2 * Author......: Jens Steube <jens.steube@gmail.com>
8 #include "include/constants.h"
9 #include "include/kernel_vendor.h"
16 #include "include/kernel_functions.c"
17 #include "types_ocl.c"
21 #define COMPARE_M "check_multi_vect1_comp4.c"
25 #define COMPARE_M "check_multi_vect4_comp4.c"
28 __constant u32 padding[8] =
48 static void swap (__local RC4_KEY *rc4_key, const u8 i, const u8 j)
53 rc4_key->S[i] = rc4_key->S[j];
57 static void rc4_init_16 (__local RC4_KEY *rc4_key, const u32 data[4])
62 __local u32 *ptr = (__local u32 *) rc4_key->S;
65 for (u32 i = 0; i < 64; i++)
73 for (u32 i = 0; i < 16; i++)
81 j += rc4_key->S[idx] + (v >> 0); swap (rc4_key, idx, j); idx++;
82 j += rc4_key->S[idx] + (v >> 8); swap (rc4_key, idx, j); idx++;
83 j += rc4_key->S[idx] + (v >> 16); swap (rc4_key, idx, j); idx++;
84 j += rc4_key->S[idx] + (v >> 24); swap (rc4_key, idx, j); idx++;
88 j += rc4_key->S[idx] + (v >> 0); swap (rc4_key, idx, j); idx++;
89 j += rc4_key->S[idx] + (v >> 8); swap (rc4_key, idx, j); idx++;
90 j += rc4_key->S[idx] + (v >> 16); swap (rc4_key, idx, j); idx++;
91 j += rc4_key->S[idx] + (v >> 24); swap (rc4_key, idx, j); idx++;
95 j += rc4_key->S[idx] + (v >> 0); swap (rc4_key, idx, j); idx++;
96 j += rc4_key->S[idx] + (v >> 8); swap (rc4_key, idx, j); idx++;
97 j += rc4_key->S[idx] + (v >> 16); swap (rc4_key, idx, j); idx++;
98 j += rc4_key->S[idx] + (v >> 24); swap (rc4_key, idx, j); idx++;
102 j += rc4_key->S[idx] + (v >> 0); swap (rc4_key, idx, j); idx++;
103 j += rc4_key->S[idx] + (v >> 8); swap (rc4_key, idx, j); idx++;
104 j += rc4_key->S[idx] + (v >> 16); swap (rc4_key, idx, j); idx++;
105 j += rc4_key->S[idx] + (v >> 24); swap (rc4_key, idx, j); idx++;
109 static u8 rc4_next_16 (__local RC4_KEY *rc4_key, u8 i, u8 j, const u32 in[4], u32 out[4])
112 for (u32 k = 0; k < 4; k++)
121 swap (rc4_key, i, j);
123 idx = rc4_key->S[i] + rc4_key->S[j];
125 xor4 |= rc4_key->S[idx] << 0;
130 swap (rc4_key, i, j);
132 idx = rc4_key->S[i] + rc4_key->S[j];
134 xor4 |= rc4_key->S[idx] << 8;
139 swap (rc4_key, i, j);
141 idx = rc4_key->S[i] + rc4_key->S[j];
143 xor4 |= rc4_key->S[idx] << 16;
148 swap (rc4_key, i, j);
150 idx = rc4_key->S[i] + rc4_key->S[j];
152 xor4 |= rc4_key->S[idx] << 24;
154 out[k] = in[k] ^ xor4;
160 static void md5_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[4])
186 MD5_STEP (MD5_Fo, a, b, c, d, w0_t, MD5C00, MD5S00);
187 MD5_STEP (MD5_Fo, d, a, b, c, w1_t, MD5C01, MD5S01);
188 MD5_STEP (MD5_Fo, c, d, a, b, w2_t, MD5C02, MD5S02);
189 MD5_STEP (MD5_Fo, b, c, d, a, w3_t, MD5C03, MD5S03);
190 MD5_STEP (MD5_Fo, a, b, c, d, w4_t, MD5C04, MD5S00);
191 MD5_STEP (MD5_Fo, d, a, b, c, w5_t, MD5C05, MD5S01);
192 MD5_STEP (MD5_Fo, c, d, a, b, w6_t, MD5C06, MD5S02);
193 MD5_STEP (MD5_Fo, b, c, d, a, w7_t, MD5C07, MD5S03);
194 MD5_STEP (MD5_Fo, a, b, c, d, w8_t, MD5C08, MD5S00);
195 MD5_STEP (MD5_Fo, d, a, b, c, w9_t, MD5C09, MD5S01);
196 MD5_STEP (MD5_Fo, c, d, a, b, wa_t, MD5C0a, MD5S02);
197 MD5_STEP (MD5_Fo, b, c, d, a, wb_t, MD5C0b, MD5S03);
198 MD5_STEP (MD5_Fo, a, b, c, d, wc_t, MD5C0c, MD5S00);
199 MD5_STEP (MD5_Fo, d, a, b, c, wd_t, MD5C0d, MD5S01);
200 MD5_STEP (MD5_Fo, c, d, a, b, we_t, MD5C0e, MD5S02);
201 MD5_STEP (MD5_Fo, b, c, d, a, wf_t, MD5C0f, MD5S03);
203 MD5_STEP (MD5_Go, a, b, c, d, w1_t, MD5C10, MD5S10);
204 MD5_STEP (MD5_Go, d, a, b, c, w6_t, MD5C11, MD5S11);
205 MD5_STEP (MD5_Go, c, d, a, b, wb_t, MD5C12, MD5S12);
206 MD5_STEP (MD5_Go, b, c, d, a, w0_t, MD5C13, MD5S13);
207 MD5_STEP (MD5_Go, a, b, c, d, w5_t, MD5C14, MD5S10);
208 MD5_STEP (MD5_Go, d, a, b, c, wa_t, MD5C15, MD5S11);
209 MD5_STEP (MD5_Go, c, d, a, b, wf_t, MD5C16, MD5S12);
210 MD5_STEP (MD5_Go, b, c, d, a, w4_t, MD5C17, MD5S13);
211 MD5_STEP (MD5_Go, a, b, c, d, w9_t, MD5C18, MD5S10);
212 MD5_STEP (MD5_Go, d, a, b, c, we_t, MD5C19, MD5S11);
213 MD5_STEP (MD5_Go, c, d, a, b, w3_t, MD5C1a, MD5S12);
214 MD5_STEP (MD5_Go, b, c, d, a, w8_t, MD5C1b, MD5S13);
215 MD5_STEP (MD5_Go, a, b, c, d, wd_t, MD5C1c, MD5S10);
216 MD5_STEP (MD5_Go, d, a, b, c, w2_t, MD5C1d, MD5S11);
217 MD5_STEP (MD5_Go, c, d, a, b, w7_t, MD5C1e, MD5S12);
218 MD5_STEP (MD5_Go, b, c, d, a, wc_t, MD5C1f, MD5S13);
220 MD5_STEP (MD5_H1, a, b, c, d, w5_t, MD5C20, MD5S20);
221 MD5_STEP (MD5_H2, d, a, b, c, w8_t, MD5C21, MD5S21);
222 MD5_STEP (MD5_H1, c, d, a, b, wb_t, MD5C22, MD5S22);
223 MD5_STEP (MD5_H2, b, c, d, a, we_t, MD5C23, MD5S23);
224 MD5_STEP (MD5_H1, a, b, c, d, w1_t, MD5C24, MD5S20);
225 MD5_STEP (MD5_H2, d, a, b, c, w4_t, MD5C25, MD5S21);
226 MD5_STEP (MD5_H1, c, d, a, b, w7_t, MD5C26, MD5S22);
227 MD5_STEP (MD5_H2, b, c, d, a, wa_t, MD5C27, MD5S23);
228 MD5_STEP (MD5_H1, a, b, c, d, wd_t, MD5C28, MD5S20);
229 MD5_STEP (MD5_H2, d, a, b, c, w0_t, MD5C29, MD5S21);
230 MD5_STEP (MD5_H1, c, d, a, b, w3_t, MD5C2a, MD5S22);
231 MD5_STEP (MD5_H2, b, c, d, a, w6_t, MD5C2b, MD5S23);
232 MD5_STEP (MD5_H1, a, b, c, d, w9_t, MD5C2c, MD5S20);
233 MD5_STEP (MD5_H2, d, a, b, c, wc_t, MD5C2d, MD5S21);
234 MD5_STEP (MD5_H1, c, d, a, b, wf_t, MD5C2e, MD5S22);
235 MD5_STEP (MD5_H2, b, c, d, a, w2_t, MD5C2f, MD5S23);
237 MD5_STEP (MD5_I , a, b, c, d, w0_t, MD5C30, MD5S30);
238 MD5_STEP (MD5_I , d, a, b, c, w7_t, MD5C31, MD5S31);
239 MD5_STEP (MD5_I , c, d, a, b, we_t, MD5C32, MD5S32);
240 MD5_STEP (MD5_I , b, c, d, a, w5_t, MD5C33, MD5S33);
241 MD5_STEP (MD5_I , a, b, c, d, wc_t, MD5C34, MD5S30);
242 MD5_STEP (MD5_I , d, a, b, c, w3_t, MD5C35, MD5S31);
243 MD5_STEP (MD5_I , c, d, a, b, wa_t, MD5C36, MD5S32);
244 MD5_STEP (MD5_I , b, c, d, a, w1_t, MD5C37, MD5S33);
245 MD5_STEP (MD5_I , a, b, c, d, w8_t, MD5C38, MD5S30);
246 MD5_STEP (MD5_I , d, a, b, c, wf_t, MD5C39, MD5S31);
247 MD5_STEP (MD5_I , c, d, a, b, w6_t, MD5C3a, MD5S32);
248 MD5_STEP (MD5_I , b, c, d, a, wd_t, MD5C3b, MD5S33);
249 MD5_STEP (MD5_I , a, b, c, d, w4_t, MD5C3c, MD5S30);
250 MD5_STEP (MD5_I , d, a, b, c, wb_t, MD5C3d, MD5S31);
251 MD5_STEP (MD5_I , c, d, a, b, w2_t, MD5C3e, MD5S32);
252 MD5_STEP (MD5_I , b, c, d, a, w9_t, MD5C3f, MD5S33);
260 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10500_init (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global pdf14_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 pdf_t *pdf_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)
266 const u32 gid = get_global_id (0);
267 const u32 lid = get_local_id (0);
269 if (gid >= gid_max) return;
273 w0[0] = pws[gid].i[ 0];
274 w0[1] = pws[gid].i[ 1];
275 w0[2] = pws[gid].i[ 2];
276 w0[3] = pws[gid].i[ 3];
280 w1[0] = pws[gid].i[ 4];
281 w1[1] = pws[gid].i[ 5];
282 w1[2] = pws[gid].i[ 6];
283 w1[3] = pws[gid].i[ 7];
287 w2[0] = pws[gid].i[ 8];
288 w2[1] = pws[gid].i[ 9];
292 const u32 pw_len = pws[gid].pw_len;
298 __local RC4_KEY rc4_keys[64];
300 __local RC4_KEY *rc4_key = &rc4_keys[lid];
308 o_buf[0] = pdf_bufs[salt_pos].o_buf[0];
309 o_buf[1] = pdf_bufs[salt_pos].o_buf[1];
310 o_buf[2] = pdf_bufs[salt_pos].o_buf[2];
311 o_buf[3] = pdf_bufs[salt_pos].o_buf[3];
312 o_buf[4] = pdf_bufs[salt_pos].o_buf[4];
313 o_buf[5] = pdf_bufs[salt_pos].o_buf[5];
314 o_buf[6] = pdf_bufs[salt_pos].o_buf[6];
315 o_buf[7] = pdf_bufs[salt_pos].o_buf[7];
317 u32 P = pdf_bufs[salt_pos].P;
321 id_buf[ 0] = pdf_bufs[salt_pos].id_buf[0];
322 id_buf[ 1] = pdf_bufs[salt_pos].id_buf[1];
323 id_buf[ 2] = pdf_bufs[salt_pos].id_buf[2];
324 id_buf[ 3] = pdf_bufs[salt_pos].id_buf[3];
326 id_buf[ 4] = pdf_bufs[salt_pos].id_buf[4];
327 id_buf[ 5] = pdf_bufs[salt_pos].id_buf[5];
328 id_buf[ 6] = pdf_bufs[salt_pos].id_buf[6];
329 id_buf[ 7] = pdf_bufs[salt_pos].id_buf[7];
336 u32 id_len = pdf_bufs[salt_pos].id_len;
337 u32 id_len4 = id_len / 4;
341 rc4data[0] = pdf_bufs[salt_pos].rc4data[0];
342 rc4data[1] = pdf_bufs[salt_pos].rc4data[1];
344 u32 final_length = 68 + id_len;
349 if (pdf_bufs[salt_pos].enc_md != 1)
357 id_buf[id_len4 + 0] = w11;
358 id_buf[id_len4 + 1] = w12;
369 // max length supported by pdf11 is 32
371 w0_t[0] = padding[0];
372 w0_t[1] = padding[1];
373 w0_t[2] = padding[2];
374 w0_t[3] = padding[3];
375 w1_t[0] = padding[4];
376 w1_t[1] = padding[5];
377 w1_t[2] = padding[6];
378 w1_t[3] = padding[7];
388 switch_buffer_by_offset (w0_t, w1_t, w2_t, w3_t, pw_len);
391 // truncate at 32 is wanted, not a bug!
418 md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
421 w0_t[1] = id_buf[ 0];
422 w0_t[2] = id_buf[ 1];
423 w0_t[3] = id_buf[ 2];
424 w1_t[0] = id_buf[ 3];
425 w1_t[1] = id_buf[ 4];
426 w1_t[2] = id_buf[ 5];
427 w1_t[3] = id_buf[ 6];
428 w2_t[0] = id_buf[ 7];
429 w2_t[1] = id_buf[ 8];
430 w2_t[2] = id_buf[ 9];
431 w2_t[3] = id_buf[10];
432 w3_t[0] = id_buf[11];
434 w3_t[2] = final_length * 8;
437 md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
439 tmps[gid].digest[0] = digest[0];
440 tmps[gid].digest[1] = digest[1];
441 tmps[gid].digest[2] = digest[2];
442 tmps[gid].digest[3] = digest[3];
444 tmps[gid].out[0] = rc4data[0];
445 tmps[gid].out[1] = rc4data[1];
446 tmps[gid].out[2] = 0;
447 tmps[gid].out[3] = 0;
450 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10500_loop (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global pdf14_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 pdf_t *pdf_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)
456 const u32 gid = get_global_id (0);
457 const u32 lid = get_local_id (0);
459 if (gid >= gid_max) return;
465 __local RC4_KEY rc4_keys[64];
467 __local RC4_KEY *rc4_key = &rc4_keys[lid];
475 digest[0] = tmps[gid].digest[0];
476 digest[1] = tmps[gid].digest[1];
477 digest[2] = tmps[gid].digest[2];
478 digest[3] = tmps[gid].digest[3];
482 out[0] = tmps[gid].out[0];
483 out[1] = tmps[gid].out[1];
484 out[2] = tmps[gid].out[2];
485 out[3] = tmps[gid].out[3];
487 for (u32 i = 0, j = loop_pos; i < loop_cnt; i++, j++)
518 md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
522 const u32 x = j - 50;
524 const u32 xv = x << 0
531 tmp[0] = digest[0] ^ xv;
532 tmp[1] = digest[1] ^ xv;
533 tmp[2] = digest[2] ^ xv;
534 tmp[3] = digest[3] ^ xv;
536 rc4_init_16 (rc4_key, tmp);
538 rc4_next_16 (rc4_key, 0, 0, out, out);
542 tmps[gid].digest[0] = digest[0];
543 tmps[gid].digest[1] = digest[1];
544 tmps[gid].digest[2] = digest[2];
545 tmps[gid].digest[3] = digest[3];
547 tmps[gid].out[0] = out[0];
548 tmps[gid].out[1] = out[1];
549 tmps[gid].out[2] = out[2];
550 tmps[gid].out[3] = out[3];
553 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10500_comp (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global pdf14_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 pdf_t *pdf_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)
559 const u32 gid = get_global_id (0);
561 if (gid >= gid_max) return;
563 const u32 lid = get_local_id (0);
569 const u32 r0 = tmps[gid].out[0];
570 const u32 r1 = tmps[gid].out[1];