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"
31 #include "include/rp_gpu.h"
35 #define VECT_COMPARE_S "check_single_vect1_comp4.c"
36 #define VECT_COMPARE_M "check_multi_vect1_comp4.c"
40 #define VECT_COMPARE_S "check_single_vect4_comp4.c"
41 #define VECT_COMPARE_M "check_multi_vect4_comp4.c"
44 __constant u32 padding[8] =
64 static void swap (__local RC4_KEY *rc4_key, const u8 i, const u8 j)
69 rc4_key->S[i] = rc4_key->S[j];
73 static void rc4_init_16 (__local RC4_KEY *rc4_key, const u32 data[4])
78 __local u32 *ptr = (__local u32 *) rc4_key->S;
81 for (u32 i = 0; i < 64; i++)
86 const u32 d0 = data[0] >> 0;
87 const u32 d1 = data[0] >> 8;
88 const u32 d2 = data[0] >> 16;
89 const u32 d3 = data[0] >> 24;
90 const u32 d4 = data[1] >> 0;
95 for (u32 i = 0; i < 255; i += 5)
97 j += rc4_key->S[i + 0] + d0; swap (rc4_key, i + 0, j);
98 j += rc4_key->S[i + 1] + d1; swap (rc4_key, i + 1, j);
99 j += rc4_key->S[i + 2] + d2; swap (rc4_key, i + 2, j);
100 j += rc4_key->S[i + 3] + d3; swap (rc4_key, i + 3, j);
101 j += rc4_key->S[i + 4] + d4; swap (rc4_key, i + 4, j);
104 j += rc4_key->S[255] + d0; swap (rc4_key, 255, j);
107 static u8 rc4_next_16 (__local RC4_KEY *rc4_key, u8 i, u8 j, __constant u32 in[4], u32 out[4])
110 for (u32 k = 0; k < 4; k++)
119 swap (rc4_key, i, j);
121 idx = rc4_key->S[i] + rc4_key->S[j];
123 xor4 |= rc4_key->S[idx] << 0;
128 swap (rc4_key, i, j);
130 idx = rc4_key->S[i] + rc4_key->S[j];
132 xor4 |= rc4_key->S[idx] << 8;
137 swap (rc4_key, i, j);
139 idx = rc4_key->S[i] + rc4_key->S[j];
141 xor4 |= rc4_key->S[idx] << 16;
146 swap (rc4_key, i, j);
148 idx = rc4_key->S[i] + rc4_key->S[j];
150 xor4 |= rc4_key->S[idx] << 24;
152 out[k] = in[k] ^ xor4;
158 static void md5_transform (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u32x digest[4])
182 MD5_STEP (MD5_Fo, a, b, c, d, w0_t, MD5C00, MD5S00);
183 MD5_STEP (MD5_Fo, d, a, b, c, w1_t, MD5C01, MD5S01);
184 MD5_STEP (MD5_Fo, c, d, a, b, w2_t, MD5C02, MD5S02);
185 MD5_STEP (MD5_Fo, b, c, d, a, w3_t, MD5C03, MD5S03);
186 MD5_STEP (MD5_Fo, a, b, c, d, w4_t, MD5C04, MD5S00);
187 MD5_STEP (MD5_Fo, d, a, b, c, w5_t, MD5C05, MD5S01);
188 MD5_STEP (MD5_Fo, c, d, a, b, w6_t, MD5C06, MD5S02);
189 MD5_STEP (MD5_Fo, b, c, d, a, w7_t, MD5C07, MD5S03);
190 MD5_STEP (MD5_Fo, a, b, c, d, w8_t, MD5C08, MD5S00);
191 MD5_STEP (MD5_Fo, d, a, b, c, w9_t, MD5C09, MD5S01);
192 MD5_STEP (MD5_Fo, c, d, a, b, wa_t, MD5C0a, MD5S02);
193 MD5_STEP (MD5_Fo, b, c, d, a, wb_t, MD5C0b, MD5S03);
194 MD5_STEP (MD5_Fo, a, b, c, d, wc_t, MD5C0c, MD5S00);
195 MD5_STEP (MD5_Fo, d, a, b, c, wd_t, MD5C0d, MD5S01);
196 MD5_STEP (MD5_Fo, c, d, a, b, we_t, MD5C0e, MD5S02);
197 MD5_STEP (MD5_Fo, b, c, d, a, wf_t, MD5C0f, MD5S03);
199 MD5_STEP (MD5_Go, a, b, c, d, w1_t, MD5C10, MD5S10);
200 MD5_STEP (MD5_Go, d, a, b, c, w6_t, MD5C11, MD5S11);
201 MD5_STEP (MD5_Go, c, d, a, b, wb_t, MD5C12, MD5S12);
202 MD5_STEP (MD5_Go, b, c, d, a, w0_t, MD5C13, MD5S13);
203 MD5_STEP (MD5_Go, a, b, c, d, w5_t, MD5C14, MD5S10);
204 MD5_STEP (MD5_Go, d, a, b, c, wa_t, MD5C15, MD5S11);
205 MD5_STEP (MD5_Go, c, d, a, b, wf_t, MD5C16, MD5S12);
206 MD5_STEP (MD5_Go, b, c, d, a, w4_t, MD5C17, MD5S13);
207 MD5_STEP (MD5_Go, a, b, c, d, w9_t, MD5C18, MD5S10);
208 MD5_STEP (MD5_Go, d, a, b, c, we_t, MD5C19, MD5S11);
209 MD5_STEP (MD5_Go, c, d, a, b, w3_t, MD5C1a, MD5S12);
210 MD5_STEP (MD5_Go, b, c, d, a, w8_t, MD5C1b, MD5S13);
211 MD5_STEP (MD5_Go, a, b, c, d, wd_t, MD5C1c, MD5S10);
212 MD5_STEP (MD5_Go, d, a, b, c, w2_t, MD5C1d, MD5S11);
213 MD5_STEP (MD5_Go, c, d, a, b, w7_t, MD5C1e, MD5S12);
214 MD5_STEP (MD5_Go, b, c, d, a, wc_t, MD5C1f, MD5S13);
216 MD5_STEP (MD5_H , a, b, c, d, w5_t, MD5C20, MD5S20);
217 MD5_STEP (MD5_H , d, a, b, c, w8_t, MD5C21, MD5S21);
218 MD5_STEP (MD5_H , c, d, a, b, wb_t, MD5C22, MD5S22);
219 MD5_STEP (MD5_H , b, c, d, a, we_t, MD5C23, MD5S23);
220 MD5_STEP (MD5_H , a, b, c, d, w1_t, MD5C24, MD5S20);
221 MD5_STEP (MD5_H , d, a, b, c, w4_t, MD5C25, MD5S21);
222 MD5_STEP (MD5_H , c, d, a, b, w7_t, MD5C26, MD5S22);
223 MD5_STEP (MD5_H , b, c, d, a, wa_t, MD5C27, MD5S23);
224 MD5_STEP (MD5_H , a, b, c, d, wd_t, MD5C28, MD5S20);
225 MD5_STEP (MD5_H , d, a, b, c, w0_t, MD5C29, MD5S21);
226 MD5_STEP (MD5_H , c, d, a, b, w3_t, MD5C2a, MD5S22);
227 MD5_STEP (MD5_H , b, c, d, a, w6_t, MD5C2b, MD5S23);
228 MD5_STEP (MD5_H , a, b, c, d, w9_t, MD5C2c, MD5S20);
229 MD5_STEP (MD5_H , d, a, b, c, wc_t, MD5C2d, MD5S21);
230 MD5_STEP (MD5_H , c, d, a, b, wf_t, MD5C2e, MD5S22);
231 MD5_STEP (MD5_H , b, c, d, a, w2_t, MD5C2f, MD5S23);
233 MD5_STEP (MD5_I , a, b, c, d, w0_t, MD5C30, MD5S30);
234 MD5_STEP (MD5_I , d, a, b, c, w7_t, MD5C31, MD5S31);
235 MD5_STEP (MD5_I , c, d, a, b, we_t, MD5C32, MD5S32);
236 MD5_STEP (MD5_I , b, c, d, a, w5_t, MD5C33, MD5S33);
237 MD5_STEP (MD5_I , a, b, c, d, wc_t, MD5C34, MD5S30);
238 MD5_STEP (MD5_I , d, a, b, c, w3_t, MD5C35, MD5S31);
239 MD5_STEP (MD5_I , c, d, a, b, wa_t, MD5C36, MD5S32);
240 MD5_STEP (MD5_I , b, c, d, a, w1_t, MD5C37, MD5S33);
241 MD5_STEP (MD5_I , a, b, c, d, w8_t, MD5C38, MD5S30);
242 MD5_STEP (MD5_I , d, a, b, c, wf_t, MD5C39, MD5S31);
243 MD5_STEP (MD5_I , c, d, a, b, w6_t, MD5C3a, MD5S32);
244 MD5_STEP (MD5_I , b, c, d, a, wd_t, MD5C3b, MD5S33);
245 MD5_STEP (MD5_I , a, b, c, d, w4_t, MD5C3c, MD5S30);
246 MD5_STEP (MD5_I , d, a, b, c, wb_t, MD5C3d, MD5S31);
247 MD5_STEP (MD5_I , c, d, a, b, w2_t, MD5C3e, MD5S32);
248 MD5_STEP (MD5_I , b, c, d, a, w9_t, MD5C3f, MD5S33);
256 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10400_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 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)
262 const u32 lid = get_local_id (0);
268 const u32 gid = get_global_id (0);
270 if (gid >= gid_max) return;
274 pw_buf0[0] = pws[gid].i[ 0];
275 pw_buf0[1] = pws[gid].i[ 1];
276 pw_buf0[2] = pws[gid].i[ 2];
277 pw_buf0[3] = pws[gid].i[ 3];
281 pw_buf1[0] = pws[gid].i[ 4];
282 pw_buf1[1] = pws[gid].i[ 5];
283 pw_buf1[2] = pws[gid].i[ 6];
284 pw_buf1[3] = pws[gid].i[ 7];
286 const u32 pw_len = pws[gid].pw_len;
292 __local RC4_KEY rc4_keys[64];
294 __local RC4_KEY *rc4_key = &rc4_keys[lid];
302 o_buf[0] = pdf_bufs[salt_pos].o_buf[0];
303 o_buf[1] = pdf_bufs[salt_pos].o_buf[1];
304 o_buf[2] = pdf_bufs[salt_pos].o_buf[2];
305 o_buf[3] = pdf_bufs[salt_pos].o_buf[3];
306 o_buf[4] = pdf_bufs[salt_pos].o_buf[4];
307 o_buf[5] = pdf_bufs[salt_pos].o_buf[5];
308 o_buf[6] = pdf_bufs[salt_pos].o_buf[6];
309 o_buf[7] = pdf_bufs[salt_pos].o_buf[7];
311 u32 P = pdf_bufs[salt_pos].P;
315 id_buf[0] = pdf_bufs[salt_pos].id_buf[0];
316 id_buf[1] = pdf_bufs[salt_pos].id_buf[1];
317 id_buf[2] = pdf_bufs[salt_pos].id_buf[2];
318 id_buf[3] = pdf_bufs[salt_pos].id_buf[3];
324 for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++)
354 const u32 out_len = apply_rules (rules_buf[il_pos].cmds, w0, w1, pw_len);
361 // max length supported by pdf11 is 32
363 w0_t[0] = padding[0];
364 w0_t[1] = padding[1];
365 w0_t[2] = padding[2];
366 w0_t[3] = padding[3];
367 w1_t[0] = padding[4];
368 w1_t[1] = padding[5];
369 w1_t[2] = padding[6];
370 w1_t[3] = padding[7];
380 switch_buffer_by_offset (w0_t, w1_t, w2_t, w3_t, pw_len);
383 // truncate at 32 is wanted, not a bug!
410 md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
429 md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
436 key[1] = digest[1] & 0xff;
440 rc4_init_16 (rc4_key, key);
444 rc4_next_16 (rc4_key, 0, 0, padding, out);
446 const u32x r0 = out[0];
447 const u32x r1 = out[1];
448 const u32x r2 = out[2];
449 const u32x r3 = out[3];
451 #include VECT_COMPARE_M
455 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10400_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 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)
459 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10400_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 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)
463 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10400_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 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)
469 const u32 lid = get_local_id (0);
475 const u32 gid = get_global_id (0);
477 if (gid >= gid_max) return;
481 pw_buf0[0] = pws[gid].i[ 0];
482 pw_buf0[1] = pws[gid].i[ 1];
483 pw_buf0[2] = pws[gid].i[ 2];
484 pw_buf0[3] = pws[gid].i[ 3];
488 pw_buf1[0] = pws[gid].i[ 4];
489 pw_buf1[1] = pws[gid].i[ 5];
490 pw_buf1[2] = pws[gid].i[ 6];
491 pw_buf1[3] = pws[gid].i[ 7];
493 const u32 pw_len = pws[gid].pw_len;
499 __local RC4_KEY rc4_keys[64];
501 __local RC4_KEY *rc4_key = &rc4_keys[lid];
507 const u32 search[4] =
509 digests_buf[digests_offset].digest_buf[DGST_R0],
510 digests_buf[digests_offset].digest_buf[DGST_R1],
511 digests_buf[digests_offset].digest_buf[DGST_R2],
512 digests_buf[digests_offset].digest_buf[DGST_R3]
521 o_buf[0] = pdf_bufs[salt_pos].o_buf[0];
522 o_buf[1] = pdf_bufs[salt_pos].o_buf[1];
523 o_buf[2] = pdf_bufs[salt_pos].o_buf[2];
524 o_buf[3] = pdf_bufs[salt_pos].o_buf[3];
525 o_buf[4] = pdf_bufs[salt_pos].o_buf[4];
526 o_buf[5] = pdf_bufs[salt_pos].o_buf[5];
527 o_buf[6] = pdf_bufs[salt_pos].o_buf[6];
528 o_buf[7] = pdf_bufs[salt_pos].o_buf[7];
530 u32 P = pdf_bufs[salt_pos].P;
534 id_buf[0] = pdf_bufs[salt_pos].id_buf[0];
535 id_buf[1] = pdf_bufs[salt_pos].id_buf[1];
536 id_buf[2] = pdf_bufs[salt_pos].id_buf[2];
537 id_buf[3] = pdf_bufs[salt_pos].id_buf[3];
543 for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++)
573 const u32 out_len = apply_rules (rules_buf[il_pos].cmds, w0, w1, pw_len);
580 // max length supported by pdf11 is 32
582 w0_t[0] = padding[0];
583 w0_t[1] = padding[1];
584 w0_t[2] = padding[2];
585 w0_t[3] = padding[3];
586 w1_t[0] = padding[4];
587 w1_t[1] = padding[5];
588 w1_t[2] = padding[6];
589 w1_t[3] = padding[7];
599 switch_buffer_by_offset (w0_t, w1_t, w2_t, w3_t, pw_len);
602 // truncate at 32 is wanted, not a bug!
629 md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
648 md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
655 key[1] = digest[1] & 0xff;
659 rc4_init_16 (rc4_key, key);
663 rc4_next_16 (rc4_key, 0, 0, padding, out);
665 const u32x r0 = out[0];
666 const u32x r1 = out[1];
667 const u32x r2 = out[2];
668 const u32x r3 = out[3];
670 #include VECT_COMPARE_S
674 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10400_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 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)
678 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10400_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 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)