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 u64 k_sha384[80] =
51 SHA384C00, SHA384C01, SHA384C02, SHA384C03,
52 SHA384C04, SHA384C05, SHA384C06, SHA384C07,
53 SHA384C08, SHA384C09, SHA384C0a, SHA384C0b,
54 SHA384C0c, SHA384C0d, SHA384C0e, SHA384C0f,
55 SHA384C10, SHA384C11, SHA384C12, SHA384C13,
56 SHA384C14, SHA384C15, SHA384C16, SHA384C17,
57 SHA384C18, SHA384C19, SHA384C1a, SHA384C1b,
58 SHA384C1c, SHA384C1d, SHA384C1e, SHA384C1f,
59 SHA384C20, SHA384C21, SHA384C22, SHA384C23,
60 SHA384C24, SHA384C25, SHA384C26, SHA384C27,
61 SHA384C28, SHA384C29, SHA384C2a, SHA384C2b,
62 SHA384C2c, SHA384C2d, SHA384C2e, SHA384C2f,
63 SHA384C30, SHA384C31, SHA384C32, SHA384C33,
64 SHA384C34, SHA384C35, SHA384C36, SHA384C37,
65 SHA384C38, SHA384C39, SHA384C3a, SHA384C3b,
66 SHA384C3c, SHA384C3d, SHA384C3e, SHA384C3f,
67 SHA384C40, SHA384C41, SHA384C42, SHA384C43,
68 SHA384C44, SHA384C45, SHA384C46, SHA384C47,
69 SHA384C48, SHA384C49, SHA384C4a, SHA384C4b,
70 SHA384C4c, SHA384C4d, SHA384C4e, SHA384C4f,
73 static void sha384_transform (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u64x digest[8])
75 u64x w0_t = hl32_to_64 (w0[0], w0[1]);
76 u64x w1_t = hl32_to_64 (w0[2], w0[3]);
77 u64x w2_t = hl32_to_64 (w1[0], w1[1]);
78 u64x w3_t = hl32_to_64 (w1[2], w1[3]);
79 u64x w4_t = hl32_to_64 (w2[0], w2[1]);
80 u64x w5_t = hl32_to_64 (w2[2], w2[3]);
81 u64x w6_t = hl32_to_64 (w3[0], w3[1]);
90 u64x wf_t = hl32_to_64 (w3[2], w3[3]);
101 #define ROUND_EXPAND() \
103 w0_t = SHA384_EXPAND (we_t, w9_t, w1_t, w0_t); \
104 w1_t = SHA384_EXPAND (wf_t, wa_t, w2_t, w1_t); \
105 w2_t = SHA384_EXPAND (w0_t, wb_t, w3_t, w2_t); \
106 w3_t = SHA384_EXPAND (w1_t, wc_t, w4_t, w3_t); \
107 w4_t = SHA384_EXPAND (w2_t, wd_t, w5_t, w4_t); \
108 w5_t = SHA384_EXPAND (w3_t, we_t, w6_t, w5_t); \
109 w6_t = SHA384_EXPAND (w4_t, wf_t, w7_t, w6_t); \
110 w7_t = SHA384_EXPAND (w5_t, w0_t, w8_t, w7_t); \
111 w8_t = SHA384_EXPAND (w6_t, w1_t, w9_t, w8_t); \
112 w9_t = SHA384_EXPAND (w7_t, w2_t, wa_t, w9_t); \
113 wa_t = SHA384_EXPAND (w8_t, w3_t, wb_t, wa_t); \
114 wb_t = SHA384_EXPAND (w9_t, w4_t, wc_t, wb_t); \
115 wc_t = SHA384_EXPAND (wa_t, w5_t, wd_t, wc_t); \
116 wd_t = SHA384_EXPAND (wb_t, w6_t, we_t, wd_t); \
117 we_t = SHA384_EXPAND (wc_t, w7_t, wf_t, we_t); \
118 wf_t = SHA384_EXPAND (wd_t, w8_t, w0_t, wf_t); \
121 #define ROUND_STEP(i) \
123 SHA384_STEP (SHA384_F0o, SHA384_F1o, a, b, c, d, e, f, g, h, w0_t, k_sha384[i + 0]); \
124 SHA384_STEP (SHA384_F0o, SHA384_F1o, h, a, b, c, d, e, f, g, w1_t, k_sha384[i + 1]); \
125 SHA384_STEP (SHA384_F0o, SHA384_F1o, g, h, a, b, c, d, e, f, w2_t, k_sha384[i + 2]); \
126 SHA384_STEP (SHA384_F0o, SHA384_F1o, f, g, h, a, b, c, d, e, w3_t, k_sha384[i + 3]); \
127 SHA384_STEP (SHA384_F0o, SHA384_F1o, e, f, g, h, a, b, c, d, w4_t, k_sha384[i + 4]); \
128 SHA384_STEP (SHA384_F0o, SHA384_F1o, d, e, f, g, h, a, b, c, w5_t, k_sha384[i + 5]); \
129 SHA384_STEP (SHA384_F0o, SHA384_F1o, c, d, e, f, g, h, a, b, w6_t, k_sha384[i + 6]); \
130 SHA384_STEP (SHA384_F0o, SHA384_F1o, b, c, d, e, f, g, h, a, w7_t, k_sha384[i + 7]); \
131 SHA384_STEP (SHA384_F0o, SHA384_F1o, a, b, c, d, e, f, g, h, w8_t, k_sha384[i + 8]); \
132 SHA384_STEP (SHA384_F0o, SHA384_F1o, h, a, b, c, d, e, f, g, w9_t, k_sha384[i + 9]); \
133 SHA384_STEP (SHA384_F0o, SHA384_F1o, g, h, a, b, c, d, e, f, wa_t, k_sha384[i + 10]); \
134 SHA384_STEP (SHA384_F0o, SHA384_F1o, f, g, h, a, b, c, d, e, wb_t, k_sha384[i + 11]); \
135 SHA384_STEP (SHA384_F0o, SHA384_F1o, e, f, g, h, a, b, c, d, wc_t, k_sha384[i + 12]); \
136 SHA384_STEP (SHA384_F0o, SHA384_F1o, d, e, f, g, h, a, b, c, wd_t, k_sha384[i + 13]); \
137 SHA384_STEP (SHA384_F0o, SHA384_F1o, c, d, e, f, g, h, a, b, we_t, k_sha384[i + 14]); \
138 SHA384_STEP (SHA384_F0o, SHA384_F1o, b, c, d, e, f, g, h, a, wf_t, k_sha384[i + 15]); \
143 ROUND_EXPAND (); ROUND_STEP (16);
144 ROUND_EXPAND (); ROUND_STEP (32);
145 ROUND_EXPAND (); ROUND_STEP (48);
146 ROUND_EXPAND (); ROUND_STEP (64);
169 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10800_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 void *esalt_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)
175 const u32 lid = get_local_id (0);
181 const u32 gid = get_global_id (0);
183 if (gid >= gid_max) return;
187 pw_buf0[0] = pws[gid].i[ 0];
188 pw_buf0[1] = pws[gid].i[ 1];
189 pw_buf0[2] = pws[gid].i[ 2];
190 pw_buf0[3] = pws[gid].i[ 3];
194 pw_buf1[0] = pws[gid].i[ 4];
195 pw_buf1[1] = pws[gid].i[ 5];
196 pw_buf1[2] = pws[gid].i[ 6];
197 pw_buf1[3] = pws[gid].i[ 7];
199 const u32 pw_len = pws[gid].pw_len;
205 for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++)
235 const u32 out_len = apply_rules (rules_buf[il_pos].cmds, w0, w1, pw_len);
237 append_0x80_2 (w0, w1, out_len);
248 w0_t[0] = swap_workaround (w0[0]);
249 w0_t[1] = swap_workaround (w0[1]);
250 w0_t[2] = swap_workaround (w0[2]);
251 w0_t[3] = swap_workaround (w0[3]);
252 w1_t[0] = swap_workaround (w1[0]);
253 w1_t[1] = swap_workaround (w1[1]);
254 w1_t[2] = swap_workaround (w1[2]);
255 w1_t[3] = swap_workaround (w1[3]);
256 w2_t[0] = swap_workaround (w2[0]);
257 w2_t[1] = swap_workaround (w2[1]);
258 w2_t[2] = swap_workaround (w2[2]);
259 w2_t[3] = swap_workaround (w2[3]);
260 w3_t[0] = swap_workaround (w3[0]);
261 w3_t[1] = swap_workaround (w3[1]);
263 w3_t[3] = out_len * 8;
267 digest[0] = SHA384M_A;
268 digest[1] = SHA384M_B;
269 digest[2] = SHA384M_C;
270 digest[3] = SHA384M_D;
271 digest[4] = SHA384M_E;
272 digest[5] = SHA384M_F;
273 digest[6] = SHA384M_G;
274 digest[7] = SHA384M_H;
276 sha384_transform (w0_t, w1_t, w2_t, w3_t, digest);
278 const u32x r0 = l32_from_64 (digest[3]);
279 const u32x r1 = h32_from_64 (digest[3]);
280 const u32x r2 = l32_from_64 (digest[2]);
281 const u32x r3 = h32_from_64 (digest[2]);
283 #include VECT_COMPARE_M
287 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10800_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 void *esalt_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)
291 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10800_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 void *esalt_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)
295 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10800_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 void *esalt_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)
301 const u32 lid = get_local_id (0);
307 const u32 gid = get_global_id (0);
309 if (gid >= gid_max) return;
313 pw_buf0[0] = pws[gid].i[ 0];
314 pw_buf0[1] = pws[gid].i[ 1];
315 pw_buf0[2] = pws[gid].i[ 2];
316 pw_buf0[3] = pws[gid].i[ 3];
320 pw_buf1[0] = pws[gid].i[ 4];
321 pw_buf1[1] = pws[gid].i[ 5];
322 pw_buf1[2] = pws[gid].i[ 6];
323 pw_buf1[3] = pws[gid].i[ 7];
325 const u32 pw_len = pws[gid].pw_len;
331 const u32 search[4] =
333 digests_buf[digests_offset].digest_buf[DGST_R0],
334 digests_buf[digests_offset].digest_buf[DGST_R1],
335 digests_buf[digests_offset].digest_buf[DGST_R2],
336 digests_buf[digests_offset].digest_buf[DGST_R3]
343 for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++)
373 const u32 out_len = apply_rules (rules_buf[il_pos].cmds, w0, w1, pw_len);
375 append_0x80_2 (w0, w1, out_len);
386 w0_t[0] = swap_workaround (w0[0]);
387 w0_t[1] = swap_workaround (w0[1]);
388 w0_t[2] = swap_workaround (w0[2]);
389 w0_t[3] = swap_workaround (w0[3]);
390 w1_t[0] = swap_workaround (w1[0]);
391 w1_t[1] = swap_workaround (w1[1]);
392 w1_t[2] = swap_workaround (w1[2]);
393 w1_t[3] = swap_workaround (w1[3]);
394 w2_t[0] = swap_workaround (w2[0]);
395 w2_t[1] = swap_workaround (w2[1]);
396 w2_t[2] = swap_workaround (w2[2]);
397 w2_t[3] = swap_workaround (w2[3]);
398 w3_t[0] = swap_workaround (w3[0]);
399 w3_t[1] = swap_workaround (w3[1]);
401 w3_t[3] = out_len * 8;
405 digest[0] = SHA384M_A;
406 digest[1] = SHA384M_B;
407 digest[2] = SHA384M_C;
408 digest[3] = SHA384M_D;
409 digest[4] = SHA384M_E;
410 digest[5] = SHA384M_F;
411 digest[6] = SHA384M_G;
412 digest[7] = SHA384M_H;
414 sha384_transform (w0_t, w1_t, w2_t, w3_t, digest);
416 const u32x r0 = l32_from_64 (digest[3]);
417 const u32x r1 = h32_from_64 (digest[3]);
418 const u32x r2 = l32_from_64 (digest[2]);
419 const u32x r3 = h32_from_64 (digest[2]);
421 #include VECT_COMPARE_S
425 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10800_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 void *esalt_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)
429 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10800_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 void *esalt_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)