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_S "check_single_vect1_comp4.c"
34 #define VECT_COMPARE_M "check_multi_vect1_comp4.c"
38 #define VECT_COMPARE_S "check_single_vect2_comp4.c"
39 #define VECT_COMPARE_M "check_multi_vect2_comp4.c"
43 #define VECT_COMPARE_S "check_single_vect4_comp4.c"
44 #define VECT_COMPARE_M "check_multi_vect4_comp4.c"
47 static void sha1_transform (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u32x digest[5])
75 SHA1_STEP (SHA1_F0o, A, B, C, D, E, w0_t);
76 SHA1_STEP (SHA1_F0o, E, A, B, C, D, w1_t);
77 SHA1_STEP (SHA1_F0o, D, E, A, B, C, w2_t);
78 SHA1_STEP (SHA1_F0o, C, D, E, A, B, w3_t);
79 SHA1_STEP (SHA1_F0o, B, C, D, E, A, w4_t);
80 SHA1_STEP (SHA1_F0o, A, B, C, D, E, w5_t);
81 SHA1_STEP (SHA1_F0o, E, A, B, C, D, w6_t);
82 SHA1_STEP (SHA1_F0o, D, E, A, B, C, w7_t);
83 SHA1_STEP (SHA1_F0o, C, D, E, A, B, w8_t);
84 SHA1_STEP (SHA1_F0o, B, C, D, E, A, w9_t);
85 SHA1_STEP (SHA1_F0o, A, B, C, D, E, wa_t);
86 SHA1_STEP (SHA1_F0o, E, A, B, C, D, wb_t);
87 SHA1_STEP (SHA1_F0o, D, E, A, B, C, wc_t);
88 SHA1_STEP (SHA1_F0o, C, D, E, A, B, wd_t);
89 SHA1_STEP (SHA1_F0o, B, C, D, E, A, we_t);
90 SHA1_STEP (SHA1_F0o, A, B, C, D, E, wf_t);
91 w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F0o, E, A, B, C, D, w0_t);
92 w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F0o, D, E, A, B, C, w1_t);
93 w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F0o, C, D, E, A, B, w2_t);
94 w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F0o, B, C, D, E, A, w3_t);
99 w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w4_t);
100 w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w5_t);
101 w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w6_t);
102 w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w7_t);
103 w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w8_t);
104 w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w9_t);
105 wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wa_t);
106 wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, wb_t);
107 wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, wc_t);
108 wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, wd_t);
109 we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, we_t);
110 wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wf_t);
111 w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w0_t);
112 w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w1_t);
113 w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w2_t);
114 w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w3_t);
115 w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w4_t);
116 w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w5_t);
117 w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w6_t);
118 w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w7_t);
123 w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, w8_t);
124 w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, w9_t);
125 wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, wa_t);
126 wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, wb_t);
127 wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, wc_t);
128 wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, wd_t);
129 we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, we_t);
130 wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, wf_t);
131 w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, w0_t);
132 w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, w1_t);
133 w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, w2_t);
134 w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, w3_t);
135 w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, w4_t);
136 w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, w5_t);
137 w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, w6_t);
138 w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, w7_t);
139 w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, w8_t);
140 w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, w9_t);
141 wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, wa_t);
142 wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, wb_t);
147 wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, wc_t);
148 wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wd_t);
149 we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, we_t);
150 wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, wf_t);
151 w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w0_t);
152 w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w1_t);
153 w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w2_t);
154 w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w3_t);
155 w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w4_t);
156 w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w5_t);
157 w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w6_t);
158 w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w7_t);
159 w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w8_t);
160 w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w9_t);
161 wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, wa_t);
162 wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, wb_t);
163 wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wc_t);
164 wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, wd_t);
165 we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, we_t);
166 wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, wf_t);
175 static void hmac_sha1_pad (u32x w0[4], u32x w1[4], u32x w2[4], u32x w3[4], u32x ipad[5], u32x opad[5])
177 w0[0] = w0[0] ^ 0x36363636;
178 w0[1] = w0[1] ^ 0x36363636;
179 w0[2] = w0[2] ^ 0x36363636;
180 w0[3] = w0[3] ^ 0x36363636;
181 w1[0] = w1[0] ^ 0x36363636;
182 w1[1] = w1[1] ^ 0x36363636;
183 w1[2] = w1[2] ^ 0x36363636;
184 w1[3] = w1[3] ^ 0x36363636;
185 w2[0] = w2[0] ^ 0x36363636;
186 w2[1] = w2[1] ^ 0x36363636;
187 w2[2] = w2[2] ^ 0x36363636;
188 w2[3] = w2[3] ^ 0x36363636;
189 w3[0] = w3[0] ^ 0x36363636;
190 w3[1] = w3[1] ^ 0x36363636;
191 w3[2] = w3[2] ^ 0x36363636;
192 w3[3] = w3[3] ^ 0x36363636;
200 sha1_transform (w0, w1, w2, w3, ipad);
202 w0[0] = w0[0] ^ 0x6a6a6a6a;
203 w0[1] = w0[1] ^ 0x6a6a6a6a;
204 w0[2] = w0[2] ^ 0x6a6a6a6a;
205 w0[3] = w0[3] ^ 0x6a6a6a6a;
206 w1[0] = w1[0] ^ 0x6a6a6a6a;
207 w1[1] = w1[1] ^ 0x6a6a6a6a;
208 w1[2] = w1[2] ^ 0x6a6a6a6a;
209 w1[3] = w1[3] ^ 0x6a6a6a6a;
210 w2[0] = w2[0] ^ 0x6a6a6a6a;
211 w2[1] = w2[1] ^ 0x6a6a6a6a;
212 w2[2] = w2[2] ^ 0x6a6a6a6a;
213 w2[3] = w2[3] ^ 0x6a6a6a6a;
214 w3[0] = w3[0] ^ 0x6a6a6a6a;
215 w3[1] = w3[1] ^ 0x6a6a6a6a;
216 w3[2] = w3[2] ^ 0x6a6a6a6a;
217 w3[3] = w3[3] ^ 0x6a6a6a6a;
225 sha1_transform (w0, w1, w2, w3, opad);
228 static void hmac_sha1_run (u32x w0[4], u32x w1[4], u32x w2[4], u32x w3[4], u32x ipad[5], u32x opad[5], u32x digest[5])
236 sha1_transform (w0, w1, w2, w3, digest);
253 w3[3] = (64 + 20) * 8;
261 sha1_transform (w0, w1, w2, w3, digest);
264 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m05400_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 ikepsk_t *ikepsk_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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
270 const u32 lid = get_local_id (0);
276 const u32 gid = get_global_id (0);
280 wordl0[0] = pws[gid].i[ 0];
281 wordl0[1] = pws[gid].i[ 1];
282 wordl0[2] = pws[gid].i[ 2];
283 wordl0[3] = pws[gid].i[ 3];
287 wordl1[0] = pws[gid].i[ 4];
288 wordl1[1] = pws[gid].i[ 5];
289 wordl1[2] = pws[gid].i[ 6];
290 wordl1[3] = pws[gid].i[ 7];
306 const u32 pw_l_len = pws[gid].pw_len;
308 if (combs_mode == COMBINATOR_MODE_BASE_RIGHT)
310 switch_buffer_by_offset (wordl0, wordl1, wordl2, wordl3, combs_buf[0].pw_len);
317 const u32 nr_len = ikepsk_bufs[salt_pos].nr_len;
318 const u32 msg_len = ikepsk_bufs[salt_pos].msg_len;
322 salt_buf0[0] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[ 0]);
323 salt_buf0[1] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[ 1]);
324 salt_buf0[2] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[ 2]);
325 salt_buf0[3] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[ 3]);
329 salt_buf1[0] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[ 4]);
330 salt_buf1[1] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[ 5]);
331 salt_buf1[2] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[ 6]);
332 salt_buf1[3] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[ 7]);
336 salt_buf2[0] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[ 8]);
337 salt_buf2[1] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[ 9]);
338 salt_buf2[2] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[10]);
339 salt_buf2[3] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[11]);
343 salt_buf3[0] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[12]);
344 salt_buf3[1] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[13]);
348 __local u32 s_msg_buf[128];
350 const u32 lid2 = lid * 2;
352 s_msg_buf[lid2 + 0] = swap_workaround (ikepsk_bufs[salt_pos].msg_buf[lid2 + 0]);
353 s_msg_buf[lid2 + 1] = swap_workaround (ikepsk_bufs[salt_pos].msg_buf[lid2 + 1]);
355 barrier (CLK_LOCAL_MEM_FENCE);
357 if (gid >= gid_max) return;
363 for (u32 il_pos = 0; il_pos < combs_cnt; il_pos++)
365 const u32 pw_r_len = combs_buf[il_pos].pw_len;
367 const u32 pw_len = pw_l_len + pw_r_len;
371 wordr0[0] = combs_buf[il_pos].i[0];
372 wordr0[1] = combs_buf[il_pos].i[1];
373 wordr0[2] = combs_buf[il_pos].i[2];
374 wordr0[3] = combs_buf[il_pos].i[3];
378 wordr1[0] = combs_buf[il_pos].i[4];
379 wordr1[1] = combs_buf[il_pos].i[5];
380 wordr1[2] = combs_buf[il_pos].i[6];
381 wordr1[3] = combs_buf[il_pos].i[7];
397 if (combs_mode == COMBINATOR_MODE_BASE_LEFT)
399 switch_buffer_by_offset (wordr0, wordr1, wordr2, wordr3, pw_l_len);
404 w0[0] = wordl0[0] | wordr0[0];
405 w0[1] = wordl0[1] | wordr0[1];
406 w0[2] = wordl0[2] | wordr0[2];
407 w0[3] = wordl0[3] | wordr0[3];
411 w1[0] = wordl1[0] | wordr1[0];
412 w1[1] = wordl1[1] | wordr1[1];
413 w1[2] = wordl1[2] | wordr1[2];
414 w1[3] = wordl1[3] | wordr1[3];
418 w2[0] = wordl2[0] | wordr2[0];
419 w2[1] = wordl2[1] | wordr2[1];
420 w2[2] = wordl2[2] | wordr2[2];
421 w2[3] = wordl2[3] | wordr2[3];
425 w3[0] = wordl3[0] | wordr3[0];
426 w3[1] = wordl3[1] | wordr3[1];
427 w3[2] = wordl3[2] | wordr3[2];
428 w3[3] = wordl3[3] | wordr3[3];
436 w0_t[0] = swap_workaround (w0[0]);
437 w0_t[1] = swap_workaround (w0[1]);
438 w0_t[2] = swap_workaround (w0[2]);
439 w0_t[3] = swap_workaround (w0[3]);
443 w1_t[0] = swap_workaround (w1[0]);
444 w1_t[1] = swap_workaround (w1[1]);
445 w1_t[2] = swap_workaround (w1[2]);
446 w1_t[3] = swap_workaround (w1[3]);
465 hmac_sha1_pad (w0_t, w1_t, w2_t, w3_t, ipad, opad);
467 w0_t[0] = salt_buf0[0];
468 w0_t[1] = salt_buf0[1];
469 w0_t[2] = salt_buf0[2];
470 w0_t[3] = salt_buf0[3];
471 w1_t[0] = salt_buf1[0];
472 w1_t[1] = salt_buf1[1];
473 w1_t[2] = salt_buf1[2];
474 w1_t[3] = salt_buf1[3];
475 w2_t[0] = salt_buf2[0];
476 w2_t[1] = salt_buf2[1];
477 w2_t[2] = salt_buf2[2];
478 w2_t[3] = salt_buf2[3];
479 w3_t[0] = salt_buf3[0];
480 w3_t[1] = salt_buf3[1];
482 w3_t[3] = (64 + nr_len) * 8;
486 hmac_sha1_run (w0_t, w1_t, w2_t, w3_t, ipad, opad, digest);
505 hmac_sha1_pad (w0_t, w1_t, w2_t, w3_t, ipad, opad);
510 for (left = ikepsk_bufs[salt_pos].msg_len, off = 0; left >= 56; left -= 64, off += 16)
512 w0_t[0] = s_msg_buf[off + 0];
513 w0_t[1] = s_msg_buf[off + 1];
514 w0_t[2] = s_msg_buf[off + 2];
515 w0_t[3] = s_msg_buf[off + 3];
516 w1_t[0] = s_msg_buf[off + 4];
517 w1_t[1] = s_msg_buf[off + 5];
518 w1_t[2] = s_msg_buf[off + 6];
519 w1_t[3] = s_msg_buf[off + 7];
520 w2_t[0] = s_msg_buf[off + 8];
521 w2_t[1] = s_msg_buf[off + 9];
522 w2_t[2] = s_msg_buf[off + 10];
523 w2_t[3] = s_msg_buf[off + 11];
524 w3_t[0] = s_msg_buf[off + 12];
525 w3_t[1] = s_msg_buf[off + 13];
526 w3_t[2] = s_msg_buf[off + 14];
527 w3_t[3] = s_msg_buf[off + 15];
529 sha1_transform (w0_t, w1_t, w2_t, w3_t, ipad);
532 w0_t[0] = s_msg_buf[off + 0];
533 w0_t[1] = s_msg_buf[off + 1];
534 w0_t[2] = s_msg_buf[off + 2];
535 w0_t[3] = s_msg_buf[off + 3];
536 w1_t[0] = s_msg_buf[off + 4];
537 w1_t[1] = s_msg_buf[off + 5];
538 w1_t[2] = s_msg_buf[off + 6];
539 w1_t[3] = s_msg_buf[off + 7];
540 w2_t[0] = s_msg_buf[off + 8];
541 w2_t[1] = s_msg_buf[off + 9];
542 w2_t[2] = s_msg_buf[off + 10];
543 w2_t[3] = s_msg_buf[off + 11];
544 w3_t[0] = s_msg_buf[off + 12];
545 w3_t[1] = s_msg_buf[off + 13];
547 w3_t[3] = (64 + msg_len) * 8;
549 hmac_sha1_run (w0_t, w1_t, w2_t, w3_t, ipad, opad, digest);
551 const u32x r0 = digest[3];
552 const u32x r1 = digest[4];
553 const u32x r2 = digest[2];
554 const u32x r3 = digest[1];
556 #include VECT_COMPARE_M
560 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m05400_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 ikepsk_t *ikepsk_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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
564 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m05400_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 ikepsk_t *ikepsk_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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
568 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m05400_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 ikepsk_t *ikepsk_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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
574 const u32 lid = get_local_id (0);
580 const u32 gid = get_global_id (0);
584 wordl0[0] = pws[gid].i[ 0];
585 wordl0[1] = pws[gid].i[ 1];
586 wordl0[2] = pws[gid].i[ 2];
587 wordl0[3] = pws[gid].i[ 3];
591 wordl1[0] = pws[gid].i[ 4];
592 wordl1[1] = pws[gid].i[ 5];
593 wordl1[2] = pws[gid].i[ 6];
594 wordl1[3] = pws[gid].i[ 7];
610 const u32 pw_l_len = pws[gid].pw_len;
612 if (combs_mode == COMBINATOR_MODE_BASE_RIGHT)
614 switch_buffer_by_offset (wordl0, wordl1, wordl2, wordl3, combs_buf[0].pw_len);
621 const u32 nr_len = ikepsk_bufs[salt_pos].nr_len;
622 const u32 msg_len = ikepsk_bufs[salt_pos].msg_len;
626 salt_buf0[0] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[ 0]);
627 salt_buf0[1] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[ 1]);
628 salt_buf0[2] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[ 2]);
629 salt_buf0[3] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[ 3]);
633 salt_buf1[0] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[ 4]);
634 salt_buf1[1] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[ 5]);
635 salt_buf1[2] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[ 6]);
636 salt_buf1[3] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[ 7]);
640 salt_buf2[0] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[ 8]);
641 salt_buf2[1] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[ 9]);
642 salt_buf2[2] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[10]);
643 salt_buf2[3] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[11]);
647 salt_buf3[0] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[12]);
648 salt_buf3[1] = swap_workaround (ikepsk_bufs[salt_pos].nr_buf[13]);
652 __local u32 s_msg_buf[128];
654 const u32 lid2 = lid * 2;
656 s_msg_buf[lid2 + 0] = swap_workaround (ikepsk_bufs[salt_pos].msg_buf[lid2 + 0]);
657 s_msg_buf[lid2 + 1] = swap_workaround (ikepsk_bufs[salt_pos].msg_buf[lid2 + 1]);
659 barrier (CLK_LOCAL_MEM_FENCE);
661 if (gid >= gid_max) return;
667 const u32 search[4] =
669 digests_buf[digests_offset].digest_buf[DGST_R0],
670 digests_buf[digests_offset].digest_buf[DGST_R1],
671 digests_buf[digests_offset].digest_buf[DGST_R2],
672 digests_buf[digests_offset].digest_buf[DGST_R3]
679 for (u32 il_pos = 0; il_pos < combs_cnt; il_pos++)
681 const u32 pw_r_len = combs_buf[il_pos].pw_len;
683 const u32 pw_len = pw_l_len + pw_r_len;
687 wordr0[0] = combs_buf[il_pos].i[0];
688 wordr0[1] = combs_buf[il_pos].i[1];
689 wordr0[2] = combs_buf[il_pos].i[2];
690 wordr0[3] = combs_buf[il_pos].i[3];
694 wordr1[0] = combs_buf[il_pos].i[4];
695 wordr1[1] = combs_buf[il_pos].i[5];
696 wordr1[2] = combs_buf[il_pos].i[6];
697 wordr1[3] = combs_buf[il_pos].i[7];
713 if (combs_mode == COMBINATOR_MODE_BASE_LEFT)
715 switch_buffer_by_offset (wordr0, wordr1, wordr2, wordr3, pw_l_len);
720 w0[0] = wordl0[0] | wordr0[0];
721 w0[1] = wordl0[1] | wordr0[1];
722 w0[2] = wordl0[2] | wordr0[2];
723 w0[3] = wordl0[3] | wordr0[3];
727 w1[0] = wordl1[0] | wordr1[0];
728 w1[1] = wordl1[1] | wordr1[1];
729 w1[2] = wordl1[2] | wordr1[2];
730 w1[3] = wordl1[3] | wordr1[3];
734 w2[0] = wordl2[0] | wordr2[0];
735 w2[1] = wordl2[1] | wordr2[1];
736 w2[2] = wordl2[2] | wordr2[2];
737 w2[3] = wordl2[3] | wordr2[3];
741 w3[0] = wordl3[0] | wordr3[0];
742 w3[1] = wordl3[1] | wordr3[1];
743 w3[2] = wordl3[2] | wordr3[2];
744 w3[3] = wordl3[3] | wordr3[3];
752 w0_t[0] = swap_workaround (w0[0]);
753 w0_t[1] = swap_workaround (w0[1]);
754 w0_t[2] = swap_workaround (w0[2]);
755 w0_t[3] = swap_workaround (w0[3]);
759 w1_t[0] = swap_workaround (w1[0]);
760 w1_t[1] = swap_workaround (w1[1]);
761 w1_t[2] = swap_workaround (w1[2]);
762 w1_t[3] = swap_workaround (w1[3]);
781 hmac_sha1_pad (w0_t, w1_t, w2_t, w3_t, ipad, opad);
783 w0_t[0] = salt_buf0[0];
784 w0_t[1] = salt_buf0[1];
785 w0_t[2] = salt_buf0[2];
786 w0_t[3] = salt_buf0[3];
787 w1_t[0] = salt_buf1[0];
788 w1_t[1] = salt_buf1[1];
789 w1_t[2] = salt_buf1[2];
790 w1_t[3] = salt_buf1[3];
791 w2_t[0] = salt_buf2[0];
792 w2_t[1] = salt_buf2[1];
793 w2_t[2] = salt_buf2[2];
794 w2_t[3] = salt_buf2[3];
795 w3_t[0] = salt_buf3[0];
796 w3_t[1] = salt_buf3[1];
798 w3_t[3] = (64 + nr_len) * 8;
802 hmac_sha1_run (w0_t, w1_t, w2_t, w3_t, ipad, opad, digest);
821 hmac_sha1_pad (w0_t, w1_t, w2_t, w3_t, ipad, opad);
826 for (left = ikepsk_bufs[salt_pos].msg_len, off = 0; left >= 56; left -= 64, off += 16)
828 w0_t[0] = s_msg_buf[off + 0];
829 w0_t[1] = s_msg_buf[off + 1];
830 w0_t[2] = s_msg_buf[off + 2];
831 w0_t[3] = s_msg_buf[off + 3];
832 w1_t[0] = s_msg_buf[off + 4];
833 w1_t[1] = s_msg_buf[off + 5];
834 w1_t[2] = s_msg_buf[off + 6];
835 w1_t[3] = s_msg_buf[off + 7];
836 w2_t[0] = s_msg_buf[off + 8];
837 w2_t[1] = s_msg_buf[off + 9];
838 w2_t[2] = s_msg_buf[off + 10];
839 w2_t[3] = s_msg_buf[off + 11];
840 w3_t[0] = s_msg_buf[off + 12];
841 w3_t[1] = s_msg_buf[off + 13];
842 w3_t[2] = s_msg_buf[off + 14];
843 w3_t[3] = s_msg_buf[off + 15];
845 sha1_transform (w0_t, w1_t, w2_t, w3_t, ipad);
848 w0_t[0] = s_msg_buf[off + 0];
849 w0_t[1] = s_msg_buf[off + 1];
850 w0_t[2] = s_msg_buf[off + 2];
851 w0_t[3] = s_msg_buf[off + 3];
852 w1_t[0] = s_msg_buf[off + 4];
853 w1_t[1] = s_msg_buf[off + 5];
854 w1_t[2] = s_msg_buf[off + 6];
855 w1_t[3] = s_msg_buf[off + 7];
856 w2_t[0] = s_msg_buf[off + 8];
857 w2_t[1] = s_msg_buf[off + 9];
858 w2_t[2] = s_msg_buf[off + 10];
859 w2_t[3] = s_msg_buf[off + 11];
860 w3_t[0] = s_msg_buf[off + 12];
861 w3_t[1] = s_msg_buf[off + 13];
863 w3_t[3] = (64 + msg_len) * 8;
865 hmac_sha1_run (w0_t, w1_t, w2_t, w3_t, ipad, opad, digest);
867 const u32x r0 = digest[3];
868 const u32x r1 = digest[4];
869 const u32x r2 = digest[2];
870 const u32x r3 = digest[1];
872 #include VECT_COMPARE_S
876 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m05400_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 ikepsk_t *ikepsk_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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
880 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m05400_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 ikepsk_t *ikepsk_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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)