OpenCL/types_ocl.c

   1 /**
   2  * Authors.....: Jens Steube <jens.steube@gmail.com>
   3  *               magnum <john.magnum@hushmail.com>
   4  *
   5  * License.....: MIT
   6  */
   7
   8 typedef uchar  u8;
   9 typedef ushort u16;
  10 typedef uint   u32;
  11 typedef ulong  u64;
  12
  13 #ifndef NEW_SIMD_CODE
  14 #undef  VECT_SIZE
  15 #define VECT_SIZE 1
  16 #endif
  17
  18 #define CONCAT(a, b)       a##b
  19 #define VTYPE(type, width) CONCAT(type, width)
  20
  21 #if VECT_SIZE == 1
  22 typedef uchar   u8x;
  23 typedef ushort  u16x;
  24 typedef uint    u32x;
  25 typedef ulong   u64x;
  26 #else
  27 typedef VTYPE(uchar,  VECT_SIZE)  u8x;
  28 typedef VTYPE(ushort, VECT_SIZE) u16x;
  29 typedef VTYPE(uint,   VECT_SIZE) u32x;
  30 typedef VTYPE(ulong,  VECT_SIZE) u64x;
  31 #endif
  32
  33 inline u32 l32_from_64_S (u64 a)
  34 {
  35   const u32 r = (u32) (a);
  36
  37   return r;
  38 }
  39
  40 inline u32 h32_from_64_S (u64 a)
  41 {
  42   a >>= 32;
  43
  44   const u32 r = (u32) (a);
  45
  46   return r;
  47 }
  48
  49 inline u64 hl32_to_64_S (const u32 a, const u32 b)
  50 {
  51   return as_ulong ((uint2) (b, a));
  52 }
  53
  54 inline u32x l32_from_64 (u64x a)
  55 {
  56   u32x r;
  57
  58   #if VECT_SIZE == 1
  59   r    = (u32) a;
  60   #endif
  61
  62   #if VECT_SIZE >= 2
  63   r.s0 = (u32) a.s0;
  64   r.s1 = (u32) a.s1;
  65   #endif
  66
  67   #if VECT_SIZE >= 4
  68   r.s2 = (u32) a.s2;
  69   r.s3 = (u32) a.s3;
  70   #endif
  71
  72   #if VECT_SIZE >= 8
  73   r.s4 = (u32) a.s4;
  74   r.s5 = (u32) a.s5;
  75   r.s6 = (u32) a.s6;
  76   r.s7 = (u32) a.s7;
  77   #endif
  78
  79   #if VECT_SIZE >= 16
  80   r.s8 = (u32) a.s8;
  81   r.s9 = (u32) a.s9;
  82   r.sa = (u32) a.sa;
  83   r.sb = (u32) a.sb;
  84   r.sc = (u32) a.sc;
  85   r.sd = (u32) a.sd;
  86   r.se = (u32) a.se;
  87   r.sf = (u32) a.sf;
  88   #endif
  89
  90   return r;
  91 }
  92
  93 inline u32x h32_from_64 (u64x a)
  94 {
  95   a >>= 32;
  96
  97   u32x r;
  98
  99   #if VECT_SIZE == 1
 100   r    = (u32) a;
 101   #endif
 102
 103   #if VECT_SIZE >= 2
 104   r.s0 = (u32) a.s0;
 105   r.s1 = (u32) a.s1;
 106   #endif
 107
 108   #if VECT_SIZE >= 4
 109   r.s2 = (u32) a.s2;
 110   r.s3 = (u32) a.s3;
 111   #endif
 112
 113   #if VECT_SIZE >= 8
 114   r.s4 = (u32) a.s4;
 115   r.s5 = (u32) a.s5;
 116   r.s6 = (u32) a.s6;
 117   r.s7 = (u32) a.s7;
 118   #endif
 119
 120   #if VECT_SIZE >= 16
 121   r.s8 = (u32) a.s8;
 122   r.s9 = (u32) a.s9;
 123   r.sa = (u32) a.sa;
 124   r.sb = (u32) a.sb;
 125   r.sc = (u32) a.sc;
 126   r.sd = (u32) a.sd;
 127   r.se = (u32) a.se;
 128   r.sf = (u32) a.sf;
 129   #endif
 130
 131   return r;
 132 }
 133
 134 inline u64x hl32_to_64 (const u32x a, const u32x b)
 135 {
 136   u64x r;
 137
 138   #if VECT_SIZE == 1
 139   r    = as_ulong ((uint2) (b,    a));
 140   #endif
 141
 142   #if VECT_SIZE >= 2
 143   r.s0 = as_ulong ((uint2) (b.s0, a.s0));
 144   r.s1 = as_ulong ((uint2) (b.s1, a.s1));
 145   #endif
 146
 147   #if VECT_SIZE >= 4
 148   r.s2 = as_ulong ((uint2) (b.s2, a.s2));
 149   r.s3 = as_ulong ((uint2) (b.s3, a.s3));
 150   #endif
 151
 152   #if VECT_SIZE >= 8
 153   r.s4 = as_ulong ((uint2) (b.s4, a.s4));
 154   r.s5 = as_ulong ((uint2) (b.s5, a.s5));
 155   r.s6 = as_ulong ((uint2) (b.s6, a.s6));
 156   r.s7 = as_ulong ((uint2) (b.s7, a.s7));
 157   #endif
 158
 159   #if VECT_SIZE >= 16
 160   r.s8 = as_ulong ((uint2) (b.s8, a.s8));
 161   r.s9 = as_ulong ((uint2) (b.s9, a.s9));
 162   r.sa = as_ulong ((uint2) (b.sa, a.sa));
 163   r.sb = as_ulong ((uint2) (b.sb, a.sb));
 164   r.sc = as_ulong ((uint2) (b.sc, a.sc));
 165   r.sd = as_ulong ((uint2) (b.sd, a.sd));
 166   r.se = as_ulong ((uint2) (b.se, a.se));
 167   r.sf = as_ulong ((uint2) (b.sf, a.sf));
 168   #endif
 169
 170   return r;
 171 }
 172
 173 #ifdef IS_AMD
 174 inline u32 swap32_S (const u32 v)
 175 {
 176   return (as_uint (as_uchar4 (v).s3210));
 177 }
 178
 179 inline u64 swap64_S (const u64 v)
 180 {
 181   return (as_ulong (as_uchar8 (v).s76543210));
 182 }
 183
 184 inline u32 rotr32_S (const u32 a, const u32 n)
 185 {
 186   return rotate (a, 32 - n);
 187 }
 188
 189 inline u32 rotl32_S (const u32 a, const u32 n)
 190 {
 191   return rotate (a, n);
 192 }
 193
 194 inline u64 rotr64_S (const u64 a, const u32 n)
 195 {
 196   const u32 a0 = h32_from_64_S (a);
 197   const u32 a1 = l32_from_64_S (a);
 198
 199   const u32 t0 = (n >= 32) ? amd_bitalign (a0, a1, n - 32) : amd_bitalign (a1, a0, n);
 200   const u32 t1 = (n >= 32) ? amd_bitalign (a1, a0, n - 32) : amd_bitalign (a0, a1, n);
 201
 202   const u64 r = hl32_to_64_S (t0, t1);
 203
 204   return r;
 205 }
 206
 207 inline u64 rotl64_S (const u64 a, const u32 n)
 208 {
 209   return rotr64_S (a, 64 - n);
 210 }
 211
 212 inline u32x swap32 (const u32x v)
 213 {
 214   return ((v >> 24) & 0x000000ff)
 215        | ((v >>  8) & 0x0000ff00)
 216        | ((v <<  8) & 0x00ff0000)
 217        | ((v << 24) & 0xff000000);
 218 }
 219
 220 inline u64x swap64 (const u64x v)
 221 {
 222   return ((v >> 56) & 0x00000000000000ff)
 223        | ((v >> 40) & 0x000000000000ff00)
 224        | ((v >> 24) & 0x0000000000ff0000)
 225        | ((v >>  8) & 0x00000000ff000000)
 226        | ((v <<  8) & 0x000000ff00000000)
 227        | ((v << 24) & 0x0000ff0000000000)
 228        | ((v << 40) & 0x00ff000000000000)
 229        | ((v << 56) & 0xff00000000000000);
 230 }
 231
 232 inline u32x rotr32 (const u32x a, const u32 n)
 233 {
 234   return rotate (a, 32 - n);
 235 }
 236
 237 inline u32x rotl32 (const u32x a, const u32 n)
 238 {
 239   return rotate (a, n);
 240 }
 241
 242 inline u64x rotr64 (const u64x a, const u32 n)
 243 {
 244   const u32x a0 = h32_from_64 (a);
 245   const u32x a1 = l32_from_64 (a);
 246
 247   const u32x t0 = (n >= 32) ? amd_bitalign (a0, a1, n - 32) : amd_bitalign (a1, a0, n);
 248   const u32x t1 = (n >= 32) ? amd_bitalign (a1, a0, n - 32) : amd_bitalign (a0, a1, n);
 249
 250   const u64x r = hl32_to_64 (t0, t1);
 251
 252   return r;
 253 }
 254
 255 inline u64x rotl64 (const u64x a, const u32 n)
 256 {
 257   return rotr64 (a, 64 - n);
 258 }
 259
 260 inline u32x __bfe (const u32x a, const u32x b, const u32x c)
 261 {
 262   return amd_bfe (a, b, c);
 263 }
 264
 265 inline u32 __bfe_S (const u32 a, const u32 b, const u32 c)
 266 {
 267   return amd_bfe (a, b, c);
 268 }
 269
 270 inline u32 amd_bytealign_S (const u32 a, const u32 b, const u32 c)
 271 {
 272   return amd_bytealign (a, b, c);
 273 }
 274 #endif
 275
 276 #ifdef IS_NV
 277 inline u32 swap32_S (const u32 v)
 278 {
 279   u32 r;
 280
 281   asm ("prmt.b32 %0, %1, 0, 0x0123;" : "=r"(r) : "r"(v));
 282
 283   return r;
 284 }
 285
 286 inline u64 swap64_S (const u64 v)
 287 {
 288   u32 il;
 289   u32 ir;
 290
 291   asm ("mov.b64 {%0, %1}, %2;" : "=r"(il), "=r"(ir) : "l"(v));
 292
 293   u32 tl;
 294   u32 tr;
 295
 296   asm ("prmt.b32 %0, %1, 0, 0x0123;" : "=r"(tl) : "r"(il));
 297   asm ("prmt.b32 %0, %1, 0, 0x0123;" : "=r"(tr) : "r"(ir));
 298
 299   u64 r;
 300
 301   asm ("mov.b64 %0, {%1, %2};" : "=l"(r) : "r"(tr), "r"(tl));
 302
 303   return r;
 304 }
 305
 306 inline u32 rotr32_S (const u32 a, const u32 n)
 307 {
 308   return rotate (a, 32 - n);
 309 }
 310
 311 inline u32 rotl32_S (const u32 a, const u32 n)
 312 {
 313   return rotate (a, n);
 314 }
 315
 316 inline u64 rotr64_S (const u64 a, const u32 n)
 317 {
 318   return rotate (a, (u64) 64 - n);
 319 }
 320
 321 inline u64 rotl64_S (const u64 a, const u32 n)
 322 {
 323   return rotr64_S (a, 64 - n);
 324 }
 325
 326 inline u32x swap32 (const u32x v)
 327 {
 328   return ((v >> 24) & 0x000000ff)
 329        | ((v >>  8) & 0x0000ff00)
 330        | ((v <<  8) & 0x00ff0000)
 331        | ((v << 24) & 0xff000000);
 332 }
 333
 334 inline u64x swap64 (const u64x v)
 335 {
 336   return ((v >> 56) & 0x00000000000000ff)
 337        | ((v >> 40) & 0x000000000000ff00)
 338        | ((v >> 24) & 0x0000000000ff0000)
 339        | ((v >>  8) & 0x00000000ff000000)
 340        | ((v <<  8) & 0x000000ff00000000)
 341        | ((v << 24) & 0x0000ff0000000000)
 342        | ((v << 40) & 0x00ff000000000000)
 343        | ((v << 56) & 0xff00000000000000);
 344 }
 345
 346 inline u32x rotr32 (const u32x a, const u32 n)
 347 {
 348   return rotate (a, 32 - n);
 349 }
 350
 351 inline u32x rotl32 (const u32x a, const u32 n)
 352 {
 353   return rotate (a, n);
 354 }
 355
 356 inline u64x rotr64 (const u64x a, const u32 n)
 357 {
 358   return rotate (a, (u64) 64 - n);
 359 }
 360
 361 inline u64x rotl64 (const u64x a, const u32 n)
 362 {
 363   return rotate (a, (u64) n);
 364 }
 365
 366 inline u32x __byte_perm (const u32x a, const u32x b, const u32x c)
 367 {
 368   u32x r;
 369
 370   #if VECT_SIZE == 1
 371   asm ("prmt.b32 %0, %1, %2, %3;" : "=r"(r)    : "r"(a),    "r"(b),    "r"(c)   );
 372   #endif
 373
 374   #if VECT_SIZE >= 2
 375   asm ("prmt.b32 %0, %1, %2, %3;" : "=r"(r.s0) : "r"(a.s0), "r"(b.s0), "r"(c.s0));
 376   asm ("prmt.b32 %0, %1, %2, %3;" : "=r"(r.s1) : "r"(a.s1), "r"(b.s1), "r"(c.s1));
 377   #endif
 378
 379   #if VECT_SIZE >= 4
 380   asm ("prmt.b32 %0, %1, %2, %3;" : "=r"(r.s2) : "r"(a.s2), "r"(b.s2), "r"(c.s2));
 381   asm ("prmt.b32 %0, %1, %2, %3;" : "=r"(r.s3) : "r"(a.s3), "r"(b.s3), "r"(c.s3));
 382   #endif
 383
 384   #if VECT_SIZE >= 8
 385   asm ("prmt.b32 %0, %1, %2, %3;" : "=r"(r.s4) : "r"(a.s4), "r"(b.s4), "r"(c.s4));
 386   asm ("prmt.b32 %0, %1, %2, %3;" : "=r"(r.s5) : "r"(a.s5), "r"(b.s5), "r"(c.s5));
 387   asm ("prmt.b32 %0, %1, %2, %3;" : "=r"(r.s6) : "r"(a.s6), "r"(b.s6), "r"(c.s6));
 388   asm ("prmt.b32 %0, %1, %2, %3;" : "=r"(r.s7) : "r"(a.s7), "r"(b.s7), "r"(c.s7));
 389   #endif
 390
 391   #if VECT_SIZE >= 16
 392   asm ("prmt.b32 %0, %1, %2, %3;" : "=r"(r.s8) : "r"(a.s8), "r"(b.s8), "r"(c.s8));
 393   asm ("prmt.b32 %0, %1, %2, %3;" : "=r"(r.s9) : "r"(a.s9), "r"(b.s9), "r"(c.s9));
 394   asm ("prmt.b32 %0, %1, %2, %3;" : "=r"(r.sa) : "r"(a.sa), "r"(b.sa), "r"(c.sa));
 395   asm ("prmt.b32 %0, %1, %2, %3;" : "=r"(r.sb) : "r"(a.sb), "r"(b.sb), "r"(c.sb));
 396   asm ("prmt.b32 %0, %1, %2, %3;" : "=r"(r.sc) : "r"(a.sc), "r"(b.sc), "r"(c.sc));
 397   asm ("prmt.b32 %0, %1, %2, %3;" : "=r"(r.sd) : "r"(a.sd), "r"(b.sd), "r"(c.sd));
 398   asm ("prmt.b32 %0, %1, %2, %3;" : "=r"(r.se) : "r"(a.se), "r"(b.se), "r"(c.se));
 399   asm ("prmt.b32 %0, %1, %2, %3;" : "=r"(r.sf) : "r"(a.sf), "r"(b.sf), "r"(c.sf));
 400   #endif
 401
 402   return r;
 403 }
 404
 405 inline u32 __byte_perm_S (const u32 a, const u32 b, const u32 c)
 406 {
 407   u32 r;
 408
 409   asm ("prmt.b32 %0, %1, %2, %3;" : "=r"(r) : "r"(a), "r"(b), "r"(c));
 410
 411   return r;
 412 }
 413
 414 inline u32x __bfe (const u32x a, const u32x b, const u32x c)
 415 {
 416   u32x r;
 417
 418   #if VECT_SIZE == 1
 419   asm ("bfe.u32 %0, %1, %2, %3;" : "=r"(r)    : "r"(a),    "r"(b),    "r"(c));
 420   #endif
 421
 422   #if VECT_SIZE >= 2
 423   asm ("bfe.u32 %0, %1, %2, %3;" : "=r"(r.s0) : "r"(a.s0), "r"(b.s0), "r"(c.s0));
 424   asm ("bfe.u32 %0, %1, %2, %3;" : "=r"(r.s1) : "r"(a.s1), "r"(b.s1), "r"(c.s1));
 425   #endif
 426
 427   #if VECT_SIZE >= 4
 428   asm ("bfe.u32 %0, %1, %2, %3;" : "=r"(r.s2) : "r"(a.s2), "r"(b.s2), "r"(c.s2));
 429   asm ("bfe.u32 %0, %1, %2, %3;" : "=r"(r.s3) : "r"(a.s3), "r"(b.s3), "r"(c.s3));
 430   #endif
 431
 432   #if VECT_SIZE >= 8
 433   asm ("bfe.u32 %0, %1, %2, %3;" : "=r"(r.s4) : "r"(a.s4), "r"(b.s4), "r"(c.s4));
 434   asm ("bfe.u32 %0, %1, %2, %3;" : "=r"(r.s5) : "r"(a.s5), "r"(b.s5), "r"(c.s5));
 435   asm ("bfe.u32 %0, %1, %2, %3;" : "=r"(r.s6) : "r"(a.s6), "r"(b.s6), "r"(c.s6));
 436   asm ("bfe.u32 %0, %1, %2, %3;" : "=r"(r.s7) : "r"(a.s7), "r"(b.s7), "r"(c.s7));
 437   #endif
 438
 439   #if VECT_SIZE >= 16
 440   asm ("bfe.u32 %0, %1, %2, %3;" : "=r"(r.s8) : "r"(a.s8), "r"(b.s8), "r"(c.s8));
 441   asm ("bfe.u32 %0, %1, %2, %3;" : "=r"(r.s9) : "r"(a.s9), "r"(b.s9), "r"(c.s9));
 442   asm ("bfe.u32 %0, %1, %2, %3;" : "=r"(r.sa) : "r"(a.sa), "r"(b.sa), "r"(c.sa));
 443   asm ("bfe.u32 %0, %1, %2, %3;" : "=r"(r.sb) : "r"(a.sb), "r"(b.sb), "r"(c.sb));
 444   asm ("bfe.u32 %0, %1, %2, %3;" : "=r"(r.sc) : "r"(a.sc), "r"(b.sc), "r"(c.sc));
 445   asm ("bfe.u32 %0, %1, %2, %3;" : "=r"(r.sd) : "r"(a.sd), "r"(b.sd), "r"(c.sd));
 446   asm ("bfe.u32 %0, %1, %2, %3;" : "=r"(r.se) : "r"(a.se), "r"(b.se), "r"(c.se));
 447   asm ("bfe.u32 %0, %1, %2, %3;" : "=r"(r.sf) : "r"(a.sf), "r"(b.sf), "r"(c.sf));
 448   #endif
 449
 450   return r;
 451 }
 452
 453 inline u32 __bfe_S (const u32 a, const u32 b, const u32 c)
 454 {
 455   u32 r;
 456
 457   asm ("bfe.u32 %0, %1, %2, %3;" : "=r"(r) : "r"(a), "r"(b), "r"(c));
 458
 459   return r;
 460 }
 461
 462 inline u32x amd_bytealign (const u32x a, const u32x b, const u32x c)
 463 {
 464   u32x r;
 465
 466   #if CUDA_ARCH >= 350
 467
 468   #if VECT_SIZE == 1
 469   asm ("shf.r.wrap.b32 %0, %1, %2, %3;" : "=r"(r)    : "r"(b),    "r"(a),    "r"((c & 3) * 8));
 470   #endif
 471
 472   #if VECT_SIZE >= 2
 473   asm ("shf.r.wrap.b32 %0, %1, %2, %3;" : "=r"(r.s0) : "r"(b.s0), "r"(a.s0), "r"((c.s0 & 3) * 8));
 474   asm ("shf.r.wrap.b32 %0, %1, %2, %3;" : "=r"(r.s1) : "r"(b.s1), "r"(a.s1), "r"((c.s1 & 3) * 8));
 475   #endif
 476
 477   #if VECT_SIZE >= 4
 478   asm ("shf.r.wrap.b32 %0, %1, %2, %3;" : "=r"(r.s2) : "r"(b.s2), "r"(a.s2), "r"((c.s2 & 3) * 8));
 479   asm ("shf.r.wrap.b32 %0, %1, %2, %3;" : "=r"(r.s3) : "r"(b.s3), "r"(a.s3), "r"((c.s3 & 3) * 8));
 480   #endif
 481
 482   #if VECT_SIZE >= 8
 483   asm ("shf.r.wrap.b32 %0, %1, %2, %3;" : "=r"(r.s4) : "r"(b.s4), "r"(a.s4), "r"((c.s4 & 3) * 8));
 484   asm ("shf.r.wrap.b32 %0, %1, %2, %3;" : "=r"(r.s5) : "r"(b.s5), "r"(a.s5), "r"((c.s5 & 3) * 8));
 485   asm ("shf.r.wrap.b32 %0, %1, %2, %3;" : "=r"(r.s6) : "r"(b.s6), "r"(a.s6), "r"((c.s6 & 3) * 8));
 486   asm ("shf.r.wrap.b32 %0, %1, %2, %3;" : "=r"(r.s7) : "r"(b.s7), "r"(a.s7), "r"((c.s7 & 3) * 8));
 487   #endif
 488
 489   #if VECT_SIZE >= 16
 490   asm ("shf.r.wrap.b32 %0, %1, %2, %3;" : "=r"(r.s8) : "r"(b.s8), "r"(a.s8), "r"((c.s8 & 3) * 8));
 491   asm ("shf.r.wrap.b32 %0, %1, %2, %3;" : "=r"(r.s9) : "r"(b.s9), "r"(a.s9), "r"((c.s9 & 3) * 8));
 492   asm ("shf.r.wrap.b32 %0, %1, %2, %3;" : "=r"(r.sa) : "r"(b.sa), "r"(a.sa), "r"((c.sa & 3) * 8));
 493   asm ("shf.r.wrap.b32 %0, %1, %2, %3;" : "=r"(r.sb) : "r"(b.sb), "r"(a.sb), "r"((c.sb & 3) * 8));
 494   asm ("shf.r.wrap.b32 %0, %1, %2, %3;" : "=r"(r.sc) : "r"(b.sc), "r"(a.sc), "r"((c.sc & 3) * 8));
 495   asm ("shf.r.wrap.b32 %0, %1, %2, %3;" : "=r"(r.sd) : "r"(b.sd), "r"(a.sd), "r"((c.sd & 3) * 8));
 496   asm ("shf.r.wrap.b32 %0, %1, %2, %3;" : "=r"(r.se) : "r"(b.se), "r"(a.se), "r"((c.se & 3) * 8));
 497   asm ("shf.r.wrap.b32 %0, %1, %2, %3;" : "=r"(r.sf) : "r"(b.sf), "r"(a.sf), "r"((c.sf & 3) * 8));
 498   #endif
 499
 500   #else
 501
 502   r = __byte_perm (b, a, ((u32x) (0x76543210) >> ((c & 3) * 4)) & 0xffff);
 503
 504   #endif
 505
 506   return r;
 507 }
 508
 509 inline u32 amd_bytealign_S (const u32 a, const u32 b, const u32 c)
 510 {
 511   u32 r;
 512
 513   #if CUDA_ARCH >= 350
 514
 515   asm ("shf.r.wrap.b32 %0, %1, %2, %3;" : "=r"(r) : "r"(b), "r"(a), "r"((c & 3) * 8));
 516
 517   #else
 518
 519   r = __byte_perm_S (b, a, (0x76543210 >> ((c & 3) * 4)) & 0xffff);
 520
 521   #endif
 522
 523   return r;
 524 }
 525 #endif
 526
 527 #ifdef IS_GENERIC
 528 inline u32 swap32_S (const u32 v)
 529 {
 530   return (as_uint (as_uchar4 (v).s3210));
 531 }
 532
 533 inline u64 swap64_S (const u64 v)
 534 {
 535   return (as_ulong (as_uchar8 (v).s76543210));
 536 }
 537
 538 inline u32 rotr32_S (const u32 a, const u32 n)
 539 {
 540   return rotate (a, 32 - n);
 541 }
 542
 543 inline u32 rotl32_S (const u32 a, const u32 n)
 544 {
 545   return rotate (a, n);
 546 }
 547
 548 inline u64 rotr64_S (const u64 a, const u32 n)
 549 {
 550   return rotate (a, (u64) 64 - n);
 551 }
 552
 553 inline u64 rotl64_S (const u64 a, const u32 n)
 554 {
 555   return rotate (a, (u64) n);
 556 }
 557
 558 inline u32x swap32 (const u32x v)
 559 {
 560   return ((v >> 24) & 0x000000ff)
 561        | ((v >>  8) & 0x0000ff00)
 562        | ((v <<  8) & 0x00ff0000)
 563        | ((v << 24) & 0xff000000);
 564 }
 565
 566 inline u64x swap64 (const u64x v)
 567 {
 568   return ((v >> 56) & 0x00000000000000ff)
 569        | ((v >> 40) & 0x000000000000ff00)
 570        | ((v >> 24) & 0x0000000000ff0000)
 571        | ((v >>  8) & 0x00000000ff000000)
 572        | ((v <<  8) & 0x000000ff00000000)
 573        | ((v << 24) & 0x0000ff0000000000)
 574        | ((v << 40) & 0x00ff000000000000)
 575        | ((v << 56) & 0xff00000000000000);
 576 }
 577
 578 inline u32x rotr32 (const u32x a, const u32 n)
 579 {
 580   return rotate (a, 32 - n);
 581 }
 582
 583 inline u32x rotl32 (const u32x a, const u32 n)
 584 {
 585   return rotate (a, n);
 586 }
 587
 588 inline u64x rotr64 (const u64x a, const u32 n)
 589 {
 590   return rotate (a, (u64) 64 - n);
 591 }
 592
 593 inline u64x rotl64 (const u64x a, const u32 n)
 594 {
 595   return rotate (a, (u64) n);
 596 }
 597
 598 inline u32x __bfe (const u32x a, const u32x b, const u32x c)
 599 {
 600   #define BIT(x)      ((u32x) (1u) << (x))
 601   #define BIT_MASK(x) (BIT (x) - 1)
 602   #define BFE(x,y,z)  (((x) >> (y)) & BIT_MASK (z))
 603
 604   return BFE (a, b, c);
 605
 606   #undef BIT
 607   #undef BIT_MASK
 608   #undef BFE
 609 }
 610
 611 inline u32 __bfe_S (const u32 a, const u32 b, const u32 c)
 612 {
 613   #define BIT(x)      (1u << (x))
 614   #define BIT_MASK(x) (BIT (x) - 1)
 615   #define BFE(x,y,z)  (((x) >> (y)) & BIT_MASK (z))
 616
 617   return BFE (a, b, c);
 618
 619   #undef BIT
 620   #undef BIT_MASK
 621   #undef BFE
 622 }
 623
 624 inline u32x amd_bytealign (const u32x a, const u32x b, const u32 c)
 625 {
 626   #if VECT_SIZE == 1
 627   const u64x tmp = ((((u64x) (a)) << 32) | ((u64x) (b))) >> ((c & 3) * 8);
 628
 629   return (u32x) (tmp);
 630   #endif
 631
 632   #if VECT_SIZE == 2
 633   const u64x tmp = ((((u64x) (a.s0, a.s1)) << 32) | ((u64x) (b.s0, b.s1))) >> ((c & 3) * 8);
 634
 635   return (u32x) (tmp.s0, tmp.s1);
 636   #endif
 637
 638   #if VECT_SIZE == 4
 639   const u64x tmp = ((((u64x) (a.s0, a.s1, a.s2, a.s3)) << 32) | ((u64x) (b.s0, b.s1, b.s2, b.s3))) >> ((c & 3) * 8);
 640
 641   return (u32x) (tmp.s0, tmp.s1, tmp.s2, tmp.s3);
 642   #endif
 643
 644   #if VECT_SIZE == 8
 645   const u64x tmp = ((((u64x) (a.s0, a.s1, a.s2, a.s3, a.s4, a.s5, a.s6, a.s7)) << 32) | ((u64x) (b.s0, b.s1, b.s2, b.s3, b.s4, b.s5, b.s6, b.s7))) >> ((c & 3) * 8);
 646
 647   return (u32x) (tmp.s0, tmp.s1, tmp.s2, tmp.s3, tmp.s4, tmp.s5, tmp.s6, tmp.s7);
 648   #endif
 649
 650   #if VECT_SIZE == 16
 651   const u64x tmp = ((((u64x) (a.s0, a.s1, a.s2, a.s3, a.s4, a.s5, a.s6, a.s7, a.s8, a.s9, a.sa, a.sb, a.sc, a.sd, a.se, a.sf)) << 32) | ((u64x) (b.s0, b.s1, b.s2, b.s3, b.s4, b.s5, b.s6, b.s7, b.s8, b.s9, b.sa, b.sb, b.sc, b.sd, b.se, b.sf))) >> ((c & 3) * 8);
 652
 653   return (u32x) (tmp.s0, tmp.s1, tmp.s2, tmp.s3, tmp.s4, tmp.s5, tmp.s6, tmp.s7, tmp.s8, tmp.s9, tmp.sa, tmp.sb, tmp.sc, tmp.sd, tmp.se, tmp.sf);
 654   #endif
 655 }
 656
 657 inline u32 amd_bytealign_S (const u32 a, const u32 b, const u32 c)
 658 {
 659   const u64 tmp = ((((u64) a) << 32) | ((u64) b)) >> ((c & 3) * 8);
 660
 661   return (u32) (tmp);
 662 }
 663
 664 #endif
 665
 666 typedef struct
 667 {
 668   #if   defined _DES_
 669   u32  digest_buf[4];
 670   #elif defined _MD4_
 671   u32  digest_buf[4];
 672   #elif defined _MD5_
 673   u32  digest_buf[4];
 674   #elif defined _MD5H_
 675   u32  digest_buf[4];
 676   #elif defined _SHA1_
 677   u32  digest_buf[5];
 678   #elif defined _BCRYPT_
 679   u32  digest_buf[6];
 680   #elif defined _SHA256_
 681   u32  digest_buf[8];
 682   #elif defined _SHA384_
 683   u32  digest_buf[16];
 684   #elif defined _SHA512_
 685   u32  digest_buf[16];
 686   #elif defined _KECCAK_
 687   u32  digest_buf[50];
 688   #elif defined _RIPEMD160_
 689   u32  digest_buf[5];
 690   #elif defined _WHIRLPOOL_
 691   u32  digest_buf[16];
 692   #elif defined _GOST_
 693   u32  digest_buf[8];
 694   #elif defined _GOST2012_256_
 695   u32  digest_buf[8];
 696   #elif defined _GOST2012_512_
 697   u32  digest_buf[16];
 698   #elif defined _SAPB_
 699   u32  digest_buf[4];
 700   #elif defined _SAPG_
 701   u32  digest_buf[5];
 702   #elif defined _MYSQL323_
 703   u32  digest_buf[4];
 704   #elif defined _LOTUS5_
 705   u32  digest_buf[4];
 706   #elif defined _LOTUS6_
 707   u32  digest_buf[4];
 708   #elif defined _SCRYPT_
 709   u32  digest_buf[8];
 710   #elif defined _LOTUS8_
 711   u32  digest_buf[4];
 712   #elif defined _OFFICE2007_
 713   u32  digest_buf[4];
 714   #elif defined _OFFICE2010_
 715   u32  digest_buf[4];
 716   #elif defined _OFFICE2013_
 717   u32  digest_buf[4];
 718   #elif defined _OLDOFFICE01_
 719   u32  digest_buf[4];
 720   #elif defined _OLDOFFICE34_
 721   u32  digest_buf[4];
 722   #elif defined _SIPHASH_
 723   u32  digest_buf[4];
 724   #elif defined _PBKDF2_MD5_
 725   u32  digest_buf[32];
 726   #elif defined _PBKDF2_SHA1_
 727   u32  digest_buf[32];
 728   #elif defined _PBKDF2_SHA256_
 729   u32  digest_buf[32];
 730   #elif defined _PBKDF2_SHA512_
 731   u32  digest_buf[32];
 732   #elif defined _PDF17L8_
 733   u32  digest_buf[8];
 734   #elif defined _CRC32_
 735   u32  digest_buf[4];
 736   #elif defined _SEVEN_ZIP_
 737   u32  digest_buf[4];
 738   #elif defined _ANDROIDFDE_
 739   u32  digest_buf[4];
 740   #elif defined _DCC2_
 741   u32  digest_buf[4];
 742   #elif defined _WPA_
 743   u32  digest_buf[4];
 744   #elif defined _MD5_SHA1_
 745   u32  digest_buf[4];
 746   #elif defined _SHA1_MD5_
 747   u32  digest_buf[5];
 748   #elif defined _NETNTLMV2_
 749   u32  digest_buf[4];
 750   #elif defined _KRB5PA_
 751   u32  digest_buf[4];
 752   #elif defined _CLOUDKEY_
 753   u32  digest_buf[8];
 754   #elif defined _SCRYPT_
 755   u32  digest_buf[4];
 756   #elif defined _PSAFE2_
 757   u32  digest_buf[5];
 758   #elif defined _LOTUS8_
 759   u32  digest_buf[4];
 760   #elif defined _RAR3_
 761   u32  digest_buf[4];
 762   #elif defined _SHA256_SHA1_
 763   u32  digest_buf[8];
 764   #elif defined _MS_DRSR_
 765   u32  digest_buf[8];
 766   #elif defined _ANDROIDFDE_SAMSUNG_
 767   u32  digest_buf[8];
 768   #elif defined _RAR5_
 769   u32  digest_buf[4];
 770   #elif defined _KRB5TGS_
 771   u32  digest_buf[4];
 772   #elif defined _AXCRYPT_
 773   u32  digest_buf[4];
 774   #elif defined _KEEPASS_
 775   u32  digest_buf[4];
 776   #elif defined _ZIP2_
 777   u32  digest_buf[4];
 778   #endif
 779
 780 } digest_t;
 781
 782 typedef struct
 783 {
 784   u32 salt_buf[16];
 785   u32 salt_buf_pc[8];
 786
 787   u32 salt_len;
 788   u32 salt_iter;
 789   u32 salt_sign[2];
 790
 791   u32 keccak_mdlen;
 792   u32 truecrypt_mdlen;
 793
 794   u32 digests_cnt;
 795   u32 digests_done;
 796
 797   u32 digests_offset;
 798
 799   u32 scrypt_N;
 800   u32 scrypt_r;
 801   u32 scrypt_p;
 802   u32 scrypt_tmto;
 803   u32 scrypt_phy;
 804
 805 } salt_t;
 806
 807 typedef struct
 808 {
 809   int V;
 810   int R;
 811   int P;
 812
 813   int enc_md;
 814
 815   u32 id_buf[8];
 816   u32 u_buf[32];
 817   u32 o_buf[32];
 818
 819   int id_len;
 820   int o_len;
 821   int u_len;
 822
 823   u32 rc4key[2];
 824   u32 rc4data[2];
 825
 826 } pdf_t;
 827
 828 typedef struct
 829 {
 830   u32 pke[25];
 831   u32 eapol[64];
 832   int eapol_size;
 833   int keyver;
 834   u8  orig_mac1[6];
 835   u8  orig_mac2[6];
 836   u8  orig_nonce1[32];
 837   u8  orig_nonce2[32];
 838
 839 } wpa_t;
 840
 841 typedef struct
 842 {
 843   u32 cry_master_buf[64];
 844   u32 ckey_buf[64];
 845   u32 public_key_buf[64];
 846
 847   u32 cry_master_len;
 848   u32 ckey_len;
 849   u32 public_key_len;
 850
 851 } bitcoin_wallet_t;
 852
 853 typedef struct
 854 {
 855   u32 salt_buf[30];
 856   u32 salt_len;
 857
 858   u32 esalt_buf[38];
 859   u32 esalt_len;
 860
 861 } sip_t;
 862
 863 typedef struct
 864 {
 865   u32 data[384];
 866
 867 } androidfde_t;
 868
 869 typedef struct
 870 {
 871   u32 nr_buf[16];
 872   u32 nr_len;
 873
 874   u32 msg_buf[128];
 875   u32 msg_len;
 876
 877 } ikepsk_t;
 878
 879 typedef struct
 880 {
 881   u32 user_len;
 882   u32 domain_len;
 883   u32 srvchall_len;
 884   u32 clichall_len;
 885
 886   u32 userdomain_buf[64];
 887   u32 chall_buf[256];
 888
 889 } netntlm_t;
 890
 891 typedef struct
 892 {
 893   u32 user[16];
 894   u32 realm[16];
 895   u32 salt[32];
 896   u32 timestamp[16];
 897   u32 checksum[4];
 898
 899 } krb5pa_t;
 900
 901 typedef struct
 902 {
 903   u32 account_info[512];
 904   u32 checksum[4];
 905   u32 edata2[2560];
 906   u32 edata2_len;
 907
 908 } krb5tgs_t;
 909
 910 typedef struct
 911 {
 912   u32 salt_buf[16];
 913   u32 data_buf[112];
 914   u32 keyfile_buf[16];
 915   u32 signature;
 916
 917 } tc_t;
 918
 919 typedef struct
 920 {
 921   u32 salt_buf[16];
 922
 923 } pbkdf2_md5_t;
 924
 925 typedef struct
 926 {
 927   u32 salt_buf[16];
 928
 929 } pbkdf2_sha1_t;
 930
 931 typedef struct
 932 {
 933   u32 salt_buf[16];
 934
 935 } pbkdf2_sha256_t;
 936
 937 typedef struct
 938 {
 939   u32 salt_buf[32];
 940
 941 } pbkdf2_sha512_t;
 942
 943 typedef struct
 944 {
 945   u32 salt_buf[128];
 946   u32 salt_len;
 947
 948 } rakp_t;
 949
 950 typedef struct
 951 {
 952   u32 data_len;
 953   u32 data_buf[512];
 954
 955 } cloudkey_t;
 956
 957 typedef struct
 958 {
 959   u32 encryptedVerifier[4];
 960   u32 encryptedVerifierHash[5];
 961
 962   u32 keySize;
 963
 964 } office2007_t;
 965
 966 typedef struct
 967 {
 968   u32 encryptedVerifier[4];
 969   u32 encryptedVerifierHash[8];
 970
 971 } office2010_t;
 972
 973 typedef struct
 974 {
 975   u32 encryptedVerifier[4];
 976   u32 encryptedVerifierHash[8];
 977
 978 } office2013_t;
 979
 980 typedef struct
 981 {
 982   u32 version;
 983   u32 encryptedVerifier[4];
 984   u32 encryptedVerifierHash[4];
 985   u32 rc4key[2];
 986
 987 } oldoffice01_t;
 988
 989 typedef struct
 990 {
 991   u32 version;
 992   u32 encryptedVerifier[4];
 993   u32 encryptedVerifierHash[5];
 994   u32 rc4key[2];
 995
 996 } oldoffice34_t;
 997
 998 typedef struct
 999 {
1000   u32 salt_buf[128];
1001   u32 salt_len;
1002
1003   u32 pc_digest[5];
1004   u32 pc_offset;
1005
1006 } pstoken_t;
1007
1008 typedef struct
1009 {
1010   u32 type;
1011   u32 mode;
1012   u32 magic;
1013   u32 salt_len;
1014   u32 salt_buf[4];
1015   u32 verify_bytes;
1016   u32 compress_length;
1017   u32 data_len;
1018   u32 data_buf[2048];
1019   u32 auth_len;
1020   u32 auth_buf[4];
1021
1022 } zip2_t;
1023
1024 typedef struct
1025 {
1026   u32 version;
1027   u32 algorithm;
1028
1029   /* key-file handling */
1030   u32 keyfile_len;
1031   u32 keyfile[8];
1032
1033   u32 final_random_seed[8];
1034   u32 transf_random_seed[8];
1035   u32 enc_iv[4];
1036   u32 contents_hash[8];
1037
1038   /* specific to version 1 */
1039   u32 contents_len;
1040   u32 contents[75000];
1041
1042   /* specific to version 2 */
1043   u32 expected_bytes[8];
1044
1045 } keepass_t;
1046
1047 typedef struct
1048 {
1049   u32 digest[4];
1050   u32 out[4];
1051
1052 } pdf14_tmp_t;
1053
1054 typedef struct
1055 {
1056   union
1057   {
1058     u32 dgst32[16];
1059     u64 dgst64[8];
1060   };
1061
1062   u32 dgst_len;
1063   u32 W_len;
1064
1065 } pdf17l8_tmp_t;
1066
1067 typedef struct
1068 {
1069   u32 digest_buf[4];
1070
1071 } phpass_tmp_t;
1072
1073 typedef struct
1074 {
1075   u32 digest_buf[4];
1076
1077 } md5crypt_tmp_t;
1078
1079 typedef struct
1080 {
1081   u32 alt_result[8];
1082
1083   u32 p_bytes[4];
1084   u32 s_bytes[4];
1085
1086 } sha256crypt_tmp_t;
1087
1088 typedef struct
1089 {
1090   u64 l_alt_result[8];
1091
1092   u64 l_p_bytes[2];
1093   u64 l_s_bytes[2];
1094
1095 } sha512crypt_tmp_t;
1096
1097 typedef struct
1098 {
1099   u32 ipad[5];
1100   u32 opad[5];
1101
1102   u32 dgst[10];
1103   u32 out[10];
1104
1105 } wpa_tmp_t;
1106
1107 typedef struct
1108 {
1109   u64 dgst[8];
1110
1111 } bitcoin_wallet_tmp_t;
1112
1113 typedef struct
1114 {
1115   u32 ipad[5];
1116   u32 opad[5];
1117
1118   u32 dgst[5];
1119   u32 out[4];
1120
1121 } dcc2_tmp_t;
1122
1123 typedef struct
1124 {
1125   u32 E[18];
1126
1127   u32 P[18];
1128
1129   u32 S0[256];
1130   u32 S1[256];
1131   u32 S2[256];
1132   u32 S3[256];
1133
1134 } bcrypt_tmp_t;
1135
1136 typedef struct
1137 {
1138   u32 digest[2];
1139
1140   u32 P[18];
1141
1142   u32 S0[256];
1143   u32 S1[256];
1144   u32 S2[256];
1145   u32 S3[256];
1146
1147 } pwsafe2_tmp_t;
1148
1149 typedef struct
1150 {
1151   u32 digest_buf[8];
1152
1153 } pwsafe3_tmp_t;
1154
1155 typedef struct
1156 {
1157   u32 digest_buf[5];
1158
1159 } androidpin_tmp_t;
1160
1161 typedef struct
1162 {
1163   u32 ipad[5];
1164   u32 opad[5];
1165
1166   u32 dgst[10];
1167   u32 out[10];
1168
1169 } androidfde_tmp_t;
1170
1171 typedef struct
1172 {
1173   u32 ipad[16];
1174   u32 opad[16];
1175
1176   u32 dgst[64];
1177   u32 out[64];
1178
1179 } tc_tmp_t;
1180
1181 typedef struct
1182 {
1183   u64 ipad[8];
1184   u64 opad[8];
1185
1186   u64 dgst[32];
1187   u64 out[32];
1188
1189 } tc64_tmp_t;
1190
1191 typedef struct
1192 {
1193   u32 ipad[4];
1194   u32 opad[4];
1195
1196   u32 dgst[32];
1197   u32 out[32];
1198
1199 } pbkdf2_md5_tmp_t;
1200
1201 typedef struct
1202 {
1203   u32 ipad[5];
1204   u32 opad[5];
1205
1206   u32 dgst[32];
1207   u32 out[32];
1208
1209 } pbkdf2_sha1_tmp_t;
1210
1211 typedef struct
1212 {
1213   u32 ipad[8];
1214   u32 opad[8];
1215
1216   u32 dgst[32];
1217   u32 out[32];
1218
1219 } pbkdf2_sha256_tmp_t;
1220
1221 typedef struct
1222 {
1223   u64 ipad[8];
1224   u64 opad[8];
1225
1226   u64 dgst[16];
1227   u64 out[16];
1228
1229 } pbkdf2_sha512_tmp_t;
1230
1231 typedef struct
1232 {
1233   u64 out[8];
1234
1235 } ecryptfs_tmp_t;
1236
1237 typedef struct
1238 {
1239   u64 ipad[8];
1240   u64 opad[8];
1241
1242   u64 dgst[16];
1243   u64 out[16];
1244
1245 } oraclet_tmp_t;
1246
1247 typedef struct
1248 {
1249   u32 ipad[5];
1250   u32 opad[5];
1251
1252   u32 dgst[5];
1253   u32 out[5];
1254
1255 } agilekey_tmp_t;
1256
1257 typedef struct
1258 {
1259   u32 ipad[5];
1260   u32 opad[5];
1261
1262   u32 dgst1[5];
1263   u32 out1[5];
1264
1265   u32 dgst2[5];
1266   u32 out2[5];
1267
1268 } mywallet_tmp_t;
1269
1270 typedef struct
1271 {
1272   u32 ipad[5];
1273   u32 opad[5];
1274
1275   u32 dgst[5];
1276   u32 out[5];
1277
1278 } sha1aix_tmp_t;
1279
1280 typedef struct
1281 {
1282   u32 ipad[8];
1283   u32 opad[8];
1284
1285   u32 dgst[8];
1286   u32 out[8];
1287
1288 } sha256aix_tmp_t;
1289
1290 typedef struct
1291 {
1292   u64 ipad[8];
1293   u64 opad[8];
1294
1295   u64 dgst[8];
1296   u64 out[8];
1297
1298 } sha512aix_tmp_t;
1299
1300 typedef struct
1301 {
1302   u32 ipad[8];
1303   u32 opad[8];
1304
1305   u32 dgst[8];
1306   u32 out[8];
1307
1308 } lastpass_tmp_t;
1309
1310 typedef struct
1311 {
1312   u64 digest_buf[8];
1313
1314 } drupal7_tmp_t;
1315
1316 typedef struct
1317 {
1318   u32 ipad[5];
1319   u32 opad[5];
1320
1321   u32 dgst[5];
1322   u32 out[5];
1323
1324 } lotus8_tmp_t;
1325
1326 typedef struct
1327 {
1328   u32 out[5];
1329
1330 } office2007_tmp_t;
1331
1332 typedef struct
1333 {
1334   u32 out[5];
1335
1336 } office2010_tmp_t;
1337
1338 typedef struct
1339 {
1340   u64 out[8];
1341
1342 } office2013_tmp_t;
1343
1344 typedef struct
1345 {
1346   u32 digest_buf[5];
1347
1348 } saph_sha1_tmp_t;
1349
1350 typedef struct
1351 {
1352   u32 block[16];
1353
1354   u32 dgst[8];
1355
1356   u32 block_len;
1357   u32 final_len;
1358
1359 } seven_zip_tmp_t;
1360
1361 typedef struct
1362 {
1363   u32 KEK[5];
1364
1365   u32 lsb[4];
1366   u32 cipher[4];
1367
1368 } axcrypt_tmp_t;
1369
1370 typedef struct
1371 {
1372   u32 tmp_digest[8];
1373
1374 } keepass_tmp_t;
1375
1376 typedef struct
1377 {
1378   u32 Kc[16];
1379   u32 Kd[16];
1380
1381   u32 iv[2];
1382
1383 } bsdicrypt_tmp_t;
1384
1385 typedef struct
1386 {
1387   u32 dgst[17][5];
1388
1389 } rar3_tmp_t;
1390
1391 typedef struct
1392 {
1393   u32 user[16];
1394
1395 } cram_md5_t;
1396
1397 typedef struct
1398 {
1399   u32 iv_buf[4];
1400   u32 iv_len;
1401
1402   u32 salt_buf[4];
1403   u32 salt_len;
1404
1405   u32 crc;
1406
1407   u32 data_buf[96];
1408   u32 data_len;
1409
1410   u32 unpack_size;
1411
1412 } seven_zip_t;
1413
1414 typedef struct
1415 {
1416   u32 key;
1417   u64 val;
1418
1419 } hcstat_table_t;
1420
1421 typedef struct
1422 {
1423   u32 cs_buf[0x100];
1424   u32 cs_len;
1425
1426 } cs_t;
1427
1428 typedef struct
1429 {
1430   u32 cmds[0x100];
1431
1432 } kernel_rule_t;
1433
1434 typedef struct
1435 {
1436   u32 salt_pos;
1437   u32 digest_pos;
1438   u32 hash_pos;
1439   u32 gidvid;
1440   u32 il_pos;
1441
1442 } plain_t;
1443
1444 typedef struct
1445 {
1446   u32 i[16];
1447
1448   u32 pw_len;
1449
1450   u32 alignment_placeholder_1;
1451   u32 alignment_placeholder_2;
1452   u32 alignment_placeholder_3;
1453
1454 } pw_t;
1455
1456 typedef struct
1457 {
1458   u32 i;
1459
1460 } bf_t;
1461
1462 typedef struct
1463 {
1464   u32 i[8];
1465
1466   u32 pw_len;
1467
1468 } comb_t;
1469
1470 typedef struct
1471 {
1472   u32 b[32];
1473
1474 } bs_word_t;
1475
1476 typedef struct
1477 {
1478   uint4 P[64];
1479
1480 } scrypt_tmp_t;