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"
27 #include "common_nv.c"
30 #define VECT_COMPARE_M "check_multi_vect1_comp4.c"
33 __device__ static void md5_transform (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u32x digest[4])
59 MD5_STEP (MD5_Fo, a, b, c, d, w0_t, MD5C00, MD5S00);
60 MD5_STEP (MD5_Fo, d, a, b, c, w1_t, MD5C01, MD5S01);
61 MD5_STEP (MD5_Fo, c, d, a, b, w2_t, MD5C02, MD5S02);
62 MD5_STEP (MD5_Fo, b, c, d, a, w3_t, MD5C03, MD5S03);
63 MD5_STEP (MD5_Fo, a, b, c, d, w4_t, MD5C04, MD5S00);
64 MD5_STEP (MD5_Fo, d, a, b, c, w5_t, MD5C05, MD5S01);
65 MD5_STEP (MD5_Fo, c, d, a, b, w6_t, MD5C06, MD5S02);
66 MD5_STEP (MD5_Fo, b, c, d, a, w7_t, MD5C07, MD5S03);
67 MD5_STEP (MD5_Fo, a, b, c, d, w8_t, MD5C08, MD5S00);
68 MD5_STEP (MD5_Fo, d, a, b, c, w9_t, MD5C09, MD5S01);
69 MD5_STEP (MD5_Fo, c, d, a, b, wa_t, MD5C0a, MD5S02);
70 MD5_STEP (MD5_Fo, b, c, d, a, wb_t, MD5C0b, MD5S03);
71 MD5_STEP (MD5_Fo, a, b, c, d, wc_t, MD5C0c, MD5S00);
72 MD5_STEP (MD5_Fo, d, a, b, c, wd_t, MD5C0d, MD5S01);
73 MD5_STEP (MD5_Fo, c, d, a, b, we_t, MD5C0e, MD5S02);
74 MD5_STEP (MD5_Fo, b, c, d, a, wf_t, MD5C0f, MD5S03);
76 MD5_STEP (MD5_Go, a, b, c, d, w1_t, MD5C10, MD5S10);
77 MD5_STEP (MD5_Go, d, a, b, c, w6_t, MD5C11, MD5S11);
78 MD5_STEP (MD5_Go, c, d, a, b, wb_t, MD5C12, MD5S12);
79 MD5_STEP (MD5_Go, b, c, d, a, w0_t, MD5C13, MD5S13);
80 MD5_STEP (MD5_Go, a, b, c, d, w5_t, MD5C14, MD5S10);
81 MD5_STEP (MD5_Go, d, a, b, c, wa_t, MD5C15, MD5S11);
82 MD5_STEP (MD5_Go, c, d, a, b, wf_t, MD5C16, MD5S12);
83 MD5_STEP (MD5_Go, b, c, d, a, w4_t, MD5C17, MD5S13);
84 MD5_STEP (MD5_Go, a, b, c, d, w9_t, MD5C18, MD5S10);
85 MD5_STEP (MD5_Go, d, a, b, c, we_t, MD5C19, MD5S11);
86 MD5_STEP (MD5_Go, c, d, a, b, w3_t, MD5C1a, MD5S12);
87 MD5_STEP (MD5_Go, b, c, d, a, w8_t, MD5C1b, MD5S13);
88 MD5_STEP (MD5_Go, a, b, c, d, wd_t, MD5C1c, MD5S10);
89 MD5_STEP (MD5_Go, d, a, b, c, w2_t, MD5C1d, MD5S11);
90 MD5_STEP (MD5_Go, c, d, a, b, w7_t, MD5C1e, MD5S12);
91 MD5_STEP (MD5_Go, b, c, d, a, wc_t, MD5C1f, MD5S13);
93 MD5_STEP (MD5_H1, a, b, c, d, w5_t, MD5C20, MD5S20);
94 MD5_STEP (MD5_H2, d, a, b, c, w8_t, MD5C21, MD5S21);
95 MD5_STEP (MD5_H1, c, d, a, b, wb_t, MD5C22, MD5S22);
96 MD5_STEP (MD5_H2, b, c, d, a, we_t, MD5C23, MD5S23);
97 MD5_STEP (MD5_H1, a, b, c, d, w1_t, MD5C24, MD5S20);
98 MD5_STEP (MD5_H2, d, a, b, c, w4_t, MD5C25, MD5S21);
99 MD5_STEP (MD5_H1, c, d, a, b, w7_t, MD5C26, MD5S22);
100 MD5_STEP (MD5_H2, b, c, d, a, wa_t, MD5C27, MD5S23);
101 MD5_STEP (MD5_H1, a, b, c, d, wd_t, MD5C28, MD5S20);
102 MD5_STEP (MD5_H2, d, a, b, c, w0_t, MD5C29, MD5S21);
103 MD5_STEP (MD5_H1, c, d, a, b, w3_t, MD5C2a, MD5S22);
104 MD5_STEP (MD5_H2, b, c, d, a, w6_t, MD5C2b, MD5S23);
105 MD5_STEP (MD5_H1, a, b, c, d, w9_t, MD5C2c, MD5S20);
106 MD5_STEP (MD5_H2, d, a, b, c, wc_t, MD5C2d, MD5S21);
107 MD5_STEP (MD5_H1, c, d, a, b, wf_t, MD5C2e, MD5S22);
108 MD5_STEP (MD5_H2, b, c, d, a, w2_t, MD5C2f, MD5S23);
110 MD5_STEP (MD5_I , a, b, c, d, w0_t, MD5C30, MD5S30);
111 MD5_STEP (MD5_I , d, a, b, c, w7_t, MD5C31, MD5S31);
112 MD5_STEP (MD5_I , c, d, a, b, we_t, MD5C32, MD5S32);
113 MD5_STEP (MD5_I , b, c, d, a, w5_t, MD5C33, MD5S33);
114 MD5_STEP (MD5_I , a, b, c, d, wc_t, MD5C34, MD5S30);
115 MD5_STEP (MD5_I , d, a, b, c, w3_t, MD5C35, MD5S31);
116 MD5_STEP (MD5_I , c, d, a, b, wa_t, MD5C36, MD5S32);
117 MD5_STEP (MD5_I , b, c, d, a, w1_t, MD5C37, MD5S33);
118 MD5_STEP (MD5_I , a, b, c, d, w8_t, MD5C38, MD5S30);
119 MD5_STEP (MD5_I , d, a, b, c, wf_t, MD5C39, MD5S31);
120 MD5_STEP (MD5_I , c, d, a, b, w6_t, MD5C3a, MD5S32);
121 MD5_STEP (MD5_I , b, c, d, a, wd_t, MD5C3b, MD5S33);
122 MD5_STEP (MD5_I , a, b, c, d, w4_t, MD5C3c, MD5S30);
123 MD5_STEP (MD5_I , d, a, b, c, wb_t, MD5C3d, MD5S31);
124 MD5_STEP (MD5_I , c, d, a, b, w2_t, MD5C3e, MD5S32);
125 MD5_STEP (MD5_I , b, c, d, a, w9_t, MD5C3f, MD5S33);
133 __device__ static void hmac_md5_pad (u32x w0[4], u32x w1[4], u32x w2[4], u32x w3[4], u32x ipad[4], u32x opad[4])
135 w0[0] = w0[0] ^ 0x36363636;
136 w0[1] = w0[1] ^ 0x36363636;
137 w0[2] = w0[2] ^ 0x36363636;
138 w0[3] = w0[3] ^ 0x36363636;
139 w1[0] = w1[0] ^ 0x36363636;
140 w1[1] = w1[1] ^ 0x36363636;
141 w1[2] = w1[2] ^ 0x36363636;
142 w1[3] = w1[3] ^ 0x36363636;
143 w2[0] = w2[0] ^ 0x36363636;
144 w2[1] = w2[1] ^ 0x36363636;
145 w2[2] = w2[2] ^ 0x36363636;
146 w2[3] = w2[3] ^ 0x36363636;
147 w3[0] = w3[0] ^ 0x36363636;
148 w3[1] = w3[1] ^ 0x36363636;
149 w3[2] = w3[2] ^ 0x36363636;
150 w3[3] = w3[3] ^ 0x36363636;
157 md5_transform (w0, w1, w2, w3, ipad);
159 w0[0] = w0[0] ^ 0x6a6a6a6a;
160 w0[1] = w0[1] ^ 0x6a6a6a6a;
161 w0[2] = w0[2] ^ 0x6a6a6a6a;
162 w0[3] = w0[3] ^ 0x6a6a6a6a;
163 w1[0] = w1[0] ^ 0x6a6a6a6a;
164 w1[1] = w1[1] ^ 0x6a6a6a6a;
165 w1[2] = w1[2] ^ 0x6a6a6a6a;
166 w1[3] = w1[3] ^ 0x6a6a6a6a;
167 w2[0] = w2[0] ^ 0x6a6a6a6a;
168 w2[1] = w2[1] ^ 0x6a6a6a6a;
169 w2[2] = w2[2] ^ 0x6a6a6a6a;
170 w2[3] = w2[3] ^ 0x6a6a6a6a;
171 w3[0] = w3[0] ^ 0x6a6a6a6a;
172 w3[1] = w3[1] ^ 0x6a6a6a6a;
173 w3[2] = w3[2] ^ 0x6a6a6a6a;
174 w3[3] = w3[3] ^ 0x6a6a6a6a;
181 md5_transform (w0, w1, w2, w3, opad);
184 __device__ static void hmac_md5_run (u32x w0[4], u32x w1[4], u32x w2[4], u32x w3[4], u32x ipad[4], u32x opad[4], u32x digest[4])
191 md5_transform (w0, w1, w2, w3, digest);
207 w3[2] = (64 + 16) * 8;
215 md5_transform (w0, w1, w2, w3, digest);
218 extern "C" __global__ void __launch_bounds__ (256, 1) m11900_init (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, pbkdf2_md5_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 pbkdf2_md5_t *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)
224 const u32 gid = (blockIdx.x * blockDim.x) + threadIdx.x;
226 if (gid >= gid_max) return;
230 w0[0] = pws[gid].i[ 0];
231 w0[1] = pws[gid].i[ 1];
232 w0[2] = pws[gid].i[ 2];
233 w0[3] = pws[gid].i[ 3];
237 w1[0] = pws[gid].i[ 4];
238 w1[1] = pws[gid].i[ 5];
239 w1[2] = pws[gid].i[ 6];
240 w1[3] = pws[gid].i[ 7];
244 w2[0] = pws[gid].i[ 8];
245 w2[1] = pws[gid].i[ 9];
246 w2[2] = pws[gid].i[10];
247 w2[3] = pws[gid].i[11];
251 w3[0] = pws[gid].i[12];
252 w3[1] = pws[gid].i[13];
253 w3[2] = pws[gid].i[14];
254 w3[3] = pws[gid].i[15];
260 const u32 salt_len = salt_bufs[salt_pos].salt_len;
267 esalt_buf0[0] = esalt_bufs[salt_pos].salt_buf[ 0];
268 esalt_buf0[1] = esalt_bufs[salt_pos].salt_buf[ 1];
269 esalt_buf0[2] = esalt_bufs[salt_pos].salt_buf[ 2];
270 esalt_buf0[3] = esalt_bufs[salt_pos].salt_buf[ 3];
271 esalt_buf1[0] = esalt_bufs[salt_pos].salt_buf[ 4];
272 esalt_buf1[1] = esalt_bufs[salt_pos].salt_buf[ 5];
273 esalt_buf1[2] = esalt_bufs[salt_pos].salt_buf[ 6];
274 esalt_buf1[3] = esalt_bufs[salt_pos].salt_buf[ 7];
275 esalt_buf2[0] = esalt_bufs[salt_pos].salt_buf[ 8];
276 esalt_buf2[1] = esalt_bufs[salt_pos].salt_buf[ 9];
277 esalt_buf2[2] = esalt_bufs[salt_pos].salt_buf[10];
278 esalt_buf2[3] = esalt_bufs[salt_pos].salt_buf[11];
279 esalt_buf3[0] = esalt_bufs[salt_pos].salt_buf[12];
280 esalt_buf3[1] = esalt_bufs[salt_pos].salt_buf[13];
281 esalt_buf3[2] = (64 + salt_len + 4) * 8;
287 hmac_md5_pad (w0, w1, w2, w3, ipad, opad);
289 tmps[gid].ipad[0] = ipad[0];
290 tmps[gid].ipad[1] = ipad[1];
291 tmps[gid].ipad[2] = ipad[2];
292 tmps[gid].ipad[3] = ipad[3];
294 tmps[gid].opad[0] = opad[0];
295 tmps[gid].opad[1] = opad[1];
296 tmps[gid].opad[2] = opad[2];
297 tmps[gid].opad[3] = opad[3];
299 for (u32 i = 0, j = 1; i < 4; i += 4, j += 1)
303 hmac_md5_run (esalt_buf0, esalt_buf1, esalt_buf2, esalt_buf3, ipad, opad, dgst);
305 tmps[gid].dgst[i + 0] = dgst[0];
306 tmps[gid].dgst[i + 1] = dgst[1];
307 tmps[gid].dgst[i + 2] = dgst[2];
308 tmps[gid].dgst[i + 3] = dgst[3];
310 tmps[gid].out[i + 0] = dgst[0];
311 tmps[gid].out[i + 1] = dgst[1];
312 tmps[gid].out[i + 2] = dgst[2];
313 tmps[gid].out[i + 3] = dgst[3];
317 extern "C" __global__ void __launch_bounds__ (256, 1) m11900_loop (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, pbkdf2_md5_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 pbkdf2_md5_t *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)
319 const u32 gid = (blockIdx.x * blockDim.x) + threadIdx.x;
321 if (gid >= gid_max) return;
325 ipad[0] = tmps[gid].ipad[0];
326 ipad[1] = tmps[gid].ipad[1];
327 ipad[2] = tmps[gid].ipad[2];
328 ipad[3] = tmps[gid].ipad[3];
332 opad[0] = tmps[gid].opad[0];
333 opad[1] = tmps[gid].opad[1];
334 opad[2] = tmps[gid].opad[2];
335 opad[3] = tmps[gid].opad[3];
337 for (u32 i = 0; i < 4; i += 4)
341 dgst[0] = tmps[gid].dgst[i + 0];
342 dgst[1] = tmps[gid].dgst[i + 1];
343 dgst[2] = tmps[gid].dgst[i + 2];
344 dgst[3] = tmps[gid].dgst[i + 3];
348 out[0] = tmps[gid].out[i + 0];
349 out[1] = tmps[gid].out[i + 1];
350 out[2] = tmps[gid].out[i + 2];
351 out[3] = tmps[gid].out[i + 3];
353 for (u32 j = 0; j < loop_cnt; j++)
374 w3[2] = (64 + 16) * 8;
377 hmac_md5_run (w0, w1, w2, w3, ipad, opad, dgst);
385 tmps[gid].dgst[i + 0] = dgst[0];
386 tmps[gid].dgst[i + 1] = dgst[1];
387 tmps[gid].dgst[i + 2] = dgst[2];
388 tmps[gid].dgst[i + 3] = dgst[3];
390 tmps[gid].out[i + 0] = out[0];
391 tmps[gid].out[i + 1] = out[1];
392 tmps[gid].out[i + 2] = out[2];
393 tmps[gid].out[i + 3] = out[3];
397 extern "C" __global__ void __launch_bounds__ (256, 1) m11900_comp (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, pbkdf2_md5_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 pbkdf2_md5_t *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;
409 const u32x r0 = tmps[gid].out[DGST_R0];
410 const u32x r1 = tmps[gid].out[DGST_R1];
411 const u32x r2 = tmps[gid].out[DGST_R2];
412 const u32x r3 = tmps[gid].out[DGST_R3];
416 #include VECT_COMPARE_M