116 wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, wc_t);
117 wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, wd_t);
118 we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, we_t);
119 wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, wf_t);
120 w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, w0_t);
121 w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, w1_t);
122 w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, w2_t);
123 w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, w3_t);
124 w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, w4_t);
125 w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, w5_t);
126 w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, w6_t);
127 w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, w7_t);
128 w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, w8_t);
129 w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, w9_t);
130 wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, wa_t);
131 wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, wb_t);
136 wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, wc_t);
137 wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wd_t);
138 we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, we_t);
139 wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, wf_t);
140 w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w0_t);
141 w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w1_t);
142 w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w2_t);
143 w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w3_t);
144 w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w4_t);
145 w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w5_t);
146 w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w6_t);
147 w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w7_t);
148 w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w8_t);
149 w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w9_t);
150 wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, wa_t);
151 wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, wb_t);
152 wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wc_t);
153 wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, wd_t);
154 we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, we_t);
155 wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, wf_t);
164 extern "C" __global__ void __launch_bounds__ (256, 1) m10300_init (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, saph_sha1_tmp_t *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)
170 const u32 gid = (blockIdx.x * blockDim.x) + threadIdx.x;
172 if (gid >= gid_max) return;
176 word_buf0[0] = pws[gid].i[0];
177 word_buf0[1] = pws[gid].i[1];
178 word_buf0[2] = pws[gid].i[2];
179 word_buf0[3] = pws[gid].i[3];
183 word_buf1[0] = pws[gid].i[4];
184 word_buf1[1] = pws[gid].i[5];
185 word_buf1[2] = pws[gid].i[6];
186 word_buf1[3] = pws[gid].i[7];
190 word_buf2[0] = pws[gid].i[8];
191 word_buf2[1] = pws[gid].i[9];
193 const u32 pw_len = pws[gid].pw_len;
199 u32 salt_len = salt_bufs[salt_pos].salt_len;
203 salt_buf[0] = salt_bufs[salt_pos].salt_buf[0];
204 salt_buf[1] = salt_bufs[salt_pos].salt_buf[1];
205 salt_buf[2] = salt_bufs[salt_pos].salt_buf[2];
206 salt_buf[3] = salt_bufs[salt_pos].salt_buf[3];
240 switch_buffer_by_offset (w0, w1, w2, w3, pw_len);
242 w0[0] |= word_buf0[0];
243 w0[1] |= word_buf0[1];
244 w0[2] |= word_buf0[2];
245 w0[3] |= word_buf0[3];
247 w1[0] |= word_buf1[0];
248 w1[1] |= word_buf1[1];
249 w1[2] |= word_buf1[2];
250 w1[3] |= word_buf1[3];
252 w2[0] |= word_buf2[0];
253 w2[1] |= word_buf2[1];
255 const u32 pw_salt_len = pw_len + salt_len;
257 append_0x80_4 (w0, w1, w2, w3, pw_salt_len);
261 w0[0] = swap_workaround (w0[0]);
262 w0[1] = swap_workaround (w0[1]);
263 w0[2] = swap_workaround (w0[2]);
264 w0[3] = swap_workaround (w0[3]);
266 w1[0] = swap_workaround (w1[0]);
267 w1[1] = swap_workaround (w1[1]);
268 w1[2] = swap_workaround (w1[2]);
269 w1[3] = swap_workaround (w1[3]);
271 w2[0] = swap_workaround (w2[0]);
272 w2[1] = swap_workaround (w2[1]);
273 w2[2] = swap_workaround (w2[2]);
274 w2[3] = swap_workaround (w2[3]);
276 w3[0] = swap_workaround (w3[0]);
277 w3[1] = swap_workaround (w3[1]);
278 w3[2] = swap_workaround (w3[2]);
279 w3[3] = pw_salt_len * 8;
289 sha1_transform (w0, w1, w2, w3, digest);
291 tmps[gid].digest_buf[0] = digest[0];
292 tmps[gid].digest_buf[1] = digest[1];
293 tmps[gid].digest_buf[2] = digest[2];
294 tmps[gid].digest_buf[3] = digest[3];
295 tmps[gid].digest_buf[4] = digest[4];
298 extern "C" __global__ void __launch_bounds__ (256, 1) m10300_loop (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, saph_sha1_tmp_t *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)
304 const u32 gid = (blockIdx.x * blockDim.x) + threadIdx.x;
308 word_buf0[0] = swap_workaround (pws[gid].i[0]);
309 word_buf0[1] = swap_workaround (pws[gid].i[1]);
310 word_buf0[2] = swap_workaround (pws[gid].i[2]);
311 word_buf0[3] = swap_workaround (pws[gid].i[3]);
315 word_buf1[0] = swap_workaround (pws[gid].i[4]);
316 word_buf1[1] = swap_workaround (pws[gid].i[5]);
317 word_buf1[2] = swap_workaround (pws[gid].i[6]);
318 word_buf1[3] = swap_workaround (pws[gid].i[7]);
322 word_buf2[0] = swap_workaround (pws[gid].i[8]);
323 word_buf2[1] = swap_workaround (pws[gid].i[9]);
325 const u32 pw_len = pws[gid].pw_len;
329 digest[0] = tmps[gid].digest_buf[0];
330 digest[1] = tmps[gid].digest_buf[1];
331 digest[2] = tmps[gid].digest_buf[2];
332 digest[3] = tmps[gid].digest_buf[3];
333 digest[4] = tmps[gid].digest_buf[4];
339 for (u32 i = 0; i < loop_cnt; i++)
363 switch_buffer_by_offset_be (w0, w1, w2, w3, pw_len);
365 w0[0] |= word_buf0[0];
366 w0[1] |= word_buf0[1];
367 w0[2] |= word_buf0[2];
368 w0[3] |= word_buf0[3];
369 w1[0] |= word_buf1[0];
370 w1[1] |= word_buf1[1];
371 w1[2] |= word_buf1[2];
372 w1[3] |= word_buf1[3];
373 w2[0] |= word_buf2[0];
374 w2[1] |= word_buf2[1];
379 w3[3] = (pw_len + 20) * 8;
387 sha1_transform (w0, w1, w2, w3, digest);
390 tmps[gid].digest_buf[0] = digest[0];
391 tmps[gid].digest_buf[1] = digest[1];
392 tmps[gid].digest_buf[2] = digest[2];
393 tmps[gid].digest_buf[3] = digest[3];
394 tmps[gid].digest_buf[4] = digest[4];
397 extern "C" __global__ void __launch_bounds__ (256, 1) m10300_comp (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, saph_sha1_tmp_t *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)
403 const u32 gid = (blockIdx.x * blockDim.x) + threadIdx.x;
405 if (gid >= gid_max) return;
407 const u32 lid = threadIdx.x;
413 const u32x r0 = tmps[gid].digest_buf[0];
414 const u32x r1 = tmps[gid].digest_buf[1];
415 const u32x r2 = tmps[gid].digest_buf[2];
416 const u32x r3 = tmps[gid].digest_buf[3];
420 #include VECT_COMPARE_M