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_vect2_comp4.c"
41 #define VECT_COMPARE_M "check_multi_vect2_comp4.c"
45 #define VECT_COMPARE_S "check_single_vect4_comp4.c"
46 #define VECT_COMPARE_M "check_multi_vect4_comp4.c"
49 __constant u32 padding[8] =
61 static void md5_transform (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u32x digest[4])
85 MD5_STEP (MD5_Fo, a, b, c, d, w0_t, MD5C00, MD5S00);
86 MD5_STEP (MD5_Fo, d, a, b, c, w1_t, MD5C01, MD5S01);
87 MD5_STEP (MD5_Fo, c, d, a, b, w2_t, MD5C02, MD5S02);
88 MD5_STEP (MD5_Fo, b, c, d, a, w3_t, MD5C03, MD5S03);
89 MD5_STEP (MD5_Fo, a, b, c, d, w4_t, MD5C04, MD5S00);
90 MD5_STEP (MD5_Fo, d, a, b, c, w5_t, MD5C05, MD5S01);
91 MD5_STEP (MD5_Fo, c, d, a, b, w6_t, MD5C06, MD5S02);
92 MD5_STEP (MD5_Fo, b, c, d, a, w7_t, MD5C07, MD5S03);
93 MD5_STEP (MD5_Fo, a, b, c, d, w8_t, MD5C08, MD5S00);
94 MD5_STEP (MD5_Fo, d, a, b, c, w9_t, MD5C09, MD5S01);
95 MD5_STEP (MD5_Fo, c, d, a, b, wa_t, MD5C0a, MD5S02);
96 MD5_STEP (MD5_Fo, b, c, d, a, wb_t, MD5C0b, MD5S03);
97 MD5_STEP (MD5_Fo, a, b, c, d, wc_t, MD5C0c, MD5S00);
98 MD5_STEP (MD5_Fo, d, a, b, c, wd_t, MD5C0d, MD5S01);
99 MD5_STEP (MD5_Fo, c, d, a, b, we_t, MD5C0e, MD5S02);
100 MD5_STEP (MD5_Fo, b, c, d, a, wf_t, MD5C0f, MD5S03);
102 MD5_STEP (MD5_Go, a, b, c, d, w1_t, MD5C10, MD5S10);
103 MD5_STEP (MD5_Go, d, a, b, c, w6_t, MD5C11, MD5S11);
104 MD5_STEP (MD5_Go, c, d, a, b, wb_t, MD5C12, MD5S12);
105 MD5_STEP (MD5_Go, b, c, d, a, w0_t, MD5C13, MD5S13);
106 MD5_STEP (MD5_Go, a, b, c, d, w5_t, MD5C14, MD5S10);
107 MD5_STEP (MD5_Go, d, a, b, c, wa_t, MD5C15, MD5S11);
108 MD5_STEP (MD5_Go, c, d, a, b, wf_t, MD5C16, MD5S12);
109 MD5_STEP (MD5_Go, b, c, d, a, w4_t, MD5C17, MD5S13);
110 MD5_STEP (MD5_Go, a, b, c, d, w9_t, MD5C18, MD5S10);
111 MD5_STEP (MD5_Go, d, a, b, c, we_t, MD5C19, MD5S11);
112 MD5_STEP (MD5_Go, c, d, a, b, w3_t, MD5C1a, MD5S12);
113 MD5_STEP (MD5_Go, b, c, d, a, w8_t, MD5C1b, MD5S13);
114 MD5_STEP (MD5_Go, a, b, c, d, wd_t, MD5C1c, MD5S10);
115 MD5_STEP (MD5_Go, d, a, b, c, w2_t, MD5C1d, MD5S11);
116 MD5_STEP (MD5_Go, c, d, a, b, w7_t, MD5C1e, MD5S12);
117 MD5_STEP (MD5_Go, b, c, d, a, wc_t, MD5C1f, MD5S13);
119 MD5_STEP (MD5_H , a, b, c, d, w5_t, MD5C20, MD5S20);
120 MD5_STEP (MD5_H , d, a, b, c, w8_t, MD5C21, MD5S21);
121 MD5_STEP (MD5_H , c, d, a, b, wb_t, MD5C22, MD5S22);
122 MD5_STEP (MD5_H , b, c, d, a, we_t, MD5C23, MD5S23);
123 MD5_STEP (MD5_H , a, b, c, d, w1_t, MD5C24, MD5S20);
124 MD5_STEP (MD5_H , d, a, b, c, w4_t, MD5C25, MD5S21);
125 MD5_STEP (MD5_H , c, d, a, b, w7_t, MD5C26, MD5S22);
126 MD5_STEP (MD5_H , b, c, d, a, wa_t, MD5C27, MD5S23);
127 MD5_STEP (MD5_H , a, b, c, d, wd_t, MD5C28, MD5S20);
128 MD5_STEP (MD5_H , d, a, b, c, w0_t, MD5C29, MD5S21);
129 MD5_STEP (MD5_H , c, d, a, b, w3_t, MD5C2a, MD5S22);
130 MD5_STEP (MD5_H , b, c, d, a, w6_t, MD5C2b, MD5S23);
131 MD5_STEP (MD5_H , a, b, c, d, w9_t, MD5C2c, MD5S20);
132 MD5_STEP (MD5_H , d, a, b, c, wc_t, MD5C2d, MD5S21);
133 MD5_STEP (MD5_H , c, d, a, b, wf_t, MD5C2e, MD5S22);
134 MD5_STEP (MD5_H , b, c, d, a, w2_t, MD5C2f, MD5S23);
136 MD5_STEP (MD5_I , a, b, c, d, w0_t, MD5C30, MD5S30);
137 MD5_STEP (MD5_I , d, a, b, c, w7_t, MD5C31, MD5S31);
138 MD5_STEP (MD5_I , c, d, a, b, we_t, MD5C32, MD5S32);
139 MD5_STEP (MD5_I , b, c, d, a, w5_t, MD5C33, MD5S33);
140 MD5_STEP (MD5_I , a, b, c, d, wc_t, MD5C34, MD5S30);
141 MD5_STEP (MD5_I , d, a, b, c, w3_t, MD5C35, MD5S31);
142 MD5_STEP (MD5_I , c, d, a, b, wa_t, MD5C36, MD5S32);
143 MD5_STEP (MD5_I , b, c, d, a, w1_t, MD5C37, MD5S33);
144 MD5_STEP (MD5_I , a, b, c, d, w8_t, MD5C38, MD5S30);
145 MD5_STEP (MD5_I , d, a, b, c, wf_t, MD5C39, MD5S31);
146 MD5_STEP (MD5_I , c, d, a, b, w6_t, MD5C3a, MD5S32);
147 MD5_STEP (MD5_I , b, c, d, a, wd_t, MD5C3b, MD5S33);
148 MD5_STEP (MD5_I , a, b, c, d, w4_t, MD5C3c, MD5S30);
149 MD5_STEP (MD5_I , d, a, b, c, wb_t, MD5C3d, MD5S31);
150 MD5_STEP (MD5_I , c, d, a, b, w2_t, MD5C3e, MD5S32);
151 MD5_STEP (MD5_I , b, c, d, a, w9_t, MD5C3f, MD5S33);
159 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10420_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)
165 const u32 lid = get_local_id (0);
171 const u32 gid = get_global_id (0);
173 if (gid >= gid_max) return;
177 pw_buf0[0] = pws[gid].i[ 0];
178 pw_buf0[1] = pws[gid].i[ 1];
179 pw_buf0[2] = pws[gid].i[ 2];
180 pw_buf0[3] = pws[gid].i[ 3];
184 pw_buf1[0] = pws[gid].i[ 4];
185 pw_buf1[1] = pws[gid].i[ 5];
186 pw_buf1[2] = pws[gid].i[ 6];
187 pw_buf1[3] = pws[gid].i[ 7];
189 const u32 pw_len = pws[gid].pw_len;
197 o_buf[0] = pdf_bufs[salt_pos].o_buf[0];
198 o_buf[1] = pdf_bufs[salt_pos].o_buf[1];
199 o_buf[2] = pdf_bufs[salt_pos].o_buf[2];
200 o_buf[3] = pdf_bufs[salt_pos].o_buf[3];
201 o_buf[4] = pdf_bufs[salt_pos].o_buf[4];
202 o_buf[5] = pdf_bufs[salt_pos].o_buf[5];
203 o_buf[6] = pdf_bufs[salt_pos].o_buf[6];
204 o_buf[7] = pdf_bufs[salt_pos].o_buf[7];
206 u32 P = pdf_bufs[salt_pos].P;
210 id_buf[0] = pdf_bufs[salt_pos].id_buf[0];
211 id_buf[1] = pdf_bufs[salt_pos].id_buf[1];
212 id_buf[2] = pdf_bufs[salt_pos].id_buf[2];
213 id_buf[3] = pdf_bufs[salt_pos].id_buf[3];
219 for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++)
249 const u32 out_len = apply_rules (rules_buf[il_pos].cmds, w0, w1, pw_len);
256 // max length supported by pdf11 is 32
258 w0_t[0] = padding[0];
259 w0_t[1] = padding[1];
260 w0_t[2] = padding[2];
261 w0_t[3] = padding[3];
262 w1_t[0] = padding[4];
263 w1_t[1] = padding[5];
264 w1_t[2] = padding[6];
265 w1_t[3] = padding[7];
275 switch_buffer_by_offset (w0_t, w1_t, w2_t, w3_t, pw_len);
278 // truncate at 32 is wanted, not a bug!
305 md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
324 md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
327 u32x b = digest[1] & 0xff;
334 #include VECT_COMPARE_M
338 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10420_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)
342 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10420_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)
346 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10420_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)
352 const u32 lid = get_local_id (0);
358 const u32 gid = get_global_id (0);
360 if (gid >= gid_max) return;
364 pw_buf0[0] = pws[gid].i[ 0];
365 pw_buf0[1] = pws[gid].i[ 1];
366 pw_buf0[2] = pws[gid].i[ 2];
367 pw_buf0[3] = pws[gid].i[ 3];
371 pw_buf1[0] = pws[gid].i[ 4];
372 pw_buf1[1] = pws[gid].i[ 5];
373 pw_buf1[2] = pws[gid].i[ 6];
374 pw_buf1[3] = pws[gid].i[ 7];
376 const u32 pw_len = pws[gid].pw_len;
382 const u32 search[4] =
384 digests_buf[digests_offset].digest_buf[DGST_R0],
385 digests_buf[digests_offset].digest_buf[DGST_R1],
386 digests_buf[digests_offset].digest_buf[DGST_R2],
387 digests_buf[digests_offset].digest_buf[DGST_R3]
396 o_buf[0] = pdf_bufs[salt_pos].o_buf[0];
397 o_buf[1] = pdf_bufs[salt_pos].o_buf[1];
398 o_buf[2] = pdf_bufs[salt_pos].o_buf[2];
399 o_buf[3] = pdf_bufs[salt_pos].o_buf[3];
400 o_buf[4] = pdf_bufs[salt_pos].o_buf[4];
401 o_buf[5] = pdf_bufs[salt_pos].o_buf[5];
402 o_buf[6] = pdf_bufs[salt_pos].o_buf[6];
403 o_buf[7] = pdf_bufs[salt_pos].o_buf[7];
405 u32 P = pdf_bufs[salt_pos].P;
409 id_buf[0] = pdf_bufs[salt_pos].id_buf[0];
410 id_buf[1] = pdf_bufs[salt_pos].id_buf[1];
411 id_buf[2] = pdf_bufs[salt_pos].id_buf[2];
412 id_buf[3] = pdf_bufs[salt_pos].id_buf[3];
418 for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++)
448 const u32 out_len = apply_rules (rules_buf[il_pos].cmds, w0, w1, pw_len);
455 // max length supported by pdf11 is 32
457 w0_t[0] = padding[0];
458 w0_t[1] = padding[1];
459 w0_t[2] = padding[2];
460 w0_t[3] = padding[3];
461 w1_t[0] = padding[4];
462 w1_t[1] = padding[5];
463 w1_t[2] = padding[6];
464 w1_t[3] = padding[7];
474 switch_buffer_by_offset (w0_t, w1_t, w2_t, w3_t, pw_len);
477 // truncate at 32 is wanted, not a bug!
504 md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
523 md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
526 u32x b = digest[1] & 0xff;
533 #include VECT_COMPARE_S
537 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10420_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)
541 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10420_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)