gcc/config/arm/arm_neon.h

   1 /* LLVM LOCAL file Changed to use preprocessor macros.  */
   2 /* APPLE LOCAL file v7 support. Merge from Codesourcery */
   3 /* ARM NEON intrinsics include file. This file is generated automatically
   4    using neon-gen.ml.  Please do not edit manually.
   5
   6    Copyright (C) 2006, 2007 Free Software Foundation, Inc.
   7    Contributed by CodeSourcery.
   8
   9    This file is part of GCC.
  10
  11    GCC is free software; you can redistribute it and/or modify it
  12    under the terms of the GNU General Public License as published
  13    by the Free Software Foundation; either version 2, or (at your
  14    option) any later version.
  15
  16    GCC is distributed in the hope that it will be useful, but WITHOUT
  17    ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
  18    or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public
  19    License for more details.
  20
  21    You should have received a copy of the GNU General Public License
  22    along with GCC; see the file COPYING.  If not, write to the
  23    Free Software Foundation, 51 Franklin Street, Fifth Floor, Boston,
  24    MA 02110-1301, USA.  */
  25
  26 /* As a special exception, if you include this header file into source
  27    files compiled by GCC, this header file does not by itself cause
  28    the resulting executable to be covered by the GNU General Public
  29    License.  This exception does not however invalidate any other
  30    reasons why the executable file might be covered by the GNU General
  31    Public License.  */
  32
  33 #ifndef _GCC_ARM_NEON_H
  34 #define _GCC_ARM_NEON_H 1
  35
  36 #ifndef __ARM_NEON__
  37 #error You must enable NEON instructions (e.g. -mfloat-abi=softfp -mfpu=neon) to use arm_neon.h
  38 #else
  39
  40 #ifdef __cplusplus
  41 extern "C" {
  42 #endif
  43
  44 #include <stdint.h>
  45
  46 typedef __builtin_neon_qi int8x8_t      __attribute__ ((__vector_size__ (8)));
  47 typedef __builtin_neon_hi int16x4_t     __attribute__ ((__vector_size__ (8)));
  48 typedef __builtin_neon_si int32x2_t     __attribute__ ((__vector_size__ (8)));
  49 typedef __builtin_neon_di int64x1_t     __attribute__ ((__vector_size__ (8)));
  50 typedef __builtin_neon_sf float32x2_t   __attribute__ ((__vector_size__ (8)));
  51 typedef __builtin_neon_poly8 poly8x8_t  __attribute__ ((__vector_size__ (8)));
  52 typedef __builtin_neon_poly16 poly16x4_t        __attribute__ ((__vector_size__ (8)));
  53 typedef __builtin_neon_uqi uint8x8_t    __attribute__ ((__vector_size__ (8)));
  54 typedef __builtin_neon_uhi uint16x4_t   __attribute__ ((__vector_size__ (8)));
  55 typedef __builtin_neon_usi uint32x2_t   __attribute__ ((__vector_size__ (8)));
  56 typedef __builtin_neon_udi uint64x1_t   __attribute__ ((__vector_size__ (8)));
  57 typedef __builtin_neon_qi int8x16_t     __attribute__ ((__vector_size__ (16)));
  58 typedef __builtin_neon_hi int16x8_t     __attribute__ ((__vector_size__ (16)));
  59 typedef __builtin_neon_si int32x4_t     __attribute__ ((__vector_size__ (16)));
  60 typedef __builtin_neon_di int64x2_t     __attribute__ ((__vector_size__ (16)));
  61 typedef __builtin_neon_sf float32x4_t   __attribute__ ((__vector_size__ (16)));
  62 typedef __builtin_neon_poly8 poly8x16_t __attribute__ ((__vector_size__ (16)));
  63 typedef __builtin_neon_poly16 poly16x8_t        __attribute__ ((__vector_size__ (16)));
  64 typedef __builtin_neon_uqi uint8x16_t   __attribute__ ((__vector_size__ (16)));
  65 typedef __builtin_neon_uhi uint16x8_t   __attribute__ ((__vector_size__ (16)));
  66 typedef __builtin_neon_usi uint32x4_t   __attribute__ ((__vector_size__ (16)));
  67 typedef __builtin_neon_udi uint64x2_t   __attribute__ ((__vector_size__ (16)));
  68
  69 typedef __builtin_neon_sf float32_t;
  70 typedef __builtin_neon_poly8 poly8_t;
  71 typedef __builtin_neon_poly16 poly16_t;
  72
  73 typedef struct int8x8x2_t
  74 {
  75   int8x8_t val[2];
  76 } int8x8x2_t;
  77
  78 typedef struct int8x16x2_t
  79 {
  80   int8x16_t val[2];
  81 } int8x16x2_t;
  82
  83 typedef struct int16x4x2_t
  84 {
  85   int16x4_t val[2];
  86 } int16x4x2_t;
  87
  88 typedef struct int16x8x2_t
  89 {
  90   int16x8_t val[2];
  91 } int16x8x2_t;
  92
  93 typedef struct int32x2x2_t
  94 {
  95   int32x2_t val[2];
  96 } int32x2x2_t;
  97
  98 typedef struct int32x4x2_t
  99 {
 100   int32x4_t val[2];
 101 } int32x4x2_t;
 102
 103 typedef struct int64x1x2_t
 104 {
 105   int64x1_t val[2];
 106 } int64x1x2_t;
 107
 108 typedef struct int64x2x2_t
 109 {
 110   int64x2_t val[2];
 111 } int64x2x2_t;
 112
 113 typedef struct uint8x8x2_t
 114 {
 115   uint8x8_t val[2];
 116 } uint8x8x2_t;
 117
 118 typedef struct uint8x16x2_t
 119 {
 120   uint8x16_t val[2];
 121 } uint8x16x2_t;
 122
 123 typedef struct uint16x4x2_t
 124 {
 125   uint16x4_t val[2];
 126 } uint16x4x2_t;
 127
 128 typedef struct uint16x8x2_t
 129 {
 130   uint16x8_t val[2];
 131 } uint16x8x2_t;
 132
 133 typedef struct uint32x2x2_t
 134 {
 135   uint32x2_t val[2];
 136 } uint32x2x2_t;
 137
 138 typedef struct uint32x4x2_t
 139 {
 140   uint32x4_t val[2];
 141 } uint32x4x2_t;
 142
 143 typedef struct uint64x1x2_t
 144 {
 145   uint64x1_t val[2];
 146 } uint64x1x2_t;
 147
 148 typedef struct uint64x2x2_t
 149 {
 150   uint64x2_t val[2];
 151 } uint64x2x2_t;
 152
 153 typedef struct float32x2x2_t
 154 {
 155   float32x2_t val[2];
 156 } float32x2x2_t;
 157
 158 typedef struct float32x4x2_t
 159 {
 160   float32x4_t val[2];
 161 } float32x4x2_t;
 162
 163 typedef struct poly8x8x2_t
 164 {
 165   poly8x8_t val[2];
 166 } poly8x8x2_t;
 167
 168 typedef struct poly8x16x2_t
 169 {
 170   poly8x16_t val[2];
 171 } poly8x16x2_t;
 172
 173 typedef struct poly16x4x2_t
 174 {
 175   poly16x4_t val[2];
 176 } poly16x4x2_t;
 177
 178 typedef struct poly16x8x2_t
 179 {
 180   poly16x8_t val[2];
 181 } poly16x8x2_t;
 182
 183 typedef struct int8x8x3_t
 184 {
 185   int8x8_t val[3];
 186 } int8x8x3_t;
 187
 188 typedef struct int8x16x3_t
 189 {
 190   int8x16_t val[3];
 191 } int8x16x3_t;
 192
 193 typedef struct int16x4x3_t
 194 {
 195   int16x4_t val[3];
 196 } int16x4x3_t;
 197
 198 typedef struct int16x8x3_t
 199 {
 200   int16x8_t val[3];
 201 } int16x8x3_t;
 202
 203 typedef struct int32x2x3_t
 204 {
 205   int32x2_t val[3];
 206 } int32x2x3_t;
 207
 208 typedef struct int32x4x3_t
 209 {
 210   int32x4_t val[3];
 211 } int32x4x3_t;
 212
 213 typedef struct int64x1x3_t
 214 {
 215   int64x1_t val[3];
 216 } int64x1x3_t;
 217
 218 typedef struct int64x2x3_t
 219 {
 220   int64x2_t val[3];
 221 } int64x2x3_t;
 222
 223 typedef struct uint8x8x3_t
 224 {
 225   uint8x8_t val[3];
 226 } uint8x8x3_t;
 227
 228 typedef struct uint8x16x3_t
 229 {
 230   uint8x16_t val[3];
 231 } uint8x16x3_t;
 232
 233 typedef struct uint16x4x3_t
 234 {
 235   uint16x4_t val[3];
 236 } uint16x4x3_t;
 237
 238 typedef struct uint16x8x3_t
 239 {
 240   uint16x8_t val[3];
 241 } uint16x8x3_t;
 242
 243 typedef struct uint32x2x3_t
 244 {
 245   uint32x2_t val[3];
 246 } uint32x2x3_t;
 247
 248 typedef struct uint32x4x3_t
 249 {
 250   uint32x4_t val[3];
 251 } uint32x4x3_t;
 252
 253 typedef struct uint64x1x3_t
 254 {
 255   uint64x1_t val[3];
 256 } uint64x1x3_t;
 257
 258 typedef struct uint64x2x3_t
 259 {
 260   uint64x2_t val[3];
 261 } uint64x2x3_t;
 262
 263 typedef struct float32x2x3_t
 264 {
 265   float32x2_t val[3];
 266 } float32x2x3_t;
 267
 268 typedef struct float32x4x3_t
 269 {
 270   float32x4_t val[3];
 271 } float32x4x3_t;
 272
 273 typedef struct poly8x8x3_t
 274 {
 275   poly8x8_t val[3];
 276 } poly8x8x3_t;
 277
 278 typedef struct poly8x16x3_t
 279 {
 280   poly8x16_t val[3];
 281 } poly8x16x3_t;
 282
 283 typedef struct poly16x4x3_t
 284 {
 285   poly16x4_t val[3];
 286 } poly16x4x3_t;
 287
 288 typedef struct poly16x8x3_t
 289 {
 290   poly16x8_t val[3];
 291 } poly16x8x3_t;
 292
 293 typedef struct int8x8x4_t
 294 {
 295   int8x8_t val[4];
 296 } int8x8x4_t;
 297
 298 typedef struct int8x16x4_t
 299 {
 300   int8x16_t val[4];
 301 } int8x16x4_t;
 302
 303 typedef struct int16x4x4_t
 304 {
 305   int16x4_t val[4];
 306 } int16x4x4_t;
 307
 308 typedef struct int16x8x4_t
 309 {
 310   int16x8_t val[4];
 311 } int16x8x4_t;
 312
 313 typedef struct int32x2x4_t
 314 {
 315   int32x2_t val[4];
 316 } int32x2x4_t;
 317
 318 typedef struct int32x4x4_t
 319 {
 320   int32x4_t val[4];
 321 } int32x4x4_t;
 322
 323 typedef struct int64x1x4_t
 324 {
 325   int64x1_t val[4];
 326 } int64x1x4_t;
 327
 328 typedef struct int64x2x4_t
 329 {
 330   int64x2_t val[4];
 331 } int64x2x4_t;
 332
 333 typedef struct uint8x8x4_t
 334 {
 335   uint8x8_t val[4];
 336 } uint8x8x4_t;
 337
 338 typedef struct uint8x16x4_t
 339 {
 340   uint8x16_t val[4];
 341 } uint8x16x4_t;
 342
 343 typedef struct uint16x4x4_t
 344 {
 345   uint16x4_t val[4];
 346 } uint16x4x4_t;
 347
 348 typedef struct uint16x8x4_t
 349 {
 350   uint16x8_t val[4];
 351 } uint16x8x4_t;
 352
 353 typedef struct uint32x2x4_t
 354 {
 355   uint32x2_t val[4];
 356 } uint32x2x4_t;
 357
 358 typedef struct uint32x4x4_t
 359 {
 360   uint32x4_t val[4];
 361 } uint32x4x4_t;
 362
 363 typedef struct uint64x1x4_t
 364 {
 365   uint64x1_t val[4];
 366 } uint64x1x4_t;
 367
 368 typedef struct uint64x2x4_t
 369 {
 370   uint64x2_t val[4];
 371 } uint64x2x4_t;
 372
 373 typedef struct float32x2x4_t
 374 {
 375   float32x2_t val[4];
 376 } float32x2x4_t;
 377
 378 typedef struct float32x4x4_t
 379 {
 380   float32x4_t val[4];
 381 } float32x4x4_t;
 382
 383 typedef struct poly8x8x4_t
 384 {
 385   poly8x8_t val[4];
 386 } poly8x8x4_t;
 387
 388 typedef struct poly8x16x4_t
 389 {
 390   poly8x16_t val[4];
 391 } poly8x16x4_t;
 392
 393 typedef struct poly16x4x4_t
 394 {
 395   poly16x4_t val[4];
 396 } poly16x4x4_t;
 397
 398 typedef struct poly16x8x4_t
 399 {
 400   poly16x8_t val[4];
 401 } poly16x8x4_t;
 402
 403
 404 #define vadd_s8(__a, __b) \
 405   (int8x8_t)__builtin_neon_vaddv8qi (__a, __b, 1)
 406
 407 #define vadd_s16(__a, __b) \
 408   (int16x4_t)__builtin_neon_vaddv4hi (__a, __b, 1)
 409
 410 #define vadd_s32(__a, __b) \
 411   (int32x2_t)__builtin_neon_vaddv2si (__a, __b, 1)
 412
 413 #define vadd_s64(__a, __b) \
 414   (int64x1_t)__builtin_neon_vaddv1di (__a, __b, 1)
 415
 416 #define vadd_f32(__a, __b) \
 417   (float32x2_t)__builtin_neon_vaddv2sf (__a, __b, 5)
 418
 419 #define vadd_u8(__a, __b) \
 420   (uint8x8_t)__builtin_neon_vaddv8qi (__a, __b, 0)
 421
 422 #define vadd_u16(__a, __b) \
 423   (uint16x4_t)__builtin_neon_vaddv4hi (__a, __b, 0)
 424
 425 #define vadd_u32(__a, __b) \
 426   (uint32x2_t)__builtin_neon_vaddv2si (__a, __b, 0)
 427
 428 #define vadd_u64(__a, __b) \
 429   (uint64x1_t)__builtin_neon_vaddv1di (__a, __b, 0)
 430
 431 #define vaddq_s8(__a, __b) \
 432   (int8x16_t)__builtin_neon_vaddv16qi (__a, __b, 1)
 433
 434 #define vaddq_s16(__a, __b) \
 435   (int16x8_t)__builtin_neon_vaddv8hi (__a, __b, 1)
 436
 437 #define vaddq_s32(__a, __b) \
 438   (int32x4_t)__builtin_neon_vaddv4si (__a, __b, 1)
 439
 440 #define vaddq_s64(__a, __b) \
 441   (int64x2_t)__builtin_neon_vaddv2di (__a, __b, 1)
 442
 443 #define vaddq_f32(__a, __b) \
 444   (float32x4_t)__builtin_neon_vaddv4sf (__a, __b, 5)
 445
 446 #define vaddq_u8(__a, __b) \
 447   (uint8x16_t)__builtin_neon_vaddv16qi (__a, __b, 0)
 448
 449 #define vaddq_u16(__a, __b) \
 450   (uint16x8_t)__builtin_neon_vaddv8hi (__a, __b, 0)
 451
 452 #define vaddq_u32(__a, __b) \
 453   (uint32x4_t)__builtin_neon_vaddv4si (__a, __b, 0)
 454
 455 #define vaddq_u64(__a, __b) \
 456   (uint64x2_t)__builtin_neon_vaddv2di (__a, __b, 0)
 457
 458 #define vaddl_s8(__a, __b) \
 459   (int16x8_t)__builtin_neon_vaddlv8qi (__a, __b, 1)
 460
 461 #define vaddl_s16(__a, __b) \
 462   (int32x4_t)__builtin_neon_vaddlv4hi (__a, __b, 1)
 463
 464 #define vaddl_s32(__a, __b) \
 465   (int64x2_t)__builtin_neon_vaddlv2si (__a, __b, 1)
 466
 467 #define vaddl_u8(__a, __b) \
 468   (uint16x8_t)__builtin_neon_vaddlv8qi (__a, __b, 0)
 469
 470 #define vaddl_u16(__a, __b) \
 471   (uint32x4_t)__builtin_neon_vaddlv4hi (__a, __b, 0)
 472
 473 #define vaddl_u32(__a, __b) \
 474   (uint64x2_t)__builtin_neon_vaddlv2si (__a, __b, 0)
 475
 476 #define vaddw_s8(__a, __b) \
 477   (int16x8_t)__builtin_neon_vaddwv8qi (__a, __b, 1)
 478
 479 #define vaddw_s16(__a, __b) \
 480   (int32x4_t)__builtin_neon_vaddwv4hi (__a, __b, 1)
 481
 482 #define vaddw_s32(__a, __b) \
 483   (int64x2_t)__builtin_neon_vaddwv2si (__a, __b, 1)
 484
 485 #define vaddw_u8(__a, __b) \
 486   (uint16x8_t)__builtin_neon_vaddwv8qi (__a, __b, 0)
 487
 488 #define vaddw_u16(__a, __b) \
 489   (uint32x4_t)__builtin_neon_vaddwv4hi (__a, __b, 0)
 490
 491 #define vaddw_u32(__a, __b) \
 492   (uint64x2_t)__builtin_neon_vaddwv2si (__a, __b, 0)
 493
 494 #define vhadd_s8(__a, __b) \
 495   (int8x8_t)__builtin_neon_vhaddv8qi (__a, __b, 1)
 496
 497 #define vhadd_s16(__a, __b) \
 498   (int16x4_t)__builtin_neon_vhaddv4hi (__a, __b, 1)
 499
 500 #define vhadd_s32(__a, __b) \
 501   (int32x2_t)__builtin_neon_vhaddv2si (__a, __b, 1)
 502
 503 #define vhadd_u8(__a, __b) \
 504   (uint8x8_t)__builtin_neon_vhaddv8qi (__a, __b, 0)
 505
 506 #define vhadd_u16(__a, __b) \
 507   (uint16x4_t)__builtin_neon_vhaddv4hi (__a, __b, 0)
 508
 509 #define vhadd_u32(__a, __b) \
 510   (uint32x2_t)__builtin_neon_vhaddv2si (__a, __b, 0)
 511
 512 #define vhaddq_s8(__a, __b) \
 513   (int8x16_t)__builtin_neon_vhaddv16qi (__a, __b, 1)
 514
 515 #define vhaddq_s16(__a, __b) \
 516   (int16x8_t)__builtin_neon_vhaddv8hi (__a, __b, 1)
 517
 518 #define vhaddq_s32(__a, __b) \
 519   (int32x4_t)__builtin_neon_vhaddv4si (__a, __b, 1)
 520
 521 #define vhaddq_u8(__a, __b) \
 522   (uint8x16_t)__builtin_neon_vhaddv16qi (__a, __b, 0)
 523
 524 #define vhaddq_u16(__a, __b) \
 525   (uint16x8_t)__builtin_neon_vhaddv8hi (__a, __b, 0)
 526
 527 #define vhaddq_u32(__a, __b) \
 528   (uint32x4_t)__builtin_neon_vhaddv4si (__a, __b, 0)
 529
 530 #define vrhadd_s8(__a, __b) \
 531   (int8x8_t)__builtin_neon_vhaddv8qi (__a, __b, 3)
 532
 533 #define vrhadd_s16(__a, __b) \
 534   (int16x4_t)__builtin_neon_vhaddv4hi (__a, __b, 3)
 535
 536 #define vrhadd_s32(__a, __b) \
 537   (int32x2_t)__builtin_neon_vhaddv2si (__a, __b, 3)
 538
 539 #define vrhadd_u8(__a, __b) \
 540   (uint8x8_t)__builtin_neon_vhaddv8qi (__a, __b, 2)
 541
 542 #define vrhadd_u16(__a, __b) \
 543   (uint16x4_t)__builtin_neon_vhaddv4hi (__a, __b, 2)
 544
 545 #define vrhadd_u32(__a, __b) \
 546   (uint32x2_t)__builtin_neon_vhaddv2si (__a, __b, 2)
 547
 548 #define vrhaddq_s8(__a, __b) \
 549   (int8x16_t)__builtin_neon_vhaddv16qi (__a, __b, 3)
 550
 551 #define vrhaddq_s16(__a, __b) \
 552   (int16x8_t)__builtin_neon_vhaddv8hi (__a, __b, 3)
 553
 554 #define vrhaddq_s32(__a, __b) \
 555   (int32x4_t)__builtin_neon_vhaddv4si (__a, __b, 3)
 556
 557 #define vrhaddq_u8(__a, __b) \
 558   (uint8x16_t)__builtin_neon_vhaddv16qi (__a, __b, 2)
 559
 560 #define vrhaddq_u16(__a, __b) \
 561   (uint16x8_t)__builtin_neon_vhaddv8hi (__a, __b, 2)
 562
 563 #define vrhaddq_u32(__a, __b) \
 564   (uint32x4_t)__builtin_neon_vhaddv4si (__a, __b, 2)
 565
 566 #define vqadd_s8(__a, __b) \
 567   (int8x8_t)__builtin_neon_vqaddv8qi (__a, __b, 1)
 568
 569 #define vqadd_s16(__a, __b) \
 570   (int16x4_t)__builtin_neon_vqaddv4hi (__a, __b, 1)
 571
 572 #define vqadd_s32(__a, __b) \
 573   (int32x2_t)__builtin_neon_vqaddv2si (__a, __b, 1)
 574
 575 #define vqadd_s64(__a, __b) \
 576   (int64x1_t)__builtin_neon_vqaddv1di (__a, __b, 1)
 577
 578 #define vqadd_u8(__a, __b) \
 579   (uint8x8_t)__builtin_neon_vqaddv8qi (__a, __b, 0)
 580
 581 #define vqadd_u16(__a, __b) \
 582   (uint16x4_t)__builtin_neon_vqaddv4hi (__a, __b, 0)
 583
 584 #define vqadd_u32(__a, __b) \
 585   (uint32x2_t)__builtin_neon_vqaddv2si (__a, __b, 0)
 586
 587 #define vqadd_u64(__a, __b) \
 588   (uint64x1_t)__builtin_neon_vqaddv1di (__a, __b, 0)
 589
 590 #define vqaddq_s8(__a, __b) \
 591   (int8x16_t)__builtin_neon_vqaddv16qi (__a, __b, 1)
 592
 593 #define vqaddq_s16(__a, __b) \
 594   (int16x8_t)__builtin_neon_vqaddv8hi (__a, __b, 1)
 595
 596 #define vqaddq_s32(__a, __b) \
 597   (int32x4_t)__builtin_neon_vqaddv4si (__a, __b, 1)
 598
 599 #define vqaddq_s64(__a, __b) \
 600   (int64x2_t)__builtin_neon_vqaddv2di (__a, __b, 1)
 601
 602 #define vqaddq_u8(__a, __b) \
 603   (uint8x16_t)__builtin_neon_vqaddv16qi (__a, __b, 0)
 604
 605 #define vqaddq_u16(__a, __b) \
 606   (uint16x8_t)__builtin_neon_vqaddv8hi (__a, __b, 0)
 607
 608 #define vqaddq_u32(__a, __b) \
 609   (uint32x4_t)__builtin_neon_vqaddv4si (__a, __b, 0)
 610
 611 #define vqaddq_u64(__a, __b) \
 612   (uint64x2_t)__builtin_neon_vqaddv2di (__a, __b, 0)
 613
 614 #define vaddhn_s16(__a, __b) \
 615   (int8x8_t)__builtin_neon_vaddhnv8hi (__a, __b, 1)
 616
 617 #define vaddhn_s32(__a, __b) \
 618   (int16x4_t)__builtin_neon_vaddhnv4si (__a, __b, 1)
 619
 620 #define vaddhn_s64(__a, __b) \
 621   (int32x2_t)__builtin_neon_vaddhnv2di (__a, __b, 1)
 622
 623 #define vaddhn_u16(__a, __b) \
 624   (uint8x8_t)__builtin_neon_vaddhnv8hi (__a, __b, 0)
 625
 626 #define vaddhn_u32(__a, __b) \
 627   (uint16x4_t)__builtin_neon_vaddhnv4si (__a, __b, 0)
 628
 629 #define vaddhn_u64(__a, __b) \
 630   (uint32x2_t)__builtin_neon_vaddhnv2di (__a, __b, 0)
 631
 632 #define vraddhn_s16(__a, __b) \
 633   (int8x8_t)__builtin_neon_vaddhnv8hi (__a, __b, 3)
 634
 635 #define vraddhn_s32(__a, __b) \
 636   (int16x4_t)__builtin_neon_vaddhnv4si (__a, __b, 3)
 637
 638 #define vraddhn_s64(__a, __b) \
 639   (int32x2_t)__builtin_neon_vaddhnv2di (__a, __b, 3)
 640
 641 #define vraddhn_u16(__a, __b) \
 642   (uint8x8_t)__builtin_neon_vaddhnv8hi (__a, __b, 2)
 643
 644 #define vraddhn_u32(__a, __b) \
 645   (uint16x4_t)__builtin_neon_vaddhnv4si (__a, __b, 2)
 646
 647 #define vraddhn_u64(__a, __b) \
 648   (uint32x2_t)__builtin_neon_vaddhnv2di (__a, __b, 2)
 649
 650 #define vmul_s8(__a, __b) \
 651   (int8x8_t)__builtin_neon_vmulv8qi (__a, __b, 1)
 652
 653 #define vmul_s16(__a, __b) \
 654   (int16x4_t)__builtin_neon_vmulv4hi (__a, __b, 1)
 655
 656 #define vmul_s32(__a, __b) \
 657   (int32x2_t)__builtin_neon_vmulv2si (__a, __b, 1)
 658
 659 #define vmul_f32(__a, __b) \
 660   (float32x2_t)__builtin_neon_vmulv2sf (__a, __b, 5)
 661
 662 #define vmul_u8(__a, __b) \
 663   (uint8x8_t)__builtin_neon_vmulv8qi (__a, __b, 0)
 664
 665 #define vmul_u16(__a, __b) \
 666   (uint16x4_t)__builtin_neon_vmulv4hi (__a, __b, 0)
 667
 668 #define vmul_u32(__a, __b) \
 669   (uint32x2_t)__builtin_neon_vmulv2si (__a, __b, 0)
 670
 671 #define vmul_p8(__a, __b) \
 672   (poly8x8_t)__builtin_neon_vmulv8qi (__a, __b, 4)
 673
 674 #define vmulq_s8(__a, __b) \
 675   (int8x16_t)__builtin_neon_vmulv16qi (__a, __b, 1)
 676
 677 #define vmulq_s16(__a, __b) \
 678   (int16x8_t)__builtin_neon_vmulv8hi (__a, __b, 1)
 679
 680 #define vmulq_s32(__a, __b) \
 681   (int32x4_t)__builtin_neon_vmulv4si (__a, __b, 1)
 682
 683 #define vmulq_f32(__a, __b) \
 684   (float32x4_t)__builtin_neon_vmulv4sf (__a, __b, 5)
 685
 686 #define vmulq_u8(__a, __b) \
 687   (uint8x16_t)__builtin_neon_vmulv16qi (__a, __b, 0)
 688
 689 #define vmulq_u16(__a, __b) \
 690   (uint16x8_t)__builtin_neon_vmulv8hi (__a, __b, 0)
 691
 692 #define vmulq_u32(__a, __b) \
 693   (uint32x4_t)__builtin_neon_vmulv4si (__a, __b, 0)
 694
 695 #define vmulq_p8(__a, __b) \
 696   (poly8x16_t)__builtin_neon_vmulv16qi (__a, __b, 4)
 697
 698 #define vqdmulh_s16(__a, __b) \
 699   (int16x4_t)__builtin_neon_vqdmulhv4hi (__a, __b, 1)
 700
 701 #define vqdmulh_s32(__a, __b) \
 702   (int32x2_t)__builtin_neon_vqdmulhv2si (__a, __b, 1)
 703
 704 #define vqdmulhq_s16(__a, __b) \
 705   (int16x8_t)__builtin_neon_vqdmulhv8hi (__a, __b, 1)
 706
 707 #define vqdmulhq_s32(__a, __b) \
 708   (int32x4_t)__builtin_neon_vqdmulhv4si (__a, __b, 1)
 709
 710 #define vqrdmulh_s16(__a, __b) \
 711   (int16x4_t)__builtin_neon_vqdmulhv4hi (__a, __b, 3)
 712
 713 #define vqrdmulh_s32(__a, __b) \
 714   (int32x2_t)__builtin_neon_vqdmulhv2si (__a, __b, 3)
 715
 716 #define vqrdmulhq_s16(__a, __b) \
 717   (int16x8_t)__builtin_neon_vqdmulhv8hi (__a, __b, 3)
 718
 719 #define vqrdmulhq_s32(__a, __b) \
 720   (int32x4_t)__builtin_neon_vqdmulhv4si (__a, __b, 3)
 721
 722 #define vmull_s8(__a, __b) \
 723   (int16x8_t)__builtin_neon_vmullv8qi (__a, __b, 1)
 724
 725 #define vmull_s16(__a, __b) \
 726   (int32x4_t)__builtin_neon_vmullv4hi (__a, __b, 1)
 727
 728 #define vmull_s32(__a, __b) \
 729   (int64x2_t)__builtin_neon_vmullv2si (__a, __b, 1)
 730
 731 #define vmull_u8(__a, __b) \
 732   (uint16x8_t)__builtin_neon_vmullv8qi (__a, __b, 0)
 733
 734 #define vmull_u16(__a, __b) \
 735   (uint32x4_t)__builtin_neon_vmullv4hi (__a, __b, 0)
 736
 737 #define vmull_u32(__a, __b) \
 738   (uint64x2_t)__builtin_neon_vmullv2si (__a, __b, 0)
 739
 740 #define vmull_p8(__a, __b) \
 741   (poly16x8_t)__builtin_neon_vmullv8qi (__a, __b, 4)
 742
 743 #define vqdmull_s16(__a, __b) \
 744   (int32x4_t)__builtin_neon_vqdmullv4hi (__a, __b, 1)
 745
 746 #define vqdmull_s32(__a, __b) \
 747   (int64x2_t)__builtin_neon_vqdmullv2si (__a, __b, 1)
 748
 749 #define vmla_s8(__a, __b, __c) \
 750   (int8x8_t)__builtin_neon_vmlav8qi (__a, __b, __c, 1)
 751
 752 #define vmla_s16(__a, __b, __c) \
 753   (int16x4_t)__builtin_neon_vmlav4hi (__a, __b, __c, 1)
 754
 755 #define vmla_s32(__a, __b, __c) \
 756   (int32x2_t)__builtin_neon_vmlav2si (__a, __b, __c, 1)
 757
 758 #define vmla_f32(__a, __b, __c) \
 759   (float32x2_t)__builtin_neon_vmlav2sf (__a, __b, __c, 5)
 760
 761 #define vmla_u8(__a, __b, __c) \
 762   (uint8x8_t)__builtin_neon_vmlav8qi (__a, __b, __c, 0)
 763
 764 #define vmla_u16(__a, __b, __c) \
 765   (uint16x4_t)__builtin_neon_vmlav4hi (__a, __b, __c, 0)
 766
 767 #define vmla_u32(__a, __b, __c) \
 768   (uint32x2_t)__builtin_neon_vmlav2si (__a, __b, __c, 0)
 769
 770 #define vmlaq_s8(__a, __b, __c) \
 771   (int8x16_t)__builtin_neon_vmlav16qi (__a, __b, __c, 1)
 772
 773 #define vmlaq_s16(__a, __b, __c) \
 774   (int16x8_t)__builtin_neon_vmlav8hi (__a, __b, __c, 1)
 775
 776 #define vmlaq_s32(__a, __b, __c) \
 777   (int32x4_t)__builtin_neon_vmlav4si (__a, __b, __c, 1)
 778
 779 #define vmlaq_f32(__a, __b, __c) \
 780   (float32x4_t)__builtin_neon_vmlav4sf (__a, __b, __c, 5)
 781
 782 #define vmlaq_u8(__a, __b, __c) \
 783   (uint8x16_t)__builtin_neon_vmlav16qi (__a, __b, __c, 0)
 784
 785 #define vmlaq_u16(__a, __b, __c) \
 786   (uint16x8_t)__builtin_neon_vmlav8hi (__a, __b, __c, 0)
 787
 788 #define vmlaq_u32(__a, __b, __c) \
 789   (uint32x4_t)__builtin_neon_vmlav4si (__a, __b, __c, 0)
 790
 791 #define vmlal_s8(__a, __b, __c) \
 792   (int16x8_t)__builtin_neon_vmlalv8qi (__a, __b, __c, 1)
 793
 794 #define vmlal_s16(__a, __b, __c) \
 795   (int32x4_t)__builtin_neon_vmlalv4hi (__a, __b, __c, 1)
 796
 797 #define vmlal_s32(__a, __b, __c) \
 798   (int64x2_t)__builtin_neon_vmlalv2si (__a, __b, __c, 1)
 799
 800 #define vmlal_u8(__a, __b, __c) \
 801   (uint16x8_t)__builtin_neon_vmlalv8qi (__a, __b, __c, 0)
 802
 803 #define vmlal_u16(__a, __b, __c) \
 804   (uint32x4_t)__builtin_neon_vmlalv4hi (__a, __b, __c, 0)
 805
 806 #define vmlal_u32(__a, __b, __c) \
 807   (uint64x2_t)__builtin_neon_vmlalv2si (__a, __b, __c, 0)
 808
 809 #define vqdmlal_s16(__a, __b, __c) \
 810   (int32x4_t)__builtin_neon_vqdmlalv4hi (__a, __b, __c, 1)
 811
 812 #define vqdmlal_s32(__a, __b, __c) \
 813   (int64x2_t)__builtin_neon_vqdmlalv2si (__a, __b, __c, 1)
 814
 815 #define vmls_s8(__a, __b, __c) \
 816   (int8x8_t)__builtin_neon_vmlsv8qi (__a, __b, __c, 1)
 817
 818 #define vmls_s16(__a, __b, __c) \
 819   (int16x4_t)__builtin_neon_vmlsv4hi (__a, __b, __c, 1)
 820
 821 #define vmls_s32(__a, __b, __c) \
 822   (int32x2_t)__builtin_neon_vmlsv2si (__a, __b, __c, 1)
 823
 824 #define vmls_f32(__a, __b, __c) \
 825   (float32x2_t)__builtin_neon_vmlsv2sf (__a, __b, __c, 5)
 826
 827 #define vmls_u8(__a, __b, __c) \
 828   (uint8x8_t)__builtin_neon_vmlsv8qi (__a, __b, __c, 0)
 829
 830 #define vmls_u16(__a, __b, __c) \
 831   (uint16x4_t)__builtin_neon_vmlsv4hi (__a, __b, __c, 0)
 832
 833 #define vmls_u32(__a, __b, __c) \
 834   (uint32x2_t)__builtin_neon_vmlsv2si (__a, __b, __c, 0)
 835
 836 #define vmlsq_s8(__a, __b, __c) \
 837   (int8x16_t)__builtin_neon_vmlsv16qi (__a, __b, __c, 1)
 838
 839 #define vmlsq_s16(__a, __b, __c) \
 840   (int16x8_t)__builtin_neon_vmlsv8hi (__a, __b, __c, 1)
 841
 842 #define vmlsq_s32(__a, __b, __c) \
 843   (int32x4_t)__builtin_neon_vmlsv4si (__a, __b, __c, 1)
 844
 845 #define vmlsq_f32(__a, __b, __c) \
 846   (float32x4_t)__builtin_neon_vmlsv4sf (__a, __b, __c, 5)
 847
 848 #define vmlsq_u8(__a, __b, __c) \
 849   (uint8x16_t)__builtin_neon_vmlsv16qi (__a, __b, __c, 0)
 850
 851 #define vmlsq_u16(__a, __b, __c) \
 852   (uint16x8_t)__builtin_neon_vmlsv8hi (__a, __b, __c, 0)
 853
 854 #define vmlsq_u32(__a, __b, __c) \
 855   (uint32x4_t)__builtin_neon_vmlsv4si (__a, __b, __c, 0)
 856
 857 #define vmlsl_s8(__a, __b, __c) \
 858   (int16x8_t)__builtin_neon_vmlslv8qi (__a, __b, __c, 1)
 859
 860 #define vmlsl_s16(__a, __b, __c) \
 861   (int32x4_t)__builtin_neon_vmlslv4hi (__a, __b, __c, 1)
 862
 863 #define vmlsl_s32(__a, __b, __c) \
 864   (int64x2_t)__builtin_neon_vmlslv2si (__a, __b, __c, 1)
 865
 866 #define vmlsl_u8(__a, __b, __c) \
 867   (uint16x8_t)__builtin_neon_vmlslv8qi (__a, __b, __c, 0)
 868
 869 #define vmlsl_u16(__a, __b, __c) \
 870   (uint32x4_t)__builtin_neon_vmlslv4hi (__a, __b, __c, 0)
 871
 872 #define vmlsl_u32(__a, __b, __c) \
 873   (uint64x2_t)__builtin_neon_vmlslv2si (__a, __b, __c, 0)
 874
 875 #define vqdmlsl_s16(__a, __b, __c) \
 876   (int32x4_t)__builtin_neon_vqdmlslv4hi (__a, __b, __c, 1)
 877
 878 #define vqdmlsl_s32(__a, __b, __c) \
 879   (int64x2_t)__builtin_neon_vqdmlslv2si (__a, __b, __c, 1)
 880
 881 #define vsub_s8(__a, __b) \
 882   (int8x8_t)__builtin_neon_vsubv8qi (__a, __b, 1)
 883
 884 #define vsub_s16(__a, __b) \
 885   (int16x4_t)__builtin_neon_vsubv4hi (__a, __b, 1)
 886
 887 #define vsub_s32(__a, __b) \
 888   (int32x2_t)__builtin_neon_vsubv2si (__a, __b, 1)
 889
 890 #define vsub_s64(__a, __b) \
 891   (int64x1_t)__builtin_neon_vsubv1di (__a, __b, 1)
 892
 893 #define vsub_f32(__a, __b) \
 894   (float32x2_t)__builtin_neon_vsubv2sf (__a, __b, 5)
 895
 896 #define vsub_u8(__a, __b) \
 897   (uint8x8_t)__builtin_neon_vsubv8qi (__a, __b, 0)
 898
 899 #define vsub_u16(__a, __b) \
 900   (uint16x4_t)__builtin_neon_vsubv4hi (__a, __b, 0)
 901
 902 #define vsub_u32(__a, __b) \
 903   (uint32x2_t)__builtin_neon_vsubv2si (__a, __b, 0)
 904
 905 #define vsub_u64(__a, __b) \
 906   (uint64x1_t)__builtin_neon_vsubv1di (__a, __b, 0)
 907
 908 #define vsubq_s8(__a, __b) \
 909   (int8x16_t)__builtin_neon_vsubv16qi (__a, __b, 1)
 910
 911 #define vsubq_s16(__a, __b) \
 912   (int16x8_t)__builtin_neon_vsubv8hi (__a, __b, 1)
 913
 914 #define vsubq_s32(__a, __b) \
 915   (int32x4_t)__builtin_neon_vsubv4si (__a, __b, 1)
 916
 917 #define vsubq_s64(__a, __b) \
 918   (int64x2_t)__builtin_neon_vsubv2di (__a, __b, 1)
 919
 920 #define vsubq_f32(__a, __b) \
 921   (float32x4_t)__builtin_neon_vsubv4sf (__a, __b, 5)
 922
 923 #define vsubq_u8(__a, __b) \
 924   (uint8x16_t)__builtin_neon_vsubv16qi (__a, __b, 0)
 925
 926 #define vsubq_u16(__a, __b) \
 927   (uint16x8_t)__builtin_neon_vsubv8hi (__a, __b, 0)
 928
 929 #define vsubq_u32(__a, __b) \
 930   (uint32x4_t)__builtin_neon_vsubv4si (__a, __b, 0)
 931
 932 #define vsubq_u64(__a, __b) \
 933   (uint64x2_t)__builtin_neon_vsubv2di (__a, __b, 0)
 934
 935 #define vsubl_s8(__a, __b) \
 936   (int16x8_t)__builtin_neon_vsublv8qi (__a, __b, 1)
 937
 938 #define vsubl_s16(__a, __b) \
 939   (int32x4_t)__builtin_neon_vsublv4hi (__a, __b, 1)
 940
 941 #define vsubl_s32(__a, __b) \
 942   (int64x2_t)__builtin_neon_vsublv2si (__a, __b, 1)
 943
 944 #define vsubl_u8(__a, __b) \
 945   (uint16x8_t)__builtin_neon_vsublv8qi (__a, __b, 0)
 946
 947 #define vsubl_u16(__a, __b) \
 948   (uint32x4_t)__builtin_neon_vsublv4hi (__a, __b, 0)
 949
 950 #define vsubl_u32(__a, __b) \
 951   (uint64x2_t)__builtin_neon_vsublv2si (__a, __b, 0)
 952
 953 #define vsubw_s8(__a, __b) \
 954   (int16x8_t)__builtin_neon_vsubwv8qi (__a, __b, 1)
 955
 956 #define vsubw_s16(__a, __b) \
 957   (int32x4_t)__builtin_neon_vsubwv4hi (__a, __b, 1)
 958
 959 #define vsubw_s32(__a, __b) \
 960   (int64x2_t)__builtin_neon_vsubwv2si (__a, __b, 1)
 961
 962 #define vsubw_u8(__a, __b) \
 963   (uint16x8_t)__builtin_neon_vsubwv8qi (__a, __b, 0)
 964
 965 #define vsubw_u16(__a, __b) \
 966   (uint32x4_t)__builtin_neon_vsubwv4hi (__a, __b, 0)
 967
 968 #define vsubw_u32(__a, __b) \
 969   (uint64x2_t)__builtin_neon_vsubwv2si (__a, __b, 0)
 970
 971 #define vhsub_s8(__a, __b) \
 972   (int8x8_t)__builtin_neon_vhsubv8qi (__a, __b, 1)
 973
 974 #define vhsub_s16(__a, __b) \
 975   (int16x4_t)__builtin_neon_vhsubv4hi (__a, __b, 1)
 976
 977 #define vhsub_s32(__a, __b) \
 978   (int32x2_t)__builtin_neon_vhsubv2si (__a, __b, 1)
 979
 980 #define vhsub_u8(__a, __b) \
 981   (uint8x8_t)__builtin_neon_vhsubv8qi (__a, __b, 0)
 982
 983 #define vhsub_u16(__a, __b) \
 984   (uint16x4_t)__builtin_neon_vhsubv4hi (__a, __b, 0)
 985
 986 #define vhsub_u32(__a, __b) \
 987   (uint32x2_t)__builtin_neon_vhsubv2si (__a, __b, 0)
 988
 989 #define vhsubq_s8(__a, __b) \
 990   (int8x16_t)__builtin_neon_vhsubv16qi (__a, __b, 1)
 991
 992 #define vhsubq_s16(__a, __b) \
 993   (int16x8_t)__builtin_neon_vhsubv8hi (__a, __b, 1)
 994
 995 #define vhsubq_s32(__a, __b) \
 996   (int32x4_t)__builtin_neon_vhsubv4si (__a, __b, 1)
 997
 998 #define vhsubq_u8(__a, __b) \
 999   (uint8x16_t)__builtin_neon_vhsubv16qi (__a, __b, 0)
1000
1001 #define vhsubq_u16(__a, __b) \
1002   (uint16x8_t)__builtin_neon_vhsubv8hi (__a, __b, 0)
1003
1004 #define vhsubq_u32(__a, __b) \
1005   (uint32x4_t)__builtin_neon_vhsubv4si (__a, __b, 0)
1006
1007 #define vqsub_s8(__a, __b) \
1008   (int8x8_t)__builtin_neon_vqsubv8qi (__a, __b, 1)
1009
1010 #define vqsub_s16(__a, __b) \
1011   (int16x4_t)__builtin_neon_vqsubv4hi (__a, __b, 1)
1012
1013 #define vqsub_s32(__a, __b) \
1014   (int32x2_t)__builtin_neon_vqsubv2si (__a, __b, 1)
1015
1016 #define vqsub_s64(__a, __b) \
1017   (int64x1_t)__builtin_neon_vqsubv1di (__a, __b, 1)
1018
1019 #define vqsub_u8(__a, __b) \
1020   (uint8x8_t)__builtin_neon_vqsubv8qi (__a, __b, 0)
1021
1022 #define vqsub_u16(__a, __b) \
1023   (uint16x4_t)__builtin_neon_vqsubv4hi (__a, __b, 0)
1024
1025 #define vqsub_u32(__a, __b) \
1026   (uint32x2_t)__builtin_neon_vqsubv2si (__a, __b, 0)
1027
1028 #define vqsub_u64(__a, __b) \
1029   (uint64x1_t)__builtin_neon_vqsubv1di (__a, __b, 0)
1030
1031 #define vqsubq_s8(__a, __b) \
1032   (int8x16_t)__builtin_neon_vqsubv16qi (__a, __b, 1)
1033
1034 #define vqsubq_s16(__a, __b) \
1035   (int16x8_t)__builtin_neon_vqsubv8hi (__a, __b, 1)
1036
1037 #define vqsubq_s32(__a, __b) \
1038   (int32x4_t)__builtin_neon_vqsubv4si (__a, __b, 1)
1039
1040 #define vqsubq_s64(__a, __b) \
1041   (int64x2_t)__builtin_neon_vqsubv2di (__a, __b, 1)
1042
1043 #define vqsubq_u8(__a, __b) \
1044   (uint8x16_t)__builtin_neon_vqsubv16qi (__a, __b, 0)
1045
1046 #define vqsubq_u16(__a, __b) \
1047   (uint16x8_t)__builtin_neon_vqsubv8hi (__a, __b, 0)
1048
1049 #define vqsubq_u32(__a, __b) \
1050   (uint32x4_t)__builtin_neon_vqsubv4si (__a, __b, 0)
1051
1052 #define vqsubq_u64(__a, __b) \
1053   (uint64x2_t)__builtin_neon_vqsubv2di (__a, __b, 0)
1054
1055 #define vsubhn_s16(__a, __b) \
1056   (int8x8_t)__builtin_neon_vsubhnv8hi (__a, __b, 1)
1057
1058 #define vsubhn_s32(__a, __b) \
1059   (int16x4_t)__builtin_neon_vsubhnv4si (__a, __b, 1)
1060
1061 #define vsubhn_s64(__a, __b) \
1062   (int32x2_t)__builtin_neon_vsubhnv2di (__a, __b, 1)
1063
1064 #define vsubhn_u16(__a, __b) \
1065   (uint8x8_t)__builtin_neon_vsubhnv8hi (__a, __b, 0)
1066
1067 #define vsubhn_u32(__a, __b) \
1068   (uint16x4_t)__builtin_neon_vsubhnv4si (__a, __b, 0)
1069
1070 #define vsubhn_u64(__a, __b) \
1071   (uint32x2_t)__builtin_neon_vsubhnv2di (__a, __b, 0)
1072
1073 #define vrsubhn_s16(__a, __b) \
1074   (int8x8_t)__builtin_neon_vsubhnv8hi (__a, __b, 3)
1075
1076 #define vrsubhn_s32(__a, __b) \
1077   (int16x4_t)__builtin_neon_vsubhnv4si (__a, __b, 3)
1078
1079 #define vrsubhn_s64(__a, __b) \
1080   (int32x2_t)__builtin_neon_vsubhnv2di (__a, __b, 3)
1081
1082 #define vrsubhn_u16(__a, __b) \
1083   (uint8x8_t)__builtin_neon_vsubhnv8hi (__a, __b, 2)
1084
1085 #define vrsubhn_u32(__a, __b) \
1086   (uint16x4_t)__builtin_neon_vsubhnv4si (__a, __b, 2)
1087
1088 #define vrsubhn_u64(__a, __b) \
1089   (uint32x2_t)__builtin_neon_vsubhnv2di (__a, __b, 2)
1090
1091 #define vceq_s8(__a, __b) \
1092   (uint8x8_t)__builtin_neon_vceqv8qi (__a, __b, 1)
1093
1094 #define vceq_s16(__a, __b) \
1095   (uint16x4_t)__builtin_neon_vceqv4hi (__a, __b, 1)
1096
1097 #define vceq_s32(__a, __b) \
1098   (uint32x2_t)__builtin_neon_vceqv2si (__a, __b, 1)
1099
1100 #define vceq_f32(__a, __b) \
1101   (uint32x2_t)__builtin_neon_vceqv2sf (__a, __b, 5)
1102
1103 #define vceq_u8(__a, __b) \
1104   (uint8x8_t)__builtin_neon_vceqv8qi (__a, __b, 0)
1105
1106 #define vceq_u16(__a, __b) \
1107   (uint16x4_t)__builtin_neon_vceqv4hi (__a, __b, 0)
1108
1109 #define vceq_u32(__a, __b) \
1110   (uint32x2_t)__builtin_neon_vceqv2si (__a, __b, 0)
1111
1112 #define vceq_p8(__a, __b) \
1113   (uint8x8_t)__builtin_neon_vceqv8qi (__a, __b, 4)
1114
1115 #define vceqq_s8(__a, __b) \
1116   (uint8x16_t)__builtin_neon_vceqv16qi (__a, __b, 1)
1117
1118 #define vceqq_s16(__a, __b) \
1119   (uint16x8_t)__builtin_neon_vceqv8hi (__a, __b, 1)
1120
1121 #define vceqq_s32(__a, __b) \
1122   (uint32x4_t)__builtin_neon_vceqv4si (__a, __b, 1)
1123
1124 #define vceqq_f32(__a, __b) \
1125   (uint32x4_t)__builtin_neon_vceqv4sf (__a, __b, 5)
1126
1127 #define vceqq_u8(__a, __b) \
1128   (uint8x16_t)__builtin_neon_vceqv16qi (__a, __b, 0)
1129
1130 #define vceqq_u16(__a, __b) \
1131   (uint16x8_t)__builtin_neon_vceqv8hi (__a, __b, 0)
1132
1133 #define vceqq_u32(__a, __b) \
1134   (uint32x4_t)__builtin_neon_vceqv4si (__a, __b, 0)
1135
1136 #define vceqq_p8(__a, __b) \
1137   (uint8x16_t)__builtin_neon_vceqv16qi (__a, __b, 4)
1138
1139 #define vcge_s8(__a, __b) \
1140   (uint8x8_t)__builtin_neon_vcgev8qi (__a, __b, 1)
1141
1142 #define vcge_s16(__a, __b) \
1143   (uint16x4_t)__builtin_neon_vcgev4hi (__a, __b, 1)
1144
1145 #define vcge_s32(__a, __b) \
1146   (uint32x2_t)__builtin_neon_vcgev2si (__a, __b, 1)
1147
1148 #define vcge_f32(__a, __b) \
1149   (uint32x2_t)__builtin_neon_vcgev2sf (__a, __b, 5)
1150
1151 #define vcge_u8(__a, __b) \
1152   (uint8x8_t)__builtin_neon_vcgev8qi (__a, __b, 0)
1153
1154 #define vcge_u16(__a, __b) \
1155   (uint16x4_t)__builtin_neon_vcgev4hi (__a, __b, 0)
1156
1157 #define vcge_u32(__a, __b) \
1158   (uint32x2_t)__builtin_neon_vcgev2si (__a, __b, 0)
1159
1160 #define vcgeq_s8(__a, __b) \
1161   (uint8x16_t)__builtin_neon_vcgev16qi (__a, __b, 1)
1162
1163 #define vcgeq_s16(__a, __b) \
1164   (uint16x8_t)__builtin_neon_vcgev8hi (__a, __b, 1)
1165
1166 #define vcgeq_s32(__a, __b) \
1167   (uint32x4_t)__builtin_neon_vcgev4si (__a, __b, 1)
1168
1169 #define vcgeq_f32(__a, __b) \
1170   (uint32x4_t)__builtin_neon_vcgev4sf (__a, __b, 5)
1171
1172 #define vcgeq_u8(__a, __b) \
1173   (uint8x16_t)__builtin_neon_vcgev16qi (__a, __b, 0)
1174
1175 #define vcgeq_u16(__a, __b) \
1176   (uint16x8_t)__builtin_neon_vcgev8hi (__a, __b, 0)
1177
1178 #define vcgeq_u32(__a, __b) \
1179   (uint32x4_t)__builtin_neon_vcgev4si (__a, __b, 0)
1180
1181 #define vcle_s8(__a, __b) \
1182   (uint8x8_t)__builtin_neon_vcgev8qi (__b, __a, 1)
1183
1184 #define vcle_s16(__a, __b) \
1185   (uint16x4_t)__builtin_neon_vcgev4hi (__b, __a, 1)
1186
1187 #define vcle_s32(__a, __b) \
1188   (uint32x2_t)__builtin_neon_vcgev2si (__b, __a, 1)
1189
1190 #define vcle_f32(__a, __b) \
1191   (uint32x2_t)__builtin_neon_vcgev2sf (__b, __a, 5)
1192
1193 #define vcle_u8(__a, __b) \
1194   (uint8x8_t)__builtin_neon_vcgev8qi (__b, __a, 0)
1195
1196 #define vcle_u16(__a, __b) \
1197   (uint16x4_t)__builtin_neon_vcgev4hi (__b, __a, 0)
1198
1199 #define vcle_u32(__a, __b) \
1200   (uint32x2_t)__builtin_neon_vcgev2si (__b, __a, 0)
1201
1202 #define vcleq_s8(__a, __b) \
1203   (uint8x16_t)__builtin_neon_vcgev16qi (__b, __a, 1)
1204
1205 #define vcleq_s16(__a, __b) \
1206   (uint16x8_t)__builtin_neon_vcgev8hi (__b, __a, 1)
1207
1208 #define vcleq_s32(__a, __b) \
1209   (uint32x4_t)__builtin_neon_vcgev4si (__b, __a, 1)
1210
1211 #define vcleq_f32(__a, __b) \
1212   (uint32x4_t)__builtin_neon_vcgev4sf (__b, __a, 5)
1213
1214 #define vcleq_u8(__a, __b) \
1215   (uint8x16_t)__builtin_neon_vcgev16qi (__b, __a, 0)
1216
1217 #define vcleq_u16(__a, __b) \
1218   (uint16x8_t)__builtin_neon_vcgev8hi (__b, __a, 0)
1219
1220 #define vcleq_u32(__a, __b) \
1221   (uint32x4_t)__builtin_neon_vcgev4si (__b, __a, 0)
1222
1223 #define vcgt_s8(__a, __b) \
1224   (uint8x8_t)__builtin_neon_vcgtv8qi (__a, __b, 1)
1225
1226 #define vcgt_s16(__a, __b) \
1227   (uint16x4_t)__builtin_neon_vcgtv4hi (__a, __b, 1)
1228
1229 #define vcgt_s32(__a, __b) \
1230   (uint32x2_t)__builtin_neon_vcgtv2si (__a, __b, 1)
1231
1232 #define vcgt_f32(__a, __b) \
1233   (uint32x2_t)__builtin_neon_vcgtv2sf (__a, __b, 5)
1234
1235 #define vcgt_u8(__a, __b) \
1236   (uint8x8_t)__builtin_neon_vcgtv8qi (__a, __b, 0)
1237
1238 #define vcgt_u16(__a, __b) \
1239   (uint16x4_t)__builtin_neon_vcgtv4hi (__a, __b, 0)
1240
1241 #define vcgt_u32(__a, __b) \
1242   (uint32x2_t)__builtin_neon_vcgtv2si (__a, __b, 0)
1243
1244 #define vcgtq_s8(__a, __b) \
1245   (uint8x16_t)__builtin_neon_vcgtv16qi (__a, __b, 1)
1246
1247 #define vcgtq_s16(__a, __b) \
1248   (uint16x8_t)__builtin_neon_vcgtv8hi (__a, __b, 1)
1249
1250 #define vcgtq_s32(__a, __b) \
1251   (uint32x4_t)__builtin_neon_vcgtv4si (__a, __b, 1)
1252
1253 #define vcgtq_f32(__a, __b) \
1254   (uint32x4_t)__builtin_neon_vcgtv4sf (__a, __b, 5)
1255
1256 #define vcgtq_u8(__a, __b) \
1257   (uint8x16_t)__builtin_neon_vcgtv16qi (__a, __b, 0)
1258
1259 #define vcgtq_u16(__a, __b) \
1260   (uint16x8_t)__builtin_neon_vcgtv8hi (__a, __b, 0)
1261
1262 #define vcgtq_u32(__a, __b) \
1263   (uint32x4_t)__builtin_neon_vcgtv4si (__a, __b, 0)
1264
1265 #define vclt_s8(__a, __b) \
1266   (uint8x8_t)__builtin_neon_vcgtv8qi (__b, __a, 1)
1267
1268 #define vclt_s16(__a, __b) \
1269   (uint16x4_t)__builtin_neon_vcgtv4hi (__b, __a, 1)
1270
1271 #define vclt_s32(__a, __b) \
1272   (uint32x2_t)__builtin_neon_vcgtv2si (__b, __a, 1)
1273
1274 #define vclt_f32(__a, __b) \
1275   (uint32x2_t)__builtin_neon_vcgtv2sf (__b, __a, 5)
1276
1277 #define vclt_u8(__a, __b) \
1278   (uint8x8_t)__builtin_neon_vcgtv8qi (__b, __a, 0)
1279
1280 #define vclt_u16(__a, __b) \
1281   (uint16x4_t)__builtin_neon_vcgtv4hi (__b, __a, 0)
1282
1283 #define vclt_u32(__a, __b) \
1284   (uint32x2_t)__builtin_neon_vcgtv2si (__b, __a, 0)
1285
1286 #define vcltq_s8(__a, __b) \
1287   (uint8x16_t)__builtin_neon_vcgtv16qi (__b, __a, 1)
1288
1289 #define vcltq_s16(__a, __b) \
1290   (uint16x8_t)__builtin_neon_vcgtv8hi (__b, __a, 1)
1291
1292 #define vcltq_s32(__a, __b) \
1293   (uint32x4_t)__builtin_neon_vcgtv4si (__b, __a, 1)
1294
1295 #define vcltq_f32(__a, __b) \
1296   (uint32x4_t)__builtin_neon_vcgtv4sf (__b, __a, 5)
1297
1298 #define vcltq_u8(__a, __b) \
1299   (uint8x16_t)__builtin_neon_vcgtv16qi (__b, __a, 0)
1300
1301 #define vcltq_u16(__a, __b) \
1302   (uint16x8_t)__builtin_neon_vcgtv8hi (__b, __a, 0)
1303
1304 #define vcltq_u32(__a, __b) \
1305   (uint32x4_t)__builtin_neon_vcgtv4si (__b, __a, 0)
1306
1307 #define vcage_f32(__a, __b) \
1308   (uint32x2_t)__builtin_neon_vcagev2sf (__a, __b, 5)
1309
1310 #define vcageq_f32(__a, __b) \
1311   (uint32x4_t)__builtin_neon_vcagev4sf (__a, __b, 5)
1312
1313 #define vcale_f32(__a, __b) \
1314   (uint32x2_t)__builtin_neon_vcagev2sf (__b, __a, 5)
1315
1316 #define vcaleq_f32(__a, __b) \
1317   (uint32x4_t)__builtin_neon_vcagev4sf (__b, __a, 5)
1318
1319 #define vcagt_f32(__a, __b) \
1320   (uint32x2_t)__builtin_neon_vcagtv2sf (__a, __b, 5)
1321
1322 #define vcagtq_f32(__a, __b) \
1323   (uint32x4_t)__builtin_neon_vcagtv4sf (__a, __b, 5)
1324
1325 #define vcalt_f32(__a, __b) \
1326   (uint32x2_t)__builtin_neon_vcagtv2sf (__b, __a, 5)
1327
1328 #define vcaltq_f32(__a, __b) \
1329   (uint32x4_t)__builtin_neon_vcagtv4sf (__b, __a, 5)
1330
1331 #define vtst_s8(__a, __b) \
1332   (uint8x8_t)__builtin_neon_vtstv8qi (__a, __b, 1)
1333
1334 #define vtst_s16(__a, __b) \
1335   (uint16x4_t)__builtin_neon_vtstv4hi (__a, __b, 1)
1336
1337 #define vtst_s32(__a, __b) \
1338   (uint32x2_t)__builtin_neon_vtstv2si (__a, __b, 1)
1339
1340 #define vtst_u8(__a, __b) \
1341   (uint8x8_t)__builtin_neon_vtstv8qi (__a, __b, 0)
1342
1343 #define vtst_u16(__a, __b) \
1344   (uint16x4_t)__builtin_neon_vtstv4hi (__a, __b, 0)
1345
1346 #define vtst_u32(__a, __b) \
1347   (uint32x2_t)__builtin_neon_vtstv2si (__a, __b, 0)
1348
1349 #define vtst_p8(__a, __b) \
1350   (uint8x8_t)__builtin_neon_vtstv8qi (__a, __b, 4)
1351
1352 #define vtstq_s8(__a, __b) \
1353   (uint8x16_t)__builtin_neon_vtstv16qi (__a, __b, 1)
1354
1355 #define vtstq_s16(__a, __b) \
1356   (uint16x8_t)__builtin_neon_vtstv8hi (__a, __b, 1)
1357
1358 #define vtstq_s32(__a, __b) \
1359   (uint32x4_t)__builtin_neon_vtstv4si (__a, __b, 1)
1360
1361 #define vtstq_u8(__a, __b) \
1362   (uint8x16_t)__builtin_neon_vtstv16qi (__a, __b, 0)
1363
1364 #define vtstq_u16(__a, __b) \
1365   (uint16x8_t)__builtin_neon_vtstv8hi (__a, __b, 0)
1366
1367 #define vtstq_u32(__a, __b) \
1368   (uint32x4_t)__builtin_neon_vtstv4si (__a, __b, 0)
1369
1370 #define vtstq_p8(__a, __b) \
1371   (uint8x16_t)__builtin_neon_vtstv16qi (__a, __b, 4)
1372
1373 #define vabd_s8(__a, __b) \
1374   (int8x8_t)__builtin_neon_vabdv8qi (__a, __b, 1)
1375
1376 #define vabd_s16(__a, __b) \
1377   (int16x4_t)__builtin_neon_vabdv4hi (__a, __b, 1)
1378
1379 #define vabd_s32(__a, __b) \
1380   (int32x2_t)__builtin_neon_vabdv2si (__a, __b, 1)
1381
1382 #define vabd_f32(__a, __b) \
1383   (float32x2_t)__builtin_neon_vabdv2sf (__a, __b, 5)
1384
1385 #define vabd_u8(__a, __b) \
1386   (uint8x8_t)__builtin_neon_vabdv8qi (__a, __b, 0)
1387
1388 #define vabd_u16(__a, __b) \
1389   (uint16x4_t)__builtin_neon_vabdv4hi (__a, __b, 0)
1390
1391 #define vabd_u32(__a, __b) \
1392   (uint32x2_t)__builtin_neon_vabdv2si (__a, __b, 0)
1393
1394 #define vabdq_s8(__a, __b) \
1395   (int8x16_t)__builtin_neon_vabdv16qi (__a, __b, 1)
1396
1397 #define vabdq_s16(__a, __b) \
1398   (int16x8_t)__builtin_neon_vabdv8hi (__a, __b, 1)
1399
1400 #define vabdq_s32(__a, __b) \
1401   (int32x4_t)__builtin_neon_vabdv4si (__a, __b, 1)
1402
1403 #define vabdq_f32(__a, __b) \
1404   (float32x4_t)__builtin_neon_vabdv4sf (__a, __b, 5)
1405
1406 #define vabdq_u8(__a, __b) \
1407   (uint8x16_t)__builtin_neon_vabdv16qi (__a, __b, 0)
1408
1409 #define vabdq_u16(__a, __b) \
1410   (uint16x8_t)__builtin_neon_vabdv8hi (__a, __b, 0)
1411
1412 #define vabdq_u32(__a, __b) \
1413   (uint32x4_t)__builtin_neon_vabdv4si (__a, __b, 0)
1414
1415 #define vabdl_s8(__a, __b) \
1416   (int16x8_t)__builtin_neon_vabdlv8qi (__a, __b, 1)
1417
1418 #define vabdl_s16(__a, __b) \
1419   (int32x4_t)__builtin_neon_vabdlv4hi (__a, __b, 1)
1420
1421 #define vabdl_s32(__a, __b) \
1422   (int64x2_t)__builtin_neon_vabdlv2si (__a, __b, 1)
1423
1424 #define vabdl_u8(__a, __b) \
1425   (uint16x8_t)__builtin_neon_vabdlv8qi (__a, __b, 0)
1426
1427 #define vabdl_u16(__a, __b) \
1428   (uint32x4_t)__builtin_neon_vabdlv4hi (__a, __b, 0)
1429
1430 #define vabdl_u32(__a, __b) \
1431   (uint64x2_t)__builtin_neon_vabdlv2si (__a, __b, 0)
1432
1433 #define vaba_s8(__a, __b, __c) \
1434   (int8x8_t)__builtin_neon_vabav8qi (__a, __b, __c, 1)
1435
1436 #define vaba_s16(__a, __b, __c) \
1437   (int16x4_t)__builtin_neon_vabav4hi (__a, __b, __c, 1)
1438
1439 #define vaba_s32(__a, __b, __c) \
1440   (int32x2_t)__builtin_neon_vabav2si (__a, __b, __c, 1)
1441
1442 #define vaba_u8(__a, __b, __c) \
1443   (uint8x8_t)__builtin_neon_vabav8qi (__a, __b, __c, 0)
1444
1445 #define vaba_u16(__a, __b, __c) \
1446   (uint16x4_t)__builtin_neon_vabav4hi (__a, __b, __c, 0)
1447
1448 #define vaba_u32(__a, __b, __c) \
1449   (uint32x2_t)__builtin_neon_vabav2si (__a, __b, __c, 0)
1450
1451 #define vabaq_s8(__a, __b, __c) \
1452   (int8x16_t)__builtin_neon_vabav16qi (__a, __b, __c, 1)
1453
1454 #define vabaq_s16(__a, __b, __c) \
1455   (int16x8_t)__builtin_neon_vabav8hi (__a, __b, __c, 1)
1456
1457 #define vabaq_s32(__a, __b, __c) \
1458   (int32x4_t)__builtin_neon_vabav4si (__a, __b, __c, 1)
1459
1460 #define vabaq_u8(__a, __b, __c) \
1461   (uint8x16_t)__builtin_neon_vabav16qi (__a, __b, __c, 0)
1462
1463 #define vabaq_u16(__a, __b, __c) \
1464   (uint16x8_t)__builtin_neon_vabav8hi (__a, __b, __c, 0)
1465
1466 #define vabaq_u32(__a, __b, __c) \
1467   (uint32x4_t)__builtin_neon_vabav4si (__a, __b, __c, 0)
1468
1469 #define vabal_s8(__a, __b, __c) \
1470   (int16x8_t)__builtin_neon_vabalv8qi (__a, __b, __c, 1)
1471
1472 #define vabal_s16(__a, __b, __c) \
1473   (int32x4_t)__builtin_neon_vabalv4hi (__a, __b, __c, 1)
1474
1475 #define vabal_s32(__a, __b, __c) \
1476   (int64x2_t)__builtin_neon_vabalv2si (__a, __b, __c, 1)
1477
1478 #define vabal_u8(__a, __b, __c) \
1479   (uint16x8_t)__builtin_neon_vabalv8qi (__a, __b, __c, 0)
1480
1481 #define vabal_u16(__a, __b, __c) \
1482   (uint32x4_t)__builtin_neon_vabalv4hi (__a, __b, __c, 0)
1483
1484 #define vabal_u32(__a, __b, __c) \
1485   (uint64x2_t)__builtin_neon_vabalv2si (__a, __b, __c, 0)
1486
1487 #define vmax_s8(__a, __b) \
1488   (int8x8_t)__builtin_neon_vmaxv8qi (__a, __b, 1)
1489
1490 #define vmax_s16(__a, __b) \
1491   (int16x4_t)__builtin_neon_vmaxv4hi (__a, __b, 1)
1492
1493 #define vmax_s32(__a, __b) \
1494   (int32x2_t)__builtin_neon_vmaxv2si (__a, __b, 1)
1495
1496 #define vmax_f32(__a, __b) \
1497   (float32x2_t)__builtin_neon_vmaxv2sf (__a, __b, 5)
1498
1499 #define vmax_u8(__a, __b) \
1500   (uint8x8_t)__builtin_neon_vmaxv8qi (__a, __b, 0)
1501
1502 #define vmax_u16(__a, __b) \
1503   (uint16x4_t)__builtin_neon_vmaxv4hi (__a, __b, 0)
1504
1505 #define vmax_u32(__a, __b) \
1506   (uint32x2_t)__builtin_neon_vmaxv2si (__a, __b, 0)
1507
1508 #define vmaxq_s8(__a, __b) \
1509   (int8x16_t)__builtin_neon_vmaxv16qi (__a, __b, 1)
1510
1511 #define vmaxq_s16(__a, __b) \
1512   (int16x8_t)__builtin_neon_vmaxv8hi (__a, __b, 1)
1513
1514 #define vmaxq_s32(__a, __b) \
1515   (int32x4_t)__builtin_neon_vmaxv4si (__a, __b, 1)
1516
1517 #define vmaxq_f32(__a, __b) \
1518   (float32x4_t)__builtin_neon_vmaxv4sf (__a, __b, 5)
1519
1520 #define vmaxq_u8(__a, __b) \
1521   (uint8x16_t)__builtin_neon_vmaxv16qi (__a, __b, 0)
1522
1523 #define vmaxq_u16(__a, __b) \
1524   (uint16x8_t)__builtin_neon_vmaxv8hi (__a, __b, 0)
1525
1526 #define vmaxq_u32(__a, __b) \
1527   (uint32x4_t)__builtin_neon_vmaxv4si (__a, __b, 0)
1528
1529 #define vmin_s8(__a, __b) \
1530   (int8x8_t)__builtin_neon_vminv8qi (__a, __b, 1)
1531
1532 #define vmin_s16(__a, __b) \
1533   (int16x4_t)__builtin_neon_vminv4hi (__a, __b, 1)
1534
1535 #define vmin_s32(__a, __b) \
1536   (int32x2_t)__builtin_neon_vminv2si (__a, __b, 1)
1537
1538 #define vmin_f32(__a, __b) \
1539   (float32x2_t)__builtin_neon_vminv2sf (__a, __b, 5)
1540
1541 #define vmin_u8(__a, __b) \
1542   (uint8x8_t)__builtin_neon_vminv8qi (__a, __b, 0)
1543
1544 #define vmin_u16(__a, __b) \
1545   (uint16x4_t)__builtin_neon_vminv4hi (__a, __b, 0)
1546
1547 #define vmin_u32(__a, __b) \
1548   (uint32x2_t)__builtin_neon_vminv2si (__a, __b, 0)
1549
1550 #define vminq_s8(__a, __b) \
1551   (int8x16_t)__builtin_neon_vminv16qi (__a, __b, 1)
1552
1553 #define vminq_s16(__a, __b) \
1554   (int16x8_t)__builtin_neon_vminv8hi (__a, __b, 1)
1555
1556 #define vminq_s32(__a, __b) \
1557   (int32x4_t)__builtin_neon_vminv4si (__a, __b, 1)
1558
1559 #define vminq_f32(__a, __b) \
1560   (float32x4_t)__builtin_neon_vminv4sf (__a, __b, 5)
1561
1562 #define vminq_u8(__a, __b) \
1563   (uint8x16_t)__builtin_neon_vminv16qi (__a, __b, 0)
1564
1565 #define vminq_u16(__a, __b) \
1566   (uint16x8_t)__builtin_neon_vminv8hi (__a, __b, 0)
1567
1568 #define vminq_u32(__a, __b) \
1569   (uint32x4_t)__builtin_neon_vminv4si (__a, __b, 0)
1570
1571 #define vpadd_s8(__a, __b) \
1572   (int8x8_t)__builtin_neon_vpaddv8qi (__a, __b, 1)
1573
1574 #define vpadd_s16(__a, __b) \
1575   (int16x4_t)__builtin_neon_vpaddv4hi (__a, __b, 1)
1576
1577 #define vpadd_s32(__a, __b) \
1578   (int32x2_t)__builtin_neon_vpaddv2si (__a, __b, 1)
1579
1580 #define vpadd_f32(__a, __b) \
1581   (float32x2_t)__builtin_neon_vpaddv2sf (__a, __b, 5)
1582
1583 #define vpadd_u8(__a, __b) \
1584   (uint8x8_t)__builtin_neon_vpaddv8qi (__a, __b, 0)
1585
1586 #define vpadd_u16(__a, __b) \
1587   (uint16x4_t)__builtin_neon_vpaddv4hi (__a, __b, 0)
1588
1589 #define vpadd_u32(__a, __b) \
1590   (uint32x2_t)__builtin_neon_vpaddv2si (__a, __b, 0)
1591
1592 #define vpaddl_s8(__a) \
1593   (int16x4_t)__builtin_neon_vpaddlv8qi (__a, 1)
1594
1595 #define vpaddl_s16(__a) \
1596   (int32x2_t)__builtin_neon_vpaddlv4hi (__a, 1)
1597
1598 #define vpaddl_s32(__a) \
1599   (int64x1_t)__builtin_neon_vpaddlv2si (__a, 1)
1600
1601 #define vpaddl_u8(__a) \
1602   (uint16x4_t)__builtin_neon_vpaddlv8qi (__a, 0)
1603
1604 #define vpaddl_u16(__a) \
1605   (uint32x2_t)__builtin_neon_vpaddlv4hi (__a, 0)
1606
1607 #define vpaddl_u32(__a) \
1608   (uint64x1_t)__builtin_neon_vpaddlv2si (__a, 0)
1609
1610 #define vpaddlq_s8(__a) \
1611   (int16x8_t)__builtin_neon_vpaddlv16qi (__a, 1)
1612
1613 #define vpaddlq_s16(__a) \
1614   (int32x4_t)__builtin_neon_vpaddlv8hi (__a, 1)
1615
1616 #define vpaddlq_s32(__a) \
1617   (int64x2_t)__builtin_neon_vpaddlv4si (__a, 1)
1618
1619 #define vpaddlq_u8(__a) \
1620   (uint16x8_t)__builtin_neon_vpaddlv16qi (__a, 0)
1621
1622 #define vpaddlq_u16(__a) \
1623   (uint32x4_t)__builtin_neon_vpaddlv8hi (__a, 0)
1624
1625 #define vpaddlq_u32(__a) \
1626   (uint64x2_t)__builtin_neon_vpaddlv4si (__a, 0)
1627
1628 #define vpadal_s8(__a, __b) \
1629   (int16x4_t)__builtin_neon_vpadalv8qi (__a, __b, 1)
1630
1631 #define vpadal_s16(__a, __b) \
1632   (int32x2_t)__builtin_neon_vpadalv4hi (__a, __b, 1)
1633
1634 #define vpadal_s32(__a, __b) \
1635   (int64x1_t)__builtin_neon_vpadalv2si (__a, __b, 1)
1636
1637 #define vpadal_u8(__a, __b) \
1638   (uint16x4_t)__builtin_neon_vpadalv8qi (__a, __b, 0)
1639
1640 #define vpadal_u16(__a, __b) \
1641   (uint32x2_t)__builtin_neon_vpadalv4hi (__a, __b, 0)
1642
1643 #define vpadal_u32(__a, __b) \
1644   (uint64x1_t)__builtin_neon_vpadalv2si (__a, __b, 0)
1645
1646 #define vpadalq_s8(__a, __b) \
1647   (int16x8_t)__builtin_neon_vpadalv16qi (__a, __b, 1)
1648
1649 #define vpadalq_s16(__a, __b) \
1650   (int32x4_t)__builtin_neon_vpadalv8hi (__a, __b, 1)
1651
1652 #define vpadalq_s32(__a, __b) \
1653   (int64x2_t)__builtin_neon_vpadalv4si (__a, __b, 1)
1654
1655 #define vpadalq_u8(__a, __b) \
1656   (uint16x8_t)__builtin_neon_vpadalv16qi (__a, __b, 0)
1657
1658 #define vpadalq_u16(__a, __b) \
1659   (uint32x4_t)__builtin_neon_vpadalv8hi (__a, __b, 0)
1660
1661 #define vpadalq_u32(__a, __b) \
1662   (uint64x2_t)__builtin_neon_vpadalv4si (__a, __b, 0)
1663
1664 #define vpmax_s8(__a, __b) \
1665   (int8x8_t)__builtin_neon_vpmaxv8qi (__a, __b, 1)
1666
1667 #define vpmax_s16(__a, __b) \
1668   (int16x4_t)__builtin_neon_vpmaxv4hi (__a, __b, 1)
1669
1670 #define vpmax_s32(__a, __b) \
1671   (int32x2_t)__builtin_neon_vpmaxv2si (__a, __b, 1)
1672
1673 #define vpmax_f32(__a, __b) \
1674   (float32x2_t)__builtin_neon_vpmaxv2sf (__a, __b, 5)
1675
1676 #define vpmax_u8(__a, __b) \
1677   (uint8x8_t)__builtin_neon_vpmaxv8qi (__a, __b, 0)
1678
1679 #define vpmax_u16(__a, __b) \
1680   (uint16x4_t)__builtin_neon_vpmaxv4hi (__a, __b, 0)
1681
1682 #define vpmax_u32(__a, __b) \
1683   (uint32x2_t)__builtin_neon_vpmaxv2si (__a, __b, 0)
1684
1685 #define vpmin_s8(__a, __b) \
1686   (int8x8_t)__builtin_neon_vpminv8qi (__a, __b, 1)
1687
1688 #define vpmin_s16(__a, __b) \
1689   (int16x4_t)__builtin_neon_vpminv4hi (__a, __b, 1)
1690
1691 #define vpmin_s32(__a, __b) \
1692   (int32x2_t)__builtin_neon_vpminv2si (__a, __b, 1)
1693
1694 #define vpmin_f32(__a, __b) \
1695   (float32x2_t)__builtin_neon_vpminv2sf (__a, __b, 5)
1696
1697 #define vpmin_u8(__a, __b) \
1698   (uint8x8_t)__builtin_neon_vpminv8qi (__a, __b, 0)
1699
1700 #define vpmin_u16(__a, __b) \
1701   (uint16x4_t)__builtin_neon_vpminv4hi (__a, __b, 0)
1702
1703 #define vpmin_u32(__a, __b) \
1704   (uint32x2_t)__builtin_neon_vpminv2si (__a, __b, 0)
1705
1706 #define vrecps_f32(__a, __b) \
1707   (float32x2_t)__builtin_neon_vrecpsv2sf (__a, __b, 5)
1708
1709 #define vrecpsq_f32(__a, __b) \
1710   (float32x4_t)__builtin_neon_vrecpsv4sf (__a, __b, 5)
1711
1712 #define vrsqrts_f32(__a, __b) \
1713   (float32x2_t)__builtin_neon_vrsqrtsv2sf (__a, __b, 5)
1714
1715 #define vrsqrtsq_f32(__a, __b) \
1716   (float32x4_t)__builtin_neon_vrsqrtsv4sf (__a, __b, 5)
1717
1718 #define vshl_s8(__a, __b) \
1719   (int8x8_t)__builtin_neon_vshlv8qi (__a, __b, 1)
1720
1721 #define vshl_s16(__a, __b) \
1722   (int16x4_t)__builtin_neon_vshlv4hi (__a, __b, 1)
1723
1724 #define vshl_s32(__a, __b) \
1725   (int32x2_t)__builtin_neon_vshlv2si (__a, __b, 1)
1726
1727 #define vshl_s64(__a, __b) \
1728   (int64x1_t)__builtin_neon_vshlv1di (__a, __b, 1)
1729
1730 #define vshl_u8(__a, __b) \
1731   (uint8x8_t)__builtin_neon_vshlv8qi (__a, __b, 0)
1732
1733 #define vshl_u16(__a, __b) \
1734   (uint16x4_t)__builtin_neon_vshlv4hi (__a, __b, 0)
1735
1736 #define vshl_u32(__a, __b) \
1737   (uint32x2_t)__builtin_neon_vshlv2si (__a, __b, 0)
1738
1739 #define vshl_u64(__a, __b) \
1740   (uint64x1_t)__builtin_neon_vshlv1di (__a, __b, 0)
1741
1742 #define vshlq_s8(__a, __b) \
1743   (int8x16_t)__builtin_neon_vshlv16qi (__a, __b, 1)
1744
1745 #define vshlq_s16(__a, __b) \
1746   (int16x8_t)__builtin_neon_vshlv8hi (__a, __b, 1)
1747
1748 #define vshlq_s32(__a, __b) \
1749   (int32x4_t)__builtin_neon_vshlv4si (__a, __b, 1)
1750
1751 #define vshlq_s64(__a, __b) \
1752   (int64x2_t)__builtin_neon_vshlv2di (__a, __b, 1)
1753
1754 #define vshlq_u8(__a, __b) \
1755   (uint8x16_t)__builtin_neon_vshlv16qi (__a, __b, 0)
1756
1757 #define vshlq_u16(__a, __b) \
1758   (uint16x8_t)__builtin_neon_vshlv8hi (__a, __b, 0)
1759
1760 #define vshlq_u32(__a, __b) \
1761   (uint32x4_t)__builtin_neon_vshlv4si (__a, __b, 0)
1762
1763 #define vshlq_u64(__a, __b) \
1764   (uint64x2_t)__builtin_neon_vshlv2di (__a, __b, 0)
1765
1766 #define vrshl_s8(__a, __b) \
1767   (int8x8_t)__builtin_neon_vshlv8qi (__a, __b, 3)
1768
1769 #define vrshl_s16(__a, __b) \
1770   (int16x4_t)__builtin_neon_vshlv4hi (__a, __b, 3)
1771
1772 #define vrshl_s32(__a, __b) \
1773   (int32x2_t)__builtin_neon_vshlv2si (__a, __b, 3)
1774
1775 #define vrshl_s64(__a, __b) \
1776   (int64x1_t)__builtin_neon_vshlv1di (__a, __b, 3)
1777
1778 #define vrshl_u8(__a, __b) \
1779   (uint8x8_t)__builtin_neon_vshlv8qi (__a, __b, 2)
1780
1781 #define vrshl_u16(__a, __b) \
1782   (uint16x4_t)__builtin_neon_vshlv4hi (__a, __b, 2)
1783
1784 #define vrshl_u32(__a, __b) \
1785   (uint32x2_t)__builtin_neon_vshlv2si (__a, __b, 2)
1786
1787 #define vrshl_u64(__a, __b) \
1788   (uint64x1_t)__builtin_neon_vshlv1di (__a, __b, 2)
1789
1790 #define vrshlq_s8(__a, __b) \
1791   (int8x16_t)__builtin_neon_vshlv16qi (__a, __b, 3)
1792
1793 #define vrshlq_s16(__a, __b) \
1794   (int16x8_t)__builtin_neon_vshlv8hi (__a, __b, 3)
1795
1796 #define vrshlq_s32(__a, __b) \
1797   (int32x4_t)__builtin_neon_vshlv4si (__a, __b, 3)
1798
1799 #define vrshlq_s64(__a, __b) \
1800   (int64x2_t)__builtin_neon_vshlv2di (__a, __b, 3)
1801
1802 #define vrshlq_u8(__a, __b) \
1803   (uint8x16_t)__builtin_neon_vshlv16qi (__a, __b, 2)
1804
1805 #define vrshlq_u16(__a, __b) \
1806   (uint16x8_t)__builtin_neon_vshlv8hi (__a, __b, 2)
1807
1808 #define vrshlq_u32(__a, __b) \
1809   (uint32x4_t)__builtin_neon_vshlv4si (__a, __b, 2)
1810
1811 #define vrshlq_u64(__a, __b) \
1812   (uint64x2_t)__builtin_neon_vshlv2di (__a, __b, 2)
1813
1814 #define vqshl_s8(__a, __b) \
1815   (int8x8_t)__builtin_neon_vqshlv8qi (__a, __b, 1)
1816
1817 #define vqshl_s16(__a, __b) \
1818   (int16x4_t)__builtin_neon_vqshlv4hi (__a, __b, 1)
1819
1820 #define vqshl_s32(__a, __b) \
1821   (int32x2_t)__builtin_neon_vqshlv2si (__a, __b, 1)
1822
1823 #define vqshl_s64(__a, __b) \
1824   (int64x1_t)__builtin_neon_vqshlv1di (__a, __b, 1)
1825
1826 #define vqshl_u8(__a, __b) \
1827   (uint8x8_t)__builtin_neon_vqshlv8qi (__a, __b, 0)
1828
1829 #define vqshl_u16(__a, __b) \
1830   (uint16x4_t)__builtin_neon_vqshlv4hi (__a, __b, 0)
1831
1832 #define vqshl_u32(__a, __b) \
1833   (uint32x2_t)__builtin_neon_vqshlv2si (__a, __b, 0)
1834
1835 #define vqshl_u64(__a, __b) \
1836   (uint64x1_t)__builtin_neon_vqshlv1di (__a, __b, 0)
1837
1838 #define vqshlq_s8(__a, __b) \
1839   (int8x16_t)__builtin_neon_vqshlv16qi (__a, __b, 1)
1840
1841 #define vqshlq_s16(__a, __b) \
1842   (int16x8_t)__builtin_neon_vqshlv8hi (__a, __b, 1)
1843
1844 #define vqshlq_s32(__a, __b) \
1845   (int32x4_t)__builtin_neon_vqshlv4si (__a, __b, 1)
1846
1847 #define vqshlq_s64(__a, __b) \
1848   (int64x2_t)__builtin_neon_vqshlv2di (__a, __b, 1)
1849
1850 #define vqshlq_u8(__a, __b) \
1851   (uint8x16_t)__builtin_neon_vqshlv16qi (__a, __b, 0)
1852
1853 #define vqshlq_u16(__a, __b) \
1854   (uint16x8_t)__builtin_neon_vqshlv8hi (__a, __b, 0)
1855
1856 #define vqshlq_u32(__a, __b) \
1857   (uint32x4_t)__builtin_neon_vqshlv4si (__a, __b, 0)
1858
1859 #define vqshlq_u64(__a, __b) \
1860   (uint64x2_t)__builtin_neon_vqshlv2di (__a, __b, 0)
1861
1862 #define vqrshl_s8(__a, __b) \
1863   (int8x8_t)__builtin_neon_vqshlv8qi (__a, __b, 3)
1864
1865 #define vqrshl_s16(__a, __b) \
1866   (int16x4_t)__builtin_neon_vqshlv4hi (__a, __b, 3)
1867
1868 #define vqrshl_s32(__a, __b) \
1869   (int32x2_t)__builtin_neon_vqshlv2si (__a, __b, 3)
1870
1871 #define vqrshl_s64(__a, __b) \
1872   (int64x1_t)__builtin_neon_vqshlv1di (__a, __b, 3)
1873
1874 #define vqrshl_u8(__a, __b) \
1875   (uint8x8_t)__builtin_neon_vqshlv8qi (__a, __b, 2)
1876
1877 #define vqrshl_u16(__a, __b) \
1878   (uint16x4_t)__builtin_neon_vqshlv4hi (__a, __b, 2)
1879
1880 #define vqrshl_u32(__a, __b) \
1881   (uint32x2_t)__builtin_neon_vqshlv2si (__a, __b, 2)
1882
1883 #define vqrshl_u64(__a, __b) \
1884   (uint64x1_t)__builtin_neon_vqshlv1di (__a, __b, 2)
1885
1886 #define vqrshlq_s8(__a, __b) \
1887   (int8x16_t)__builtin_neon_vqshlv16qi (__a, __b, 3)
1888
1889 #define vqrshlq_s16(__a, __b) \
1890   (int16x8_t)__builtin_neon_vqshlv8hi (__a, __b, 3)
1891
1892 #define vqrshlq_s32(__a, __b) \
1893   (int32x4_t)__builtin_neon_vqshlv4si (__a, __b, 3)
1894
1895 #define vqrshlq_s64(__a, __b) \
1896   (int64x2_t)__builtin_neon_vqshlv2di (__a, __b, 3)
1897
1898 #define vqrshlq_u8(__a, __b) \
1899   (uint8x16_t)__builtin_neon_vqshlv16qi (__a, __b, 2)
1900
1901 #define vqrshlq_u16(__a, __b) \
1902   (uint16x8_t)__builtin_neon_vqshlv8hi (__a, __b, 2)
1903
1904 #define vqrshlq_u32(__a, __b) \
1905   (uint32x4_t)__builtin_neon_vqshlv4si (__a, __b, 2)
1906
1907 #define vqrshlq_u64(__a, __b) \
1908   (uint64x2_t)__builtin_neon_vqshlv2di (__a, __b, 2)
1909
1910 #define vshr_n_s8(__a, __b) \
1911   (int8x8_t)__builtin_neon_vshr_nv8qi (__a, __b, 1)
1912
1913 #define vshr_n_s16(__a, __b) \
1914   (int16x4_t)__builtin_neon_vshr_nv4hi (__a, __b, 1)
1915
1916 #define vshr_n_s32(__a, __b) \
1917   (int32x2_t)__builtin_neon_vshr_nv2si (__a, __b, 1)
1918
1919 #define vshr_n_s64(__a, __b) \
1920   (int64x1_t)__builtin_neon_vshr_nv1di (__a, __b, 1)
1921
1922 #define vshr_n_u8(__a, __b) \
1923   (uint8x8_t)__builtin_neon_vshr_nv8qi (__a, __b, 0)
1924
1925 #define vshr_n_u16(__a, __b) \
1926   (uint16x4_t)__builtin_neon_vshr_nv4hi (__a, __b, 0)
1927
1928 #define vshr_n_u32(__a, __b) \
1929   (uint32x2_t)__builtin_neon_vshr_nv2si (__a, __b, 0)
1930
1931 #define vshr_n_u64(__a, __b) \
1932   (uint64x1_t)__builtin_neon_vshr_nv1di (__a, __b, 0)
1933
1934 #define vshrq_n_s8(__a, __b) \
1935   (int8x16_t)__builtin_neon_vshr_nv16qi (__a, __b, 1)
1936
1937 #define vshrq_n_s16(__a, __b) \
1938   (int16x8_t)__builtin_neon_vshr_nv8hi (__a, __b, 1)
1939
1940 #define vshrq_n_s32(__a, __b) \
1941   (int32x4_t)__builtin_neon_vshr_nv4si (__a, __b, 1)
1942
1943 #define vshrq_n_s64(__a, __b) \
1944   (int64x2_t)__builtin_neon_vshr_nv2di (__a, __b, 1)
1945
1946 #define vshrq_n_u8(__a, __b) \
1947   (uint8x16_t)__builtin_neon_vshr_nv16qi (__a, __b, 0)
1948
1949 #define vshrq_n_u16(__a, __b) \
1950   (uint16x8_t)__builtin_neon_vshr_nv8hi (__a, __b, 0)
1951
1952 #define vshrq_n_u32(__a, __b) \
1953   (uint32x4_t)__builtin_neon_vshr_nv4si (__a, __b, 0)
1954
1955 #define vshrq_n_u64(__a, __b) \
1956   (uint64x2_t)__builtin_neon_vshr_nv2di (__a, __b, 0)
1957
1958 #define vrshr_n_s8(__a, __b) \
1959   (int8x8_t)__builtin_neon_vshr_nv8qi (__a, __b, 3)
1960
1961 #define vrshr_n_s16(__a, __b) \
1962   (int16x4_t)__builtin_neon_vshr_nv4hi (__a, __b, 3)
1963
1964 #define vrshr_n_s32(__a, __b) \
1965   (int32x2_t)__builtin_neon_vshr_nv2si (__a, __b, 3)
1966
1967 #define vrshr_n_s64(__a, __b) \
1968   (int64x1_t)__builtin_neon_vshr_nv1di (__a, __b, 3)
1969
1970 #define vrshr_n_u8(__a, __b) \
1971   (uint8x8_t)__builtin_neon_vshr_nv8qi (__a, __b, 2)
1972
1973 #define vrshr_n_u16(__a, __b) \
1974   (uint16x4_t)__builtin_neon_vshr_nv4hi (__a, __b, 2)
1975
1976 #define vrshr_n_u32(__a, __b) \
1977   (uint32x2_t)__builtin_neon_vshr_nv2si (__a, __b, 2)
1978
1979 #define vrshr_n_u64(__a, __b) \
1980   (uint64x1_t)__builtin_neon_vshr_nv1di (__a, __b, 2)
1981
1982 #define vrshrq_n_s8(__a, __b) \
1983   (int8x16_t)__builtin_neon_vshr_nv16qi (__a, __b, 3)
1984
1985 #define vrshrq_n_s16(__a, __b) \
1986   (int16x8_t)__builtin_neon_vshr_nv8hi (__a, __b, 3)
1987
1988 #define vrshrq_n_s32(__a, __b) \
1989   (int32x4_t)__builtin_neon_vshr_nv4si (__a, __b, 3)
1990
1991 #define vrshrq_n_s64(__a, __b) \
1992   (int64x2_t)__builtin_neon_vshr_nv2di (__a, __b, 3)
1993
1994 #define vrshrq_n_u8(__a, __b) \
1995   (uint8x16_t)__builtin_neon_vshr_nv16qi (__a, __b, 2)
1996
1997 #define vrshrq_n_u16(__a, __b) \
1998   (uint16x8_t)__builtin_neon_vshr_nv8hi (__a, __b, 2)
1999
2000 #define vrshrq_n_u32(__a, __b) \
2001   (uint32x4_t)__builtin_neon_vshr_nv4si (__a, __b, 2)
2002
2003 #define vrshrq_n_u64(__a, __b) \
2004   (uint64x2_t)__builtin_neon_vshr_nv2di (__a, __b, 2)
2005
2006 #define vshrn_n_s16(__a, __b) \
2007   (int8x8_t)__builtin_neon_vshrn_nv8hi (__a, __b, 1)
2008
2009 #define vshrn_n_s32(__a, __b) \
2010   (int16x4_t)__builtin_neon_vshrn_nv4si (__a, __b, 1)
2011
2012 #define vshrn_n_s64(__a, __b) \
2013   (int32x2_t)__builtin_neon_vshrn_nv2di (__a, __b, 1)
2014
2015 #define vshrn_n_u16(__a, __b) \
2016   (uint8x8_t)__builtin_neon_vshrn_nv8hi (__a, __b, 0)
2017
2018 #define vshrn_n_u32(__a, __b) \
2019   (uint16x4_t)__builtin_neon_vshrn_nv4si (__a, __b, 0)
2020
2021 #define vshrn_n_u64(__a, __b) \
2022   (uint32x2_t)__builtin_neon_vshrn_nv2di (__a, __b, 0)
2023
2024 #define vrshrn_n_s16(__a, __b) \
2025   (int8x8_t)__builtin_neon_vshrn_nv8hi (__a, __b, 3)
2026
2027 #define vrshrn_n_s32(__a, __b) \
2028   (int16x4_t)__builtin_neon_vshrn_nv4si (__a, __b, 3)
2029
2030 #define vrshrn_n_s64(__a, __b) \
2031   (int32x2_t)__builtin_neon_vshrn_nv2di (__a, __b, 3)
2032
2033 #define vrshrn_n_u16(__a, __b) \
2034   (uint8x8_t)__builtin_neon_vshrn_nv8hi (__a, __b, 2)
2035
2036 #define vrshrn_n_u32(__a, __b) \
2037   (uint16x4_t)__builtin_neon_vshrn_nv4si (__a, __b, 2)
2038
2039 #define vrshrn_n_u64(__a, __b) \
2040   (uint32x2_t)__builtin_neon_vshrn_nv2di (__a, __b, 2)
2041
2042 #define vqshrn_n_s16(__a, __b) \
2043   (int8x8_t)__builtin_neon_vqshrn_nv8hi (__a, __b, 1)
2044
2045 #define vqshrn_n_s32(__a, __b) \
2046   (int16x4_t)__builtin_neon_vqshrn_nv4si (__a, __b, 1)
2047
2048 #define vqshrn_n_s64(__a, __b) \
2049   (int32x2_t)__builtin_neon_vqshrn_nv2di (__a, __b, 1)
2050
2051 #define vqshrn_n_u16(__a, __b) \
2052   (uint8x8_t)__builtin_neon_vqshrn_nv8hi (__a, __b, 0)
2053
2054 #define vqshrn_n_u32(__a, __b) \
2055   (uint16x4_t)__builtin_neon_vqshrn_nv4si (__a, __b, 0)
2056
2057 #define vqshrn_n_u64(__a, __b) \
2058   (uint32x2_t)__builtin_neon_vqshrn_nv2di (__a, __b, 0)
2059
2060 #define vqrshrn_n_s16(__a, __b) \
2061   (int8x8_t)__builtin_neon_vqshrn_nv8hi (__a, __b, 3)
2062
2063 #define vqrshrn_n_s32(__a, __b) \
2064   (int16x4_t)__builtin_neon_vqshrn_nv4si (__a, __b, 3)
2065
2066 #define vqrshrn_n_s64(__a, __b) \
2067   (int32x2_t)__builtin_neon_vqshrn_nv2di (__a, __b, 3)
2068
2069 #define vqrshrn_n_u16(__a, __b) \
2070   (uint8x8_t)__builtin_neon_vqshrn_nv8hi (__a, __b, 2)
2071
2072 #define vqrshrn_n_u32(__a, __b) \
2073   (uint16x4_t)__builtin_neon_vqshrn_nv4si (__a, __b, 2)
2074
2075 #define vqrshrn_n_u64(__a, __b) \
2076   (uint32x2_t)__builtin_neon_vqshrn_nv2di (__a, __b, 2)
2077
2078 #define vqshrun_n_s16(__a, __b) \
2079   (uint8x8_t)__builtin_neon_vqshrun_nv8hi (__a, __b, 1)
2080
2081 #define vqshrun_n_s32(__a, __b) \
2082   (uint16x4_t)__builtin_neon_vqshrun_nv4si (__a, __b, 1)
2083
2084 #define vqshrun_n_s64(__a, __b) \
2085   (uint32x2_t)__builtin_neon_vqshrun_nv2di (__a, __b, 1)
2086
2087 #define vqrshrun_n_s16(__a, __b) \
2088   (uint8x8_t)__builtin_neon_vqshrun_nv8hi (__a, __b, 3)
2089
2090 #define vqrshrun_n_s32(__a, __b) \
2091   (uint16x4_t)__builtin_neon_vqshrun_nv4si (__a, __b, 3)
2092
2093 #define vqrshrun_n_s64(__a, __b) \
2094   (uint32x2_t)__builtin_neon_vqshrun_nv2di (__a, __b, 3)
2095
2096 #define vshl_n_s8(__a, __b) \
2097   (int8x8_t)__builtin_neon_vshl_nv8qi (__a, __b, 1)
2098
2099 #define vshl_n_s16(__a, __b) \
2100   (int16x4_t)__builtin_neon_vshl_nv4hi (__a, __b, 1)
2101
2102 #define vshl_n_s32(__a, __b) \
2103   (int32x2_t)__builtin_neon_vshl_nv2si (__a, __b, 1)
2104
2105 #define vshl_n_s64(__a, __b) \
2106   (int64x1_t)__builtin_neon_vshl_nv1di (__a, __b, 1)
2107
2108 #define vshl_n_u8(__a, __b) \
2109   (uint8x8_t)__builtin_neon_vshl_nv8qi (__a, __b, 0)
2110
2111 #define vshl_n_u16(__a, __b) \
2112   (uint16x4_t)__builtin_neon_vshl_nv4hi (__a, __b, 0)
2113
2114 #define vshl_n_u32(__a, __b) \
2115   (uint32x2_t)__builtin_neon_vshl_nv2si (__a, __b, 0)
2116
2117 #define vshl_n_u64(__a, __b) \
2118   (uint64x1_t)__builtin_neon_vshl_nv1di (__a, __b, 0)
2119
2120 #define vshlq_n_s8(__a, __b) \
2121   (int8x16_t)__builtin_neon_vshl_nv16qi (__a, __b, 1)
2122
2123 #define vshlq_n_s16(__a, __b) \
2124   (int16x8_t)__builtin_neon_vshl_nv8hi (__a, __b, 1)
2125
2126 #define vshlq_n_s32(__a, __b) \
2127   (int32x4_t)__builtin_neon_vshl_nv4si (__a, __b, 1)
2128
2129 #define vshlq_n_s64(__a, __b) \
2130   (int64x2_t)__builtin_neon_vshl_nv2di (__a, __b, 1)
2131
2132 #define vshlq_n_u8(__a, __b) \
2133   (uint8x16_t)__builtin_neon_vshl_nv16qi (__a, __b, 0)
2134
2135 #define vshlq_n_u16(__a, __b) \
2136   (uint16x8_t)__builtin_neon_vshl_nv8hi (__a, __b, 0)
2137
2138 #define vshlq_n_u32(__a, __b) \
2139   (uint32x4_t)__builtin_neon_vshl_nv4si (__a, __b, 0)
2140
2141 #define vshlq_n_u64(__a, __b) \
2142   (uint64x2_t)__builtin_neon_vshl_nv2di (__a, __b, 0)
2143
2144 #define vqshl_n_s8(__a, __b) \
2145   (int8x8_t)__builtin_neon_vqshl_nv8qi (__a, __b, 1)
2146
2147 #define vqshl_n_s16(__a, __b) \
2148   (int16x4_t)__builtin_neon_vqshl_nv4hi (__a, __b, 1)
2149
2150 #define vqshl_n_s32(__a, __b) \
2151   (int32x2_t)__builtin_neon_vqshl_nv2si (__a, __b, 1)
2152
2153 #define vqshl_n_s64(__a, __b) \
2154   (int64x1_t)__builtin_neon_vqshl_nv1di (__a, __b, 1)
2155
2156 #define vqshl_n_u8(__a, __b) \
2157   (uint8x8_t)__builtin_neon_vqshl_nv8qi (__a, __b, 0)
2158
2159 #define vqshl_n_u16(__a, __b) \
2160   (uint16x4_t)__builtin_neon_vqshl_nv4hi (__a, __b, 0)
2161
2162 #define vqshl_n_u32(__a, __b) \
2163   (uint32x2_t)__builtin_neon_vqshl_nv2si (__a, __b, 0)
2164
2165 #define vqshl_n_u64(__a, __b) \
2166   (uint64x1_t)__builtin_neon_vqshl_nv1di (__a, __b, 0)
2167
2168 #define vqshlq_n_s8(__a, __b) \
2169   (int8x16_t)__builtin_neon_vqshl_nv16qi (__a, __b, 1)
2170
2171 #define vqshlq_n_s16(__a, __b) \
2172   (int16x8_t)__builtin_neon_vqshl_nv8hi (__a, __b, 1)
2173
2174 #define vqshlq_n_s32(__a, __b) \
2175   (int32x4_t)__builtin_neon_vqshl_nv4si (__a, __b, 1)
2176
2177 #define vqshlq_n_s64(__a, __b) \
2178   (int64x2_t)__builtin_neon_vqshl_nv2di (__a, __b, 1)
2179
2180 #define vqshlq_n_u8(__a, __b) \
2181   (uint8x16_t)__builtin_neon_vqshl_nv16qi (__a, __b, 0)
2182
2183 #define vqshlq_n_u16(__a, __b) \
2184   (uint16x8_t)__builtin_neon_vqshl_nv8hi (__a, __b, 0)
2185
2186 #define vqshlq_n_u32(__a, __b) \
2187   (uint32x4_t)__builtin_neon_vqshl_nv4si (__a, __b, 0)
2188
2189 #define vqshlq_n_u64(__a, __b) \
2190   (uint64x2_t)__builtin_neon_vqshl_nv2di (__a, __b, 0)
2191
2192 #define vqshlu_n_s8(__a, __b) \
2193   (uint8x8_t)__builtin_neon_vqshlu_nv8qi (__a, __b, 1)
2194
2195 #define vqshlu_n_s16(__a, __b) \
2196   (uint16x4_t)__builtin_neon_vqshlu_nv4hi (__a, __b, 1)
2197
2198 #define vqshlu_n_s32(__a, __b) \
2199   (uint32x2_t)__builtin_neon_vqshlu_nv2si (__a, __b, 1)
2200
2201 #define vqshlu_n_s64(__a, __b) \
2202   (uint64x1_t)__builtin_neon_vqshlu_nv1di (__a, __b, 1)
2203
2204 #define vqshluq_n_s8(__a, __b) \
2205   (uint8x16_t)__builtin_neon_vqshlu_nv16qi (__a, __b, 1)
2206
2207 #define vqshluq_n_s16(__a, __b) \
2208   (uint16x8_t)__builtin_neon_vqshlu_nv8hi (__a, __b, 1)
2209
2210 #define vqshluq_n_s32(__a, __b) \
2211   (uint32x4_t)__builtin_neon_vqshlu_nv4si (__a, __b, 1)
2212
2213 #define vqshluq_n_s64(__a, __b) \
2214   (uint64x2_t)__builtin_neon_vqshlu_nv2di (__a, __b, 1)
2215
2216 #define vshll_n_s8(__a, __b) \
2217   (int16x8_t)__builtin_neon_vshll_nv8qi (__a, __b, 1)
2218
2219 #define vshll_n_s16(__a, __b) \
2220   (int32x4_t)__builtin_neon_vshll_nv4hi (__a, __b, 1)
2221
2222 #define vshll_n_s32(__a, __b) \
2223   (int64x2_t)__builtin_neon_vshll_nv2si (__a, __b, 1)
2224
2225 #define vshll_n_u8(__a, __b) \
2226   (uint16x8_t)__builtin_neon_vshll_nv8qi (__a, __b, 0)
2227
2228 #define vshll_n_u16(__a, __b) \
2229   (uint32x4_t)__builtin_neon_vshll_nv4hi (__a, __b, 0)
2230
2231 #define vshll_n_u32(__a, __b) \
2232   (uint64x2_t)__builtin_neon_vshll_nv2si (__a, __b, 0)
2233
2234 #define vsra_n_s8(__a, __b, __c) \
2235   (int8x8_t)__builtin_neon_vsra_nv8qi (__a, __b, __c, 1)
2236
2237 #define vsra_n_s16(__a, __b, __c) \
2238   (int16x4_t)__builtin_neon_vsra_nv4hi (__a, __b, __c, 1)
2239
2240 #define vsra_n_s32(__a, __b, __c) \
2241   (int32x2_t)__builtin_neon_vsra_nv2si (__a, __b, __c, 1)
2242
2243 #define vsra_n_s64(__a, __b, __c) \
2244   (int64x1_t)__builtin_neon_vsra_nv1di (__a, __b, __c, 1)
2245
2246 #define vsra_n_u8(__a, __b, __c) \
2247   (uint8x8_t)__builtin_neon_vsra_nv8qi (__a, __b, __c, 0)
2248
2249 #define vsra_n_u16(__a, __b, __c) \
2250   (uint16x4_t)__builtin_neon_vsra_nv4hi (__a, __b, __c, 0)
2251
2252 #define vsra_n_u32(__a, __b, __c) \
2253   (uint32x2_t)__builtin_neon_vsra_nv2si (__a, __b, __c, 0)
2254
2255 #define vsra_n_u64(__a, __b, __c) \
2256   (uint64x1_t)__builtin_neon_vsra_nv1di (__a, __b, __c, 0)
2257
2258 #define vsraq_n_s8(__a, __b, __c) \
2259   (int8x16_t)__builtin_neon_vsra_nv16qi (__a, __b, __c, 1)
2260
2261 #define vsraq_n_s16(__a, __b, __c) \
2262   (int16x8_t)__builtin_neon_vsra_nv8hi (__a, __b, __c, 1)
2263
2264 #define vsraq_n_s32(__a, __b, __c) \
2265   (int32x4_t)__builtin_neon_vsra_nv4si (__a, __b, __c, 1)
2266
2267 #define vsraq_n_s64(__a, __b, __c) \
2268   (int64x2_t)__builtin_neon_vsra_nv2di (__a, __b, __c, 1)
2269
2270 #define vsraq_n_u8(__a, __b, __c) \
2271   (uint8x16_t)__builtin_neon_vsra_nv16qi (__a, __b, __c, 0)
2272
2273 #define vsraq_n_u16(__a, __b, __c) \
2274   (uint16x8_t)__builtin_neon_vsra_nv8hi (__a, __b, __c, 0)
2275
2276 #define vsraq_n_u32(__a, __b, __c) \
2277   (uint32x4_t)__builtin_neon_vsra_nv4si (__a, __b, __c, 0)
2278
2279 #define vsraq_n_u64(__a, __b, __c) \
2280   (uint64x2_t)__builtin_neon_vsra_nv2di (__a, __b, __c, 0)
2281
2282 #define vrsra_n_s8(__a, __b, __c) \
2283   (int8x8_t)__builtin_neon_vsra_nv8qi (__a, __b, __c, 3)
2284
2285 #define vrsra_n_s16(__a, __b, __c) \
2286   (int16x4_t)__builtin_neon_vsra_nv4hi (__a, __b, __c, 3)
2287
2288 #define vrsra_n_s32(__a, __b, __c) \
2289   (int32x2_t)__builtin_neon_vsra_nv2si (__a, __b, __c, 3)
2290
2291 #define vrsra_n_s64(__a, __b, __c) \
2292   (int64x1_t)__builtin_neon_vsra_nv1di (__a, __b, __c, 3)
2293
2294 #define vrsra_n_u8(__a, __b, __c) \
2295   (uint8x8_t)__builtin_neon_vsra_nv8qi (__a, __b, __c, 2)
2296
2297 #define vrsra_n_u16(__a, __b, __c) \
2298   (uint16x4_t)__builtin_neon_vsra_nv4hi (__a, __b, __c, 2)
2299
2300 #define vrsra_n_u32(__a, __b, __c) \
2301   (uint32x2_t)__builtin_neon_vsra_nv2si (__a, __b, __c, 2)
2302
2303 #define vrsra_n_u64(__a, __b, __c) \
2304   (uint64x1_t)__builtin_neon_vsra_nv1di (__a, __b, __c, 2)
2305
2306 #define vrsraq_n_s8(__a, __b, __c) \
2307   (int8x16_t)__builtin_neon_vsra_nv16qi (__a, __b, __c, 3)
2308
2309 #define vrsraq_n_s16(__a, __b, __c) \
2310   (int16x8_t)__builtin_neon_vsra_nv8hi (__a, __b, __c, 3)
2311
2312 #define vrsraq_n_s32(__a, __b, __c) \
2313   (int32x4_t)__builtin_neon_vsra_nv4si (__a, __b, __c, 3)
2314
2315 #define vrsraq_n_s64(__a, __b, __c) \
2316   (int64x2_t)__builtin_neon_vsra_nv2di (__a, __b, __c, 3)
2317
2318 #define vrsraq_n_u8(__a, __b, __c) \
2319   (uint8x16_t)__builtin_neon_vsra_nv16qi (__a, __b, __c, 2)
2320
2321 #define vrsraq_n_u16(__a, __b, __c) \
2322   (uint16x8_t)__builtin_neon_vsra_nv8hi (__a, __b, __c, 2)
2323
2324 #define vrsraq_n_u32(__a, __b, __c) \
2325   (uint32x4_t)__builtin_neon_vsra_nv4si (__a, __b, __c, 2)
2326
2327 #define vrsraq_n_u64(__a, __b, __c) \
2328   (uint64x2_t)__builtin_neon_vsra_nv2di (__a, __b, __c, 2)
2329
2330 #define vsri_n_s8(__a, __b, __c) \
2331   (int8x8_t)__builtin_neon_vsri_nv8qi (__a, __b, __c)
2332
2333 #define vsri_n_s16(__a, __b, __c) \
2334   (int16x4_t)__builtin_neon_vsri_nv4hi (__a, __b, __c)
2335
2336 #define vsri_n_s32(__a, __b, __c) \
2337   (int32x2_t)__builtin_neon_vsri_nv2si (__a, __b, __c)
2338
2339 #define vsri_n_s64(__a, __b, __c) \
2340   (int64x1_t)__builtin_neon_vsri_nv1di (__a, __b, __c)
2341
2342 #define vsri_n_u8(__a, __b, __c) \
2343   (uint8x8_t)__builtin_neon_vsri_nv8qi (__a, __b, __c)
2344
2345 #define vsri_n_u16(__a, __b, __c) \
2346   (uint16x4_t)__builtin_neon_vsri_nv4hi (__a, __b, __c)
2347
2348 #define vsri_n_u32(__a, __b, __c) \
2349   (uint32x2_t)__builtin_neon_vsri_nv2si (__a, __b, __c)
2350
2351 #define vsri_n_u64(__a, __b, __c) \
2352   (uint64x1_t)__builtin_neon_vsri_nv1di (__a, __b, __c)
2353
2354 #define vsri_n_p8(__a, __b, __c) \
2355   (poly8x8_t)__builtin_neon_vsri_nv8qi (__a, __b, __c)
2356
2357 #define vsri_n_p16(__a, __b, __c) \
2358   (poly16x4_t)__builtin_neon_vsri_nv4hi (__a, __b, __c)
2359
2360 #define vsriq_n_s8(__a, __b, __c) \
2361   (int8x16_t)__builtin_neon_vsri_nv16qi (__a, __b, __c)
2362
2363 #define vsriq_n_s16(__a, __b, __c) \
2364   (int16x8_t)__builtin_neon_vsri_nv8hi (__a, __b, __c)
2365
2366 #define vsriq_n_s32(__a, __b, __c) \
2367   (int32x4_t)__builtin_neon_vsri_nv4si (__a, __b, __c)
2368
2369 #define vsriq_n_s64(__a, __b, __c) \
2370   (int64x2_t)__builtin_neon_vsri_nv2di (__a, __b, __c)
2371
2372 #define vsriq_n_u8(__a, __b, __c) \
2373   (uint8x16_t)__builtin_neon_vsri_nv16qi (__a, __b, __c)
2374
2375 #define vsriq_n_u16(__a, __b, __c) \
2376   (uint16x8_t)__builtin_neon_vsri_nv8hi (__a, __b, __c)
2377
2378 #define vsriq_n_u32(__a, __b, __c) \
2379   (uint32x4_t)__builtin_neon_vsri_nv4si (__a, __b, __c)
2380
2381 #define vsriq_n_u64(__a, __b, __c) \
2382   (uint64x2_t)__builtin_neon_vsri_nv2di (__a, __b, __c)
2383
2384 #define vsriq_n_p8(__a, __b, __c) \
2385   (poly8x16_t)__builtin_neon_vsri_nv16qi (__a, __b, __c)
2386
2387 #define vsriq_n_p16(__a, __b, __c) \
2388   (poly16x8_t)__builtin_neon_vsri_nv8hi (__a, __b, __c)
2389
2390 #define vsli_n_s8(__a, __b, __c) \
2391   (int8x8_t)__builtin_neon_vsli_nv8qi (__a, __b, __c)
2392
2393 #define vsli_n_s16(__a, __b, __c) \
2394   (int16x4_t)__builtin_neon_vsli_nv4hi (__a, __b, __c)
2395
2396 #define vsli_n_s32(__a, __b, __c) \
2397   (int32x2_t)__builtin_neon_vsli_nv2si (__a, __b, __c)
2398
2399 #define vsli_n_s64(__a, __b, __c) \
2400   (int64x1_t)__builtin_neon_vsli_nv1di (__a, __b, __c)
2401
2402 #define vsli_n_u8(__a, __b, __c) \
2403   (uint8x8_t)__builtin_neon_vsli_nv8qi (__a, __b, __c)
2404
2405 #define vsli_n_u16(__a, __b, __c) \
2406   (uint16x4_t)__builtin_neon_vsli_nv4hi (__a, __b, __c)
2407
2408 #define vsli_n_u32(__a, __b, __c) \
2409   (uint32x2_t)__builtin_neon_vsli_nv2si (__a, __b, __c)
2410
2411 #define vsli_n_u64(__a, __b, __c) \
2412   (uint64x1_t)__builtin_neon_vsli_nv1di (__a, __b, __c)
2413
2414 #define vsli_n_p8(__a, __b, __c) \
2415   (poly8x8_t)__builtin_neon_vsli_nv8qi (__a, __b, __c)
2416
2417 #define vsli_n_p16(__a, __b, __c) \
2418   (poly16x4_t)__builtin_neon_vsli_nv4hi (__a, __b, __c)
2419
2420 #define vsliq_n_s8(__a, __b, __c) \
2421   (int8x16_t)__builtin_neon_vsli_nv16qi (__a, __b, __c)
2422
2423 #define vsliq_n_s16(__a, __b, __c) \
2424   (int16x8_t)__builtin_neon_vsli_nv8hi (__a, __b, __c)
2425
2426 #define vsliq_n_s32(__a, __b, __c) \
2427   (int32x4_t)__builtin_neon_vsli_nv4si (__a, __b, __c)
2428
2429 #define vsliq_n_s64(__a, __b, __c) \
2430   (int64x2_t)__builtin_neon_vsli_nv2di (__a, __b, __c)
2431
2432 #define vsliq_n_u8(__a, __b, __c) \
2433   (uint8x16_t)__builtin_neon_vsli_nv16qi (__a, __b, __c)
2434
2435 #define vsliq_n_u16(__a, __b, __c) \
2436   (uint16x8_t)__builtin_neon_vsli_nv8hi (__a, __b, __c)
2437
2438 #define vsliq_n_u32(__a, __b, __c) \
2439   (uint32x4_t)__builtin_neon_vsli_nv4si (__a, __b, __c)
2440
2441 #define vsliq_n_u64(__a, __b, __c) \
2442   (uint64x2_t)__builtin_neon_vsli_nv2di (__a, __b, __c)
2443
2444 #define vsliq_n_p8(__a, __b, __c) \
2445   (poly8x16_t)__builtin_neon_vsli_nv16qi (__a, __b, __c)
2446
2447 #define vsliq_n_p16(__a, __b, __c) \
2448   (poly16x8_t)__builtin_neon_vsli_nv8hi (__a, __b, __c)
2449
2450 #define vabs_s8(__a) \
2451   (int8x8_t)__builtin_neon_vabsv8qi (__a, 1)
2452
2453 #define vabs_s16(__a) \
2454   (int16x4_t)__builtin_neon_vabsv4hi (__a, 1)
2455
2456 #define vabs_s32(__a) \
2457   (int32x2_t)__builtin_neon_vabsv2si (__a, 1)
2458
2459 #define vabs_f32(__a) \
2460   (float32x2_t)__builtin_neon_vabsv2sf (__a, 5)
2461
2462 #define vabsq_s8(__a) \
2463   (int8x16_t)__builtin_neon_vabsv16qi (__a, 1)
2464
2465 #define vabsq_s16(__a) \
2466   (int16x8_t)__builtin_neon_vabsv8hi (__a, 1)
2467
2468 #define vabsq_s32(__a) \
2469   (int32x4_t)__builtin_neon_vabsv4si (__a, 1)
2470
2471 #define vabsq_f32(__a) \
2472   (float32x4_t)__builtin_neon_vabsv4sf (__a, 5)
2473
2474 #define vqabs_s8(__a) \
2475   (int8x8_t)__builtin_neon_vqabsv8qi (__a, 1)
2476
2477 #define vqabs_s16(__a) \
2478   (int16x4_t)__builtin_neon_vqabsv4hi (__a, 1)
2479
2480 #define vqabs_s32(__a) \
2481   (int32x2_t)__builtin_neon_vqabsv2si (__a, 1)
2482
2483 #define vqabsq_s8(__a) \
2484   (int8x16_t)__builtin_neon_vqabsv16qi (__a, 1)
2485
2486 #define vqabsq_s16(__a) \
2487   (int16x8_t)__builtin_neon_vqabsv8hi (__a, 1)
2488
2489 #define vqabsq_s32(__a) \
2490   (int32x4_t)__builtin_neon_vqabsv4si (__a, 1)
2491
2492 #define vneg_s8(__a) \
2493   (int8x8_t)__builtin_neon_vnegv8qi (__a, 1)
2494
2495 #define vneg_s16(__a) \
2496   (int16x4_t)__builtin_neon_vnegv4hi (__a, 1)
2497
2498 #define vneg_s32(__a) \
2499   (int32x2_t)__builtin_neon_vnegv2si (__a, 1)
2500
2501 #define vneg_f32(__a) \
2502   (float32x2_t)__builtin_neon_vnegv2sf (__a, 5)
2503
2504 #define vnegq_s8(__a) \
2505   (int8x16_t)__builtin_neon_vnegv16qi (__a, 1)
2506
2507 #define vnegq_s16(__a) \
2508   (int16x8_t)__builtin_neon_vnegv8hi (__a, 1)
2509
2510 #define vnegq_s32(__a) \
2511   (int32x4_t)__builtin_neon_vnegv4si (__a, 1)
2512
2513 #define vnegq_f32(__a) \
2514   (float32x4_t)__builtin_neon_vnegv4sf (__a, 5)
2515
2516 #define vqneg_s8(__a) \
2517   (int8x8_t)__builtin_neon_vqnegv8qi (__a, 1)
2518
2519 #define vqneg_s16(__a) \
2520   (int16x4_t)__builtin_neon_vqnegv4hi (__a, 1)
2521
2522 #define vqneg_s32(__a) \
2523   (int32x2_t)__builtin_neon_vqnegv2si (__a, 1)
2524
2525 #define vqnegq_s8(__a) \
2526   (int8x16_t)__builtin_neon_vqnegv16qi (__a, 1)
2527
2528 #define vqnegq_s16(__a) \
2529   (int16x8_t)__builtin_neon_vqnegv8hi (__a, 1)
2530
2531 #define vqnegq_s32(__a) \
2532   (int32x4_t)__builtin_neon_vqnegv4si (__a, 1)
2533
2534 #define vmvn_s8(__a) \
2535   (int8x8_t)__builtin_neon_vmvnv8qi (__a, 1)
2536
2537 #define vmvn_s16(__a) \
2538   (int16x4_t)__builtin_neon_vmvnv4hi (__a, 1)
2539
2540 #define vmvn_s32(__a) \
2541   (int32x2_t)__builtin_neon_vmvnv2si (__a, 1)
2542
2543 #define vmvn_u8(__a) \
2544   (uint8x8_t)__builtin_neon_vmvnv8qi (__a, 0)
2545
2546 #define vmvn_u16(__a) \
2547   (uint16x4_t)__builtin_neon_vmvnv4hi (__a, 0)
2548
2549 #define vmvn_u32(__a) \
2550   (uint32x2_t)__builtin_neon_vmvnv2si (__a, 0)
2551
2552 #define vmvn_p8(__a) \
2553   (poly8x8_t)__builtin_neon_vmvnv8qi (__a, 4)
2554
2555 #define vmvnq_s8(__a) \
2556   (int8x16_t)__builtin_neon_vmvnv16qi (__a, 1)
2557
2558 #define vmvnq_s16(__a) \
2559   (int16x8_t)__builtin_neon_vmvnv8hi (__a, 1)
2560
2561 #define vmvnq_s32(__a) \
2562   (int32x4_t)__builtin_neon_vmvnv4si (__a, 1)
2563
2564 #define vmvnq_u8(__a) \
2565   (uint8x16_t)__builtin_neon_vmvnv16qi (__a, 0)
2566
2567 #define vmvnq_u16(__a) \
2568   (uint16x8_t)__builtin_neon_vmvnv8hi (__a, 0)
2569
2570 #define vmvnq_u32(__a) \
2571   (uint32x4_t)__builtin_neon_vmvnv4si (__a, 0)
2572
2573 #define vmvnq_p8(__a) \
2574   (poly8x16_t)__builtin_neon_vmvnv16qi (__a, 4)
2575
2576 #define vcls_s8(__a) \
2577   (int8x8_t)__builtin_neon_vclsv8qi (__a, 1)
2578
2579 #define vcls_s16(__a) \
2580   (int16x4_t)__builtin_neon_vclsv4hi (__a, 1)
2581
2582 #define vcls_s32(__a) \
2583   (int32x2_t)__builtin_neon_vclsv2si (__a, 1)
2584
2585 #define vclsq_s8(__a) \
2586   (int8x16_t)__builtin_neon_vclsv16qi (__a, 1)
2587
2588 #define vclsq_s16(__a) \
2589   (int16x8_t)__builtin_neon_vclsv8hi (__a, 1)
2590
2591 #define vclsq_s32(__a) \
2592   (int32x4_t)__builtin_neon_vclsv4si (__a, 1)
2593
2594 #define vclz_s8(__a) \
2595   (int8x8_t)__builtin_neon_vclzv8qi (__a, 1)
2596
2597 #define vclz_s16(__a) \
2598   (int16x4_t)__builtin_neon_vclzv4hi (__a, 1)
2599
2600 #define vclz_s32(__a) \
2601   (int32x2_t)__builtin_neon_vclzv2si (__a, 1)
2602
2603 #define vclz_u8(__a) \
2604   (uint8x8_t)__builtin_neon_vclzv8qi (__a, 0)
2605
2606 #define vclz_u16(__a) \
2607   (uint16x4_t)__builtin_neon_vclzv4hi (__a, 0)
2608
2609 #define vclz_u32(__a) \
2610   (uint32x2_t)__builtin_neon_vclzv2si (__a, 0)
2611
2612 #define vclzq_s8(__a) \
2613   (int8x16_t)__builtin_neon_vclzv16qi (__a, 1)
2614
2615 #define vclzq_s16(__a) \
2616   (int16x8_t)__builtin_neon_vclzv8hi (__a, 1)
2617
2618 #define vclzq_s32(__a) \
2619   (int32x4_t)__builtin_neon_vclzv4si (__a, 1)
2620
2621 #define vclzq_u8(__a) \
2622   (uint8x16_t)__builtin_neon_vclzv16qi (__a, 0)
2623
2624 #define vclzq_u16(__a) \
2625   (uint16x8_t)__builtin_neon_vclzv8hi (__a, 0)
2626
2627 #define vclzq_u32(__a) \
2628   (uint32x4_t)__builtin_neon_vclzv4si (__a, 0)
2629
2630 #define vcnt_s8(__a) \
2631   (int8x8_t)__builtin_neon_vcntv8qi (__a, 1)
2632
2633 #define vcnt_u8(__a) \
2634   (uint8x8_t)__builtin_neon_vcntv8qi (__a, 0)
2635
2636 #define vcnt_p8(__a) \
2637   (poly8x8_t)__builtin_neon_vcntv8qi (__a, 4)
2638
2639 #define vcntq_s8(__a) \
2640   (int8x16_t)__builtin_neon_vcntv16qi (__a, 1)
2641
2642 #define vcntq_u8(__a) \
2643   (uint8x16_t)__builtin_neon_vcntv16qi (__a, 0)
2644
2645 #define vcntq_p8(__a) \
2646   (poly8x16_t)__builtin_neon_vcntv16qi (__a, 4)
2647
2648 #define vrecpe_f32(__a) \
2649   (float32x2_t)__builtin_neon_vrecpev2sf (__a, 5)
2650
2651 #define vrecpe_u32(__a) \
2652   (uint32x2_t)__builtin_neon_vrecpev2si (__a, 0)
2653
2654 #define vrecpeq_f32(__a) \
2655   (float32x4_t)__builtin_neon_vrecpev4sf (__a, 5)
2656
2657 #define vrecpeq_u32(__a) \
2658   (uint32x4_t)__builtin_neon_vrecpev4si (__a, 0)
2659
2660 #define vrsqrte_f32(__a) \
2661   (float32x2_t)__builtin_neon_vrsqrtev2sf (__a, 5)
2662
2663 #define vrsqrte_u32(__a) \
2664   (uint32x2_t)__builtin_neon_vrsqrtev2si (__a, 0)
2665
2666 #define vrsqrteq_f32(__a) \
2667   (float32x4_t)__builtin_neon_vrsqrtev4sf (__a, 5)
2668
2669 #define vrsqrteq_u32(__a) \
2670   (uint32x4_t)__builtin_neon_vrsqrtev4si (__a, 0)
2671
2672 #define vget_lane_s8(__a, __b) \
2673   (int8_t)__builtin_neon_vget_lanev8qi (__a, __b, 1)
2674
2675 #define vget_lane_s16(__a, __b) \
2676   (int16_t)__builtin_neon_vget_lanev4hi (__a, __b, 1)
2677
2678 #define vget_lane_s32(__a, __b) \
2679   (int32_t)__builtin_neon_vget_lanev2si (__a, __b, 1)
2680
2681 #define vget_lane_f32(__a, __b) \
2682   (float32_t)__builtin_neon_vget_lanev2sf (__a, __b, 5)
2683
2684 #define vget_lane_u8(__a, __b) \
2685   (uint8_t)__builtin_neon_vget_lanev8qi (__a, __b, 0)
2686
2687 #define vget_lane_u16(__a, __b) \
2688   (uint16_t)__builtin_neon_vget_lanev4hi (__a, __b, 0)
2689
2690 #define vget_lane_u32(__a, __b) \
2691   (uint32_t)__builtin_neon_vget_lanev2si (__a, __b, 0)
2692
2693 #define vget_lane_p8(__a, __b) \
2694   (poly8_t)__builtin_neon_vget_lanev8qi (__a, __b, 4)
2695
2696 #define vget_lane_p16(__a, __b) \
2697   (poly16_t)__builtin_neon_vget_lanev4hi (__a, __b, 4)
2698
2699 #define vget_lane_s64(__a, __b) \
2700   (int64_t)__builtin_neon_vget_lanev1di (__a, __b, 1)
2701
2702 #define vget_lane_u64(__a, __b) \
2703   (uint64_t)__builtin_neon_vget_lanev1di (__a, __b, 0)
2704
2705 #define vgetq_lane_s8(__a, __b) \
2706   (int8_t)__builtin_neon_vget_lanev16qi (__a, __b, 1)
2707
2708 #define vgetq_lane_s16(__a, __b) \
2709   (int16_t)__builtin_neon_vget_lanev8hi (__a, __b, 1)
2710
2711 #define vgetq_lane_s32(__a, __b) \
2712   (int32_t)__builtin_neon_vget_lanev4si (__a, __b, 1)
2713
2714 #define vgetq_lane_f32(__a, __b) \
2715   (float32_t)__builtin_neon_vget_lanev4sf (__a, __b, 5)
2716
2717 #define vgetq_lane_u8(__a, __b) \
2718   (uint8_t)__builtin_neon_vget_lanev16qi (__a, __b, 0)
2719
2720 #define vgetq_lane_u16(__a, __b) \
2721   (uint16_t)__builtin_neon_vget_lanev8hi (__a, __b, 0)
2722
2723 #define vgetq_lane_u32(__a, __b) \
2724   (uint32_t)__builtin_neon_vget_lanev4si (__a, __b, 0)
2725
2726 #define vgetq_lane_p8(__a, __b) \
2727   (poly8_t)__builtin_neon_vget_lanev16qi (__a, __b, 4)
2728
2729 #define vgetq_lane_p16(__a, __b) \
2730   (poly16_t)__builtin_neon_vget_lanev8hi (__a, __b, 4)
2731
2732 #define vgetq_lane_s64(__a, __b) \
2733   (int64_t)__builtin_neon_vget_lanev2di (__a, __b, 1)
2734
2735 #define vgetq_lane_u64(__a, __b) \
2736   (uint64_t)__builtin_neon_vget_lanev2di (__a, __b, 0)
2737
2738 #define vset_lane_s8(__a, __b, __c) \
2739   (int8x8_t)__builtin_neon_vset_lanev8qi (__a, __b, __c)
2740
2741 #define vset_lane_s16(__a, __b, __c) \
2742   (int16x4_t)__builtin_neon_vset_lanev4hi (__a, __b, __c)
2743
2744 #define vset_lane_s32(__a, __b, __c) \
2745   (int32x2_t)__builtin_neon_vset_lanev2si (__a, __b, __c)
2746
2747 #define vset_lane_f32(__a, __b, __c) \
2748   (float32x2_t)__builtin_neon_vset_lanev2sf (__a, __b, __c)
2749
2750 #define vset_lane_u8(__a, __b, __c) \
2751   (uint8x8_t)__builtin_neon_vset_lanev8qi (__a, __b, __c)
2752
2753 #define vset_lane_u16(__a, __b, __c) \
2754   (uint16x4_t)__builtin_neon_vset_lanev4hi (__a, __b, __c)
2755
2756 #define vset_lane_u32(__a, __b, __c) \
2757   (uint32x2_t)__builtin_neon_vset_lanev2si (__a, __b, __c)
2758
2759 #define vset_lane_p8(__a, __b, __c) \
2760   (poly8x8_t)__builtin_neon_vset_lanev8qi (__a, __b, __c)
2761
2762 #define vset_lane_p16(__a, __b, __c) \
2763   (poly16x4_t)__builtin_neon_vset_lanev4hi (__a, __b, __c)
2764
2765 #define vset_lane_s64(__a, __b, __c) \
2766   (int64x1_t)__builtin_neon_vset_lanev1di (__a, __b, __c)
2767
2768 #define vset_lane_u64(__a, __b, __c) \
2769   (uint64x1_t)__builtin_neon_vset_lanev1di (__a, __b, __c)
2770
2771 #define vsetq_lane_s8(__a, __b, __c) \
2772   (int8x16_t)__builtin_neon_vset_lanev16qi (__a, __b, __c)
2773
2774 #define vsetq_lane_s16(__a, __b, __c) \
2775   (int16x8_t)__builtin_neon_vset_lanev8hi (__a, __b, __c)
2776
2777 #define vsetq_lane_s32(__a, __b, __c) \
2778   (int32x4_t)__builtin_neon_vset_lanev4si (__a, __b, __c)
2779
2780 #define vsetq_lane_f32(__a, __b, __c) \
2781   (float32x4_t)__builtin_neon_vset_lanev4sf (__a, __b, __c)
2782
2783 #define vsetq_lane_u8(__a, __b, __c) \
2784   (uint8x16_t)__builtin_neon_vset_lanev16qi (__a, __b, __c)
2785
2786 #define vsetq_lane_u16(__a, __b, __c) \
2787   (uint16x8_t)__builtin_neon_vset_lanev8hi (__a, __b, __c)
2788
2789 #define vsetq_lane_u32(__a, __b, __c) \
2790   (uint32x4_t)__builtin_neon_vset_lanev4si (__a, __b, __c)
2791
2792 #define vsetq_lane_p8(__a, __b, __c) \
2793   (poly8x16_t)__builtin_neon_vset_lanev16qi (__a, __b, __c)
2794
2795 #define vsetq_lane_p16(__a, __b, __c) \
2796   (poly16x8_t)__builtin_neon_vset_lanev8hi (__a, __b, __c)
2797
2798 #define vsetq_lane_s64(__a, __b, __c) \
2799   (int64x2_t)__builtin_neon_vset_lanev2di (__a, __b, __c)
2800
2801 #define vsetq_lane_u64(__a, __b, __c) \
2802   (uint64x2_t)__builtin_neon_vset_lanev2di (__a, __b, __c)
2803
2804 #define vcreate_s8(__a) \
2805   (int8x8_t)__builtin_neon_vcreatev8qi (__a)
2806
2807 #define vcreate_s16(__a) \
2808   (int16x4_t)__builtin_neon_vcreatev4hi (__a)
2809
2810 #define vcreate_s32(__a) \
2811   (int32x2_t)__builtin_neon_vcreatev2si (__a)
2812
2813 #define vcreate_s64(__a) \
2814   (int64x1_t)__builtin_neon_vcreatev1di (__a)
2815
2816 #define vcreate_f32(__a) \
2817   (float32x2_t)__builtin_neon_vcreatev2sf (__a)
2818
2819 #define vcreate_u8(__a) \
2820   (uint8x8_t)__builtin_neon_vcreatev8qi (__a)
2821
2822 #define vcreate_u16(__a) \
2823   (uint16x4_t)__builtin_neon_vcreatev4hi (__a)
2824
2825 #define vcreate_u32(__a) \
2826   (uint32x2_t)__builtin_neon_vcreatev2si (__a)
2827
2828 #define vcreate_u64(__a) \
2829   (uint64x1_t)__builtin_neon_vcreatev1di (__a)
2830
2831 #define vcreate_p8(__a) \
2832   (poly8x8_t)__builtin_neon_vcreatev8qi (__a)
2833
2834 #define vcreate_p16(__a) \
2835   (poly16x4_t)__builtin_neon_vcreatev4hi (__a)
2836
2837 #define vdup_n_s8(__a) \
2838   (int8x8_t)__builtin_neon_vdup_nv8qi (__a)
2839
2840 #define vdup_n_s16(__a) \
2841   (int16x4_t)__builtin_neon_vdup_nv4hi (__a)
2842
2843 #define vdup_n_s32(__a) \
2844   (int32x2_t)__builtin_neon_vdup_nv2si (__a)
2845
2846 #define vdup_n_f32(__a) \
2847   (float32x2_t)__builtin_neon_vdup_nv2sf (__a)
2848
2849 #define vdup_n_u8(__a) \
2850   (uint8x8_t)__builtin_neon_vdup_nv8qi (__a)
2851
2852 #define vdup_n_u16(__a) \
2853   (uint16x4_t)__builtin_neon_vdup_nv4hi (__a)
2854
2855 #define vdup_n_u32(__a) \
2856   (uint32x2_t)__builtin_neon_vdup_nv2si (__a)
2857
2858 #define vdup_n_p8(__a) \
2859   (poly8x8_t)__builtin_neon_vdup_nv8qi (__a)
2860
2861 #define vdup_n_p16(__a) \
2862   (poly16x4_t)__builtin_neon_vdup_nv4hi (__a)
2863
2864 #define vdup_n_s64(__a) \
2865   (int64x1_t)__builtin_neon_vdup_nv1di (__a)
2866
2867 #define vdup_n_u64(__a) \
2868   (uint64x1_t)__builtin_neon_vdup_nv1di (__a)
2869
2870 #define vdupq_n_s8(__a) \
2871   (int8x16_t)__builtin_neon_vdup_nv16qi (__a)
2872
2873 #define vdupq_n_s16(__a) \
2874   (int16x8_t)__builtin_neon_vdup_nv8hi (__a)
2875
2876 #define vdupq_n_s32(__a) \
2877   (int32x4_t)__builtin_neon_vdup_nv4si (__a)
2878
2879 #define vdupq_n_f32(__a) \
2880   (float32x4_t)__builtin_neon_vdup_nv4sf (__a)
2881
2882 #define vdupq_n_u8(__a) \
2883   (uint8x16_t)__builtin_neon_vdup_nv16qi (__a)
2884
2885 #define vdupq_n_u16(__a) \
2886   (uint16x8_t)__builtin_neon_vdup_nv8hi (__a)
2887
2888 #define vdupq_n_u32(__a) \
2889   (uint32x4_t)__builtin_neon_vdup_nv4si (__a)
2890
2891 #define vdupq_n_p8(__a) \
2892   (poly8x16_t)__builtin_neon_vdup_nv16qi (__a)
2893
2894 #define vdupq_n_p16(__a) \
2895   (poly16x8_t)__builtin_neon_vdup_nv8hi (__a)
2896
2897 #define vdupq_n_s64(__a) \
2898   (int64x2_t)__builtin_neon_vdup_nv2di (__a)
2899
2900 #define vdupq_n_u64(__a) \
2901   (uint64x2_t)__builtin_neon_vdup_nv2di (__a)
2902
2903 #define vmov_n_s8(__a) \
2904   (int8x8_t)__builtin_neon_vdup_nv8qi (__a)
2905
2906 #define vmov_n_s16(__a) \
2907   (int16x4_t)__builtin_neon_vdup_nv4hi (__a)
2908
2909 #define vmov_n_s32(__a) \
2910   (int32x2_t)__builtin_neon_vdup_nv2si (__a)
2911
2912 #define vmov_n_f32(__a) \
2913   (float32x2_t)__builtin_neon_vdup_nv2sf (__a)
2914
2915 #define vmov_n_u8(__a) \
2916   (uint8x8_t)__builtin_neon_vdup_nv8qi (__a)
2917
2918 #define vmov_n_u16(__a) \
2919   (uint16x4_t)__builtin_neon_vdup_nv4hi (__a)
2920
2921 #define vmov_n_u32(__a) \
2922   (uint32x2_t)__builtin_neon_vdup_nv2si (__a)
2923
2924 #define vmov_n_p8(__a) \
2925   (poly8x8_t)__builtin_neon_vdup_nv8qi (__a)
2926
2927 #define vmov_n_p16(__a) \
2928   (poly16x4_t)__builtin_neon_vdup_nv4hi (__a)
2929
2930 #define vmov_n_s64(__a) \
2931   (int64x1_t)__builtin_neon_vdup_nv1di (__a)
2932
2933 #define vmov_n_u64(__a) \
2934   (uint64x1_t)__builtin_neon_vdup_nv1di (__a)
2935
2936 #define vmovq_n_s8(__a) \
2937   (int8x16_t)__builtin_neon_vdup_nv16qi (__a)
2938
2939 #define vmovq_n_s16(__a) \
2940   (int16x8_t)__builtin_neon_vdup_nv8hi (__a)
2941
2942 #define vmovq_n_s32(__a) \
2943   (int32x4_t)__builtin_neon_vdup_nv4si (__a)
2944
2945 #define vmovq_n_f32(__a) \
2946   (float32x4_t)__builtin_neon_vdup_nv4sf (__a)
2947
2948 #define vmovq_n_u8(__a) \
2949   (uint8x16_t)__builtin_neon_vdup_nv16qi (__a)
2950
2951 #define vmovq_n_u16(__a) \
2952   (uint16x8_t)__builtin_neon_vdup_nv8hi (__a)
2953
2954 #define vmovq_n_u32(__a) \
2955   (uint32x4_t)__builtin_neon_vdup_nv4si (__a)
2956
2957 #define vmovq_n_p8(__a) \
2958   (poly8x16_t)__builtin_neon_vdup_nv16qi (__a)
2959
2960 #define vmovq_n_p16(__a) \
2961   (poly16x8_t)__builtin_neon_vdup_nv8hi (__a)
2962
2963 #define vmovq_n_s64(__a) \
2964   (int64x2_t)__builtin_neon_vdup_nv2di (__a)
2965
2966 #define vmovq_n_u64(__a) \
2967   (uint64x2_t)__builtin_neon_vdup_nv2di (__a)
2968
2969 #define vdup_lane_s8(__a, __b) \
2970   (int8x8_t)__builtin_neon_vdup_lanev8qi (__a, __b)
2971
2972 #define vdup_lane_s16(__a, __b) \
2973   (int16x4_t)__builtin_neon_vdup_lanev4hi (__a, __b)
2974
2975 #define vdup_lane_s32(__a, __b) \
2976   (int32x2_t)__builtin_neon_vdup_lanev2si (__a, __b)
2977
2978 #define vdup_lane_f32(__a, __b) \
2979   (float32x2_t)__builtin_neon_vdup_lanev2sf (__a, __b)
2980
2981 #define vdup_lane_u8(__a, __b) \
2982   (uint8x8_t)__builtin_neon_vdup_lanev8qi (__a, __b)
2983
2984 #define vdup_lane_u16(__a, __b) \
2985   (uint16x4_t)__builtin_neon_vdup_lanev4hi (__a, __b)
2986
2987 #define vdup_lane_u32(__a, __b) \
2988   (uint32x2_t)__builtin_neon_vdup_lanev2si (__a, __b)
2989
2990 #define vdup_lane_p8(__a, __b) \
2991   (poly8x8_t)__builtin_neon_vdup_lanev8qi (__a, __b)
2992
2993 #define vdup_lane_p16(__a, __b) \
2994   (poly16x4_t)__builtin_neon_vdup_lanev4hi (__a, __b)
2995
2996 #define vdup_lane_s64(__a, __b) \
2997   (int64x1_t)__builtin_neon_vdup_lanev1di (__a, __b)
2998
2999 #define vdup_lane_u64(__a, __b) \
3000   (uint64x1_t)__builtin_neon_vdup_lanev1di (__a, __b)
3001
3002 #define vdupq_lane_s8(__a, __b) \
3003   (int8x16_t)__builtin_neon_vdup_lanev16qi (__a, __b)
3004
3005 #define vdupq_lane_s16(__a, __b) \
3006   (int16x8_t)__builtin_neon_vdup_lanev8hi (__a, __b)
3007
3008 #define vdupq_lane_s32(__a, __b) \
3009   (int32x4_t)__builtin_neon_vdup_lanev4si (__a, __b)
3010
3011 #define vdupq_lane_f32(__a, __b) \
3012   (float32x4_t)__builtin_neon_vdup_lanev4sf (__a, __b)
3013
3014 #define vdupq_lane_u8(__a, __b) \
3015   (uint8x16_t)__builtin_neon_vdup_lanev16qi (__a, __b)
3016
3017 #define vdupq_lane_u16(__a, __b) \
3018   (uint16x8_t)__builtin_neon_vdup_lanev8hi (__a, __b)
3019
3020 #define vdupq_lane_u32(__a, __b) \
3021   (uint32x4_t)__builtin_neon_vdup_lanev4si (__a, __b)
3022
3023 #define vdupq_lane_p8(__a, __b) \
3024   (poly8x16_t)__builtin_neon_vdup_lanev16qi (__a, __b)
3025
3026 #define vdupq_lane_p16(__a, __b) \
3027   (poly16x8_t)__builtin_neon_vdup_lanev8hi (__a, __b)
3028
3029 #define vdupq_lane_s64(__a, __b) \
3030   (int64x2_t)__builtin_neon_vdup_lanev2di (__a, __b)
3031
3032 #define vdupq_lane_u64(__a, __b) \
3033   (uint64x2_t)__builtin_neon_vdup_lanev2di (__a, __b)
3034
3035 #define vcombine_s8(__a, __b) \
3036   (int8x16_t)__builtin_neon_vcombinev8qi (__a, __b)
3037
3038 #define vcombine_s16(__a, __b) \
3039   (int16x8_t)__builtin_neon_vcombinev4hi (__a, __b)
3040
3041 #define vcombine_s32(__a, __b) \
3042   (int32x4_t)__builtin_neon_vcombinev2si (__a, __b)
3043
3044 #define vcombine_s64(__a, __b) \
3045   (int64x2_t)__builtin_neon_vcombinev1di (__a, __b)
3046
3047 #define vcombine_f32(__a, __b) \
3048   (float32x4_t)__builtin_neon_vcombinev2sf (__a, __b)
3049
3050 #define vcombine_u8(__a, __b) \
3051   (uint8x16_t)__builtin_neon_vcombinev8qi (__a, __b)
3052
3053 #define vcombine_u16(__a, __b) \
3054   (uint16x8_t)__builtin_neon_vcombinev4hi (__a, __b)
3055
3056 #define vcombine_u32(__a, __b) \
3057   (uint32x4_t)__builtin_neon_vcombinev2si (__a, __b)
3058
3059 #define vcombine_u64(__a, __b) \
3060   (uint64x2_t)__builtin_neon_vcombinev1di (__a, __b)
3061
3062 #define vcombine_p8(__a, __b) \
3063   (poly8x16_t)__builtin_neon_vcombinev8qi (__a, __b)
3064
3065 #define vcombine_p16(__a, __b) \
3066   (poly16x8_t)__builtin_neon_vcombinev4hi (__a, __b)
3067
3068 #define vget_high_s8(__a) \
3069   (int8x8_t)__builtin_neon_vget_highv16qi (__a)
3070
3071 #define vget_high_s16(__a) \
3072   (int16x4_t)__builtin_neon_vget_highv8hi (__a)
3073
3074 #define vget_high_s32(__a) \
3075   (int32x2_t)__builtin_neon_vget_highv4si (__a)
3076
3077 #define vget_high_s64(__a) \
3078   (int64x1_t)__builtin_neon_vget_highv2di (__a)
3079
3080 #define vget_high_f32(__a) \
3081   (float32x2_t)__builtin_neon_vget_highv4sf (__a)
3082
3083 #define vget_high_u8(__a) \
3084   (uint8x8_t)__builtin_neon_vget_highv16qi (__a)
3085
3086 #define vget_high_u16(__a) \
3087   (uint16x4_t)__builtin_neon_vget_highv8hi (__a)
3088
3089 #define vget_high_u32(__a) \
3090   (uint32x2_t)__builtin_neon_vget_highv4si (__a)
3091
3092 #define vget_high_u64(__a) \
3093   (uint64x1_t)__builtin_neon_vget_highv2di (__a)
3094
3095 #define vget_high_p8(__a) \
3096   (poly8x8_t)__builtin_neon_vget_highv16qi (__a)
3097
3098 #define vget_high_p16(__a) \
3099   (poly16x4_t)__builtin_neon_vget_highv8hi (__a)
3100
3101 #define vget_low_s8(__a) \
3102   (int8x8_t)__builtin_neon_vget_lowv16qi (__a)
3103
3104 #define vget_low_s16(__a) \
3105   (int16x4_t)__builtin_neon_vget_lowv8hi (__a)
3106
3107 #define vget_low_s32(__a) \
3108   (int32x2_t)__builtin_neon_vget_lowv4si (__a)
3109
3110 #define vget_low_s64(__a) \
3111   (int64x1_t)__builtin_neon_vget_lowv2di (__a)
3112
3113 #define vget_low_f32(__a) \
3114   (float32x2_t)__builtin_neon_vget_lowv4sf (__a)
3115
3116 #define vget_low_u8(__a) \
3117   (uint8x8_t)__builtin_neon_vget_lowv16qi (__a)
3118
3119 #define vget_low_u16(__a) \
3120   (uint16x4_t)__builtin_neon_vget_lowv8hi (__a)
3121
3122 #define vget_low_u32(__a) \
3123   (uint32x2_t)__builtin_neon_vget_lowv4si (__a)
3124
3125 #define vget_low_u64(__a) \
3126   (uint64x1_t)__builtin_neon_vget_lowv2di (__a)
3127
3128 #define vget_low_p8(__a) \
3129   (poly8x8_t)__builtin_neon_vget_lowv16qi (__a)
3130
3131 #define vget_low_p16(__a) \
3132   (poly16x4_t)__builtin_neon_vget_lowv8hi (__a)
3133
3134 #define vcvt_s32_f32(__a) \
3135   (int32x2_t)__builtin_neon_vcvtv2sf (__a, 1)
3136
3137 #define vcvt_f32_s32(__a) \
3138   (float32x2_t)__builtin_neon_vcvtv2si (__a, 1)
3139
3140 #define vcvt_f32_u32(__a) \
3141   (float32x2_t)__builtin_neon_vcvtv2si (__a, 0)
3142
3143 #define vcvt_u32_f32(__a) \
3144   (uint32x2_t)__builtin_neon_vcvtv2sf (__a, 0)
3145
3146 #define vcvtq_s32_f32(__a) \
3147   (int32x4_t)__builtin_neon_vcvtv4sf (__a, 1)
3148
3149 #define vcvtq_f32_s32(__a) \
3150   (float32x4_t)__builtin_neon_vcvtv4si (__a, 1)
3151
3152 #define vcvtq_f32_u32(__a) \
3153   (float32x4_t)__builtin_neon_vcvtv4si (__a, 0)
3154
3155 #define vcvtq_u32_f32(__a) \
3156   (uint32x4_t)__builtin_neon_vcvtv4sf (__a, 0)
3157
3158 #define vcvt_n_s32_f32(__a, __b) \
3159   (int32x2_t)__builtin_neon_vcvt_nv2sf (__a, __b, 1)
3160
3161 #define vcvt_n_f32_s32(__a, __b) \
3162   (float32x2_t)__builtin_neon_vcvt_nv2si (__a, __b, 1)
3163
3164 #define vcvt_n_f32_u32(__a, __b) \
3165   (float32x2_t)__builtin_neon_vcvt_nv2si (__a, __b, 0)
3166
3167 #define vcvt_n_u32_f32(__a, __b) \
3168   (uint32x2_t)__builtin_neon_vcvt_nv2sf (__a, __b, 0)
3169
3170 #define vcvtq_n_s32_f32(__a, __b) \
3171   (int32x4_t)__builtin_neon_vcvt_nv4sf (__a, __b, 1)
3172
3173 #define vcvtq_n_f32_s32(__a, __b) \
3174   (float32x4_t)__builtin_neon_vcvt_nv4si (__a, __b, 1)
3175
3176 #define vcvtq_n_f32_u32(__a, __b) \
3177   (float32x4_t)__builtin_neon_vcvt_nv4si (__a, __b, 0)
3178
3179 #define vcvtq_n_u32_f32(__a, __b) \
3180   (uint32x4_t)__builtin_neon_vcvt_nv4sf (__a, __b, 0)
3181
3182 #define vmovn_s16(__a) \
3183   (int8x8_t)__builtin_neon_vmovnv8hi (__a, 1)
3184
3185 #define vmovn_s32(__a) \
3186   (int16x4_t)__builtin_neon_vmovnv4si (__a, 1)
3187
3188 #define vmovn_s64(__a) \
3189   (int32x2_t)__builtin_neon_vmovnv2di (__a, 1)
3190
3191 #define vmovn_u16(__a) \
3192   (uint8x8_t)__builtin_neon_vmovnv8hi (__a, 0)
3193
3194 #define vmovn_u32(__a) \
3195   (uint16x4_t)__builtin_neon_vmovnv4si (__a, 0)
3196
3197 #define vmovn_u64(__a) \
3198   (uint32x2_t)__builtin_neon_vmovnv2di (__a, 0)
3199
3200 #define vqmovn_s16(__a) \
3201   (int8x8_t)__builtin_neon_vqmovnv8hi (__a, 1)
3202
3203 #define vqmovn_s32(__a) \
3204   (int16x4_t)__builtin_neon_vqmovnv4si (__a, 1)
3205
3206 #define vqmovn_s64(__a) \
3207   (int32x2_t)__builtin_neon_vqmovnv2di (__a, 1)
3208
3209 #define vqmovn_u16(__a) \
3210   (uint8x8_t)__builtin_neon_vqmovnv8hi (__a, 0)
3211
3212 #define vqmovn_u32(__a) \
3213   (uint16x4_t)__builtin_neon_vqmovnv4si (__a, 0)
3214
3215 #define vqmovn_u64(__a) \
3216   (uint32x2_t)__builtin_neon_vqmovnv2di (__a, 0)
3217
3218 #define vqmovun_s16(__a) \
3219   (uint8x8_t)__builtin_neon_vqmovunv8hi (__a, 1)
3220
3221 #define vqmovun_s32(__a) \
3222   (uint16x4_t)__builtin_neon_vqmovunv4si (__a, 1)
3223
3224 #define vqmovun_s64(__a) \
3225   (uint32x2_t)__builtin_neon_vqmovunv2di (__a, 1)
3226
3227 #define vmovl_s8(__a) \
3228   (int16x8_t)__builtin_neon_vmovlv8qi (__a, 1)
3229
3230 #define vmovl_s16(__a) \
3231   (int32x4_t)__builtin_neon_vmovlv4hi (__a, 1)
3232
3233 #define vmovl_s32(__a) \
3234   (int64x2_t)__builtin_neon_vmovlv2si (__a, 1)
3235
3236 #define vmovl_u8(__a) \
3237   (uint16x8_t)__builtin_neon_vmovlv8qi (__a, 0)
3238
3239 #define vmovl_u16(__a) \
3240   (uint32x4_t)__builtin_neon_vmovlv4hi (__a, 0)
3241
3242 #define vmovl_u32(__a) \
3243   (uint64x2_t)__builtin_neon_vmovlv2si (__a, 0)
3244
3245 #define vtbl1_s8(__a, __b) \
3246   (int8x8_t)__builtin_neon_vtbl1v8qi (__a, __b)
3247
3248 #define vtbl1_u8(__a, __b) \
3249   (uint8x8_t)__builtin_neon_vtbl1v8qi (__a, __b)
3250
3251 #define vtbl1_p8(__a, __b) \
3252   (poly8x8_t)__builtin_neon_vtbl1v8qi (__a, __b)
3253
3254 #define vtbl2_s8(__a, __b) \
3255   ({ \
3256      union { int8x8x2_t __i; __builtin_neon_v8qi2 __o; } __au = { __a }; \
3257      (int8x8_t)__builtin_neon_vtbl2v8qi (__au.__o, __b); \
3258    })
3259
3260 #define vtbl2_u8(__a, __b) \
3261   ({ \
3262      union { uint8x8x2_t __i; __builtin_neon_v8qi2 __o; } __au = { __a }; \
3263      (uint8x8_t)__builtin_neon_vtbl2v8qi (__au.__o, __b); \
3264    })
3265
3266 #define vtbl2_p8(__a, __b) \
3267   ({ \
3268      union { poly8x8x2_t __i; __builtin_neon_v8qi2 __o; } __au = { __a }; \
3269      (poly8x8_t)__builtin_neon_vtbl2v8qi (__au.__o, __b); \
3270    })
3271
3272 #define vtbl3_s8(__a, __b) \
3273   ({ \
3274      union { int8x8x3_t __i; __builtin_neon_v8qi3 __o; } __au = { __a }; \
3275      (int8x8_t)__builtin_neon_vtbl3v8qi (__au.__o, __b); \
3276    })
3277
3278 #define vtbl3_u8(__a, __b) \
3279   ({ \
3280      union { uint8x8x3_t __i; __builtin_neon_v8qi3 __o; } __au = { __a }; \
3281      (uint8x8_t)__builtin_neon_vtbl3v8qi (__au.__o, __b); \
3282    })
3283
3284 #define vtbl3_p8(__a, __b) \
3285   ({ \
3286      union { poly8x8x3_t __i; __builtin_neon_v8qi3 __o; } __au = { __a }; \
3287      (poly8x8_t)__builtin_neon_vtbl3v8qi (__au.__o, __b); \
3288    })
3289
3290 #define vtbl4_s8(__a, __b) \
3291   ({ \
3292      union { int8x8x4_t __i; __builtin_neon_v8qi4 __o; } __au = { __a }; \
3293      (int8x8_t)__builtin_neon_vtbl4v8qi (__au.__o, __b); \
3294    })
3295
3296 #define vtbl4_u8(__a, __b) \
3297   ({ \
3298      union { uint8x8x4_t __i; __builtin_neon_v8qi4 __o; } __au = { __a }; \
3299      (uint8x8_t)__builtin_neon_vtbl4v8qi (__au.__o, __b); \
3300    })
3301
3302 #define vtbl4_p8(__a, __b) \
3303   ({ \
3304      union { poly8x8x4_t __i; __builtin_neon_v8qi4 __o; } __au = { __a }; \
3305      (poly8x8_t)__builtin_neon_vtbl4v8qi (__au.__o, __b); \
3306    })
3307
3308 #define vtbx1_s8(__a, __b, __c) \
3309   (int8x8_t)__builtin_neon_vtbx1v8qi (__a, __b, __c)
3310
3311 #define vtbx1_u8(__a, __b, __c) \
3312   (uint8x8_t)__builtin_neon_vtbx1v8qi (__a, __b, __c)
3313
3314 #define vtbx1_p8(__a, __b, __c) \
3315   (poly8x8_t)__builtin_neon_vtbx1v8qi (__a, __b, __c)
3316
3317 #define vtbx2_s8(__a, __b, __c) \
3318   ({ \
3319      union { int8x8x2_t __i; __builtin_neon_v8qi2 __o; } __bu = { __b }; \
3320      (int8x8_t)__builtin_neon_vtbx2v8qi (__a, __bu.__o, __c); \
3321    })
3322
3323 #define vtbx2_u8(__a, __b, __c) \
3324   ({ \
3325      union { uint8x8x2_t __i; __builtin_neon_v8qi2 __o; } __bu = { __b }; \
3326      (uint8x8_t)__builtin_neon_vtbx2v8qi (__a, __bu.__o, __c); \
3327    })
3328
3329 #define vtbx2_p8(__a, __b, __c) \
3330   ({ \
3331      union { poly8x8x2_t __i; __builtin_neon_v8qi2 __o; } __bu = { __b }; \
3332      (poly8x8_t)__builtin_neon_vtbx2v8qi (__a, __bu.__o, __c); \
3333    })
3334
3335 #define vtbx3_s8(__a, __b, __c) \
3336   ({ \
3337      union { int8x8x3_t __i; __builtin_neon_v8qi3 __o; } __bu = { __b }; \
3338      (int8x8_t)__builtin_neon_vtbx3v8qi (__a, __bu.__o, __c); \
3339    })
3340
3341 #define vtbx3_u8(__a, __b, __c) \
3342   ({ \
3343      union { uint8x8x3_t __i; __builtin_neon_v8qi3 __o; } __bu = { __b }; \
3344      (uint8x8_t)__builtin_neon_vtbx3v8qi (__a, __bu.__o, __c); \
3345    })
3346
3347 #define vtbx3_p8(__a, __b, __c) \
3348   ({ \
3349      union { poly8x8x3_t __i; __builtin_neon_v8qi3 __o; } __bu = { __b }; \
3350      (poly8x8_t)__builtin_neon_vtbx3v8qi (__a, __bu.__o, __c); \
3351    })
3352
3353 #define vtbx4_s8(__a, __b, __c) \
3354   ({ \
3355      union { int8x8x4_t __i; __builtin_neon_v8qi4 __o; } __bu = { __b }; \
3356      (int8x8_t)__builtin_neon_vtbx4v8qi (__a, __bu.__o, __c); \
3357    })
3358
3359 #define vtbx4_u8(__a, __b, __c) \
3360   ({ \
3361      union { uint8x8x4_t __i; __builtin_neon_v8qi4 __o; } __bu = { __b }; \
3362      (uint8x8_t)__builtin_neon_vtbx4v8qi (__a, __bu.__o, __c); \
3363    })
3364
3365 #define vtbx4_p8(__a, __b, __c) \
3366   ({ \
3367      union { poly8x8x4_t __i; __builtin_neon_v8qi4 __o; } __bu = { __b }; \
3368      (poly8x8_t)__builtin_neon_vtbx4v8qi (__a, __bu.__o, __c); \
3369    })
3370
3371 #define vmul_lane_s16(__a, __b, __c) \
3372   (int16x4_t)__builtin_neon_vmul_lanev4hi (__a, __b, __c, 1)
3373
3374 #define vmul_lane_s32(__a, __b, __c) \
3375   (int32x2_t)__builtin_neon_vmul_lanev2si (__a, __b, __c, 1)
3376
3377 #define vmul_lane_f32(__a, __b, __c) \
3378   (float32x2_t)__builtin_neon_vmul_lanev2sf (__a, __b, __c, 5)
3379
3380 #define vmul_lane_u16(__a, __b, __c) \
3381   (uint16x4_t)__builtin_neon_vmul_lanev4hi (__a, __b, __c, 0)
3382
3383 #define vmul_lane_u32(__a, __b, __c) \
3384   (uint32x2_t)__builtin_neon_vmul_lanev2si (__a, __b, __c, 0)
3385
3386 #define vmulq_lane_s16(__a, __b, __c) \
3387   (int16x8_t)__builtin_neon_vmul_lanev8hi (__a, __b, __c, 1)
3388
3389 #define vmulq_lane_s32(__a, __b, __c) \
3390   (int32x4_t)__builtin_neon_vmul_lanev4si (__a, __b, __c, 1)
3391
3392 #define vmulq_lane_f32(__a, __b, __c) \
3393   (float32x4_t)__builtin_neon_vmul_lanev4sf (__a, __b, __c, 5)
3394
3395 #define vmulq_lane_u16(__a, __b, __c) \
3396   (uint16x8_t)__builtin_neon_vmul_lanev8hi (__a, __b, __c, 0)
3397
3398 #define vmulq_lane_u32(__a, __b, __c) \
3399   (uint32x4_t)__builtin_neon_vmul_lanev4si (__a, __b, __c, 0)
3400
3401 #define vmla_lane_s16(__a, __b, __c, __d) \
3402   (int16x4_t)__builtin_neon_vmla_lanev4hi (__a, __b, __c, __d, 1)
3403
3404 #define vmla_lane_s32(__a, __b, __c, __d) \
3405   (int32x2_t)__builtin_neon_vmla_lanev2si (__a, __b, __c, __d, 1)
3406
3407 #define vmla_lane_f32(__a, __b, __c, __d) \
3408   (float32x2_t)__builtin_neon_vmla_lanev2sf (__a, __b, __c, __d, 5)
3409
3410 #define vmla_lane_u16(__a, __b, __c, __d) \
3411   (uint16x4_t)__builtin_neon_vmla_lanev4hi (__a, __b, __c, __d, 0)
3412
3413 #define vmla_lane_u32(__a, __b, __c, __d) \
3414   (uint32x2_t)__builtin_neon_vmla_lanev2si (__a, __b, __c, __d, 0)
3415
3416 #define vmlaq_lane_s16(__a, __b, __c, __d) \
3417   (int16x8_t)__builtin_neon_vmla_lanev8hi (__a, __b, __c, __d, 1)
3418
3419 #define vmlaq_lane_s32(__a, __b, __c, __d) \
3420   (int32x4_t)__builtin_neon_vmla_lanev4si (__a, __b, __c, __d, 1)
3421
3422 #define vmlaq_lane_f32(__a, __b, __c, __d) \
3423   (float32x4_t)__builtin_neon_vmla_lanev4sf (__a, __b, __c, __d, 5)
3424
3425 #define vmlaq_lane_u16(__a, __b, __c, __d) \
3426   (uint16x8_t)__builtin_neon_vmla_lanev8hi (__a, __b, __c, __d, 0)
3427
3428 #define vmlaq_lane_u32(__a, __b, __c, __d) \
3429   (uint32x4_t)__builtin_neon_vmla_lanev4si (__a, __b, __c, __d, 0)
3430
3431 #define vmlal_lane_s16(__a, __b, __c, __d) \
3432   (int32x4_t)__builtin_neon_vmlal_lanev4hi (__a, __b, __c, __d, 1)
3433
3434 #define vmlal_lane_s32(__a, __b, __c, __d) \
3435   (int64x2_t)__builtin_neon_vmlal_lanev2si (__a, __b, __c, __d, 1)
3436
3437 #define vmlal_lane_u16(__a, __b, __c, __d) \
3438   (uint32x4_t)__builtin_neon_vmlal_lanev4hi (__a, __b, __c, __d, 0)
3439
3440 #define vmlal_lane_u32(__a, __b, __c, __d) \
3441   (uint64x2_t)__builtin_neon_vmlal_lanev2si (__a, __b, __c, __d, 0)
3442
3443 #define vqdmlal_lane_s16(__a, __b, __c, __d) \
3444   (int32x4_t)__builtin_neon_vqdmlal_lanev4hi (__a, __b, __c, __d, 1)
3445
3446 #define vqdmlal_lane_s32(__a, __b, __c, __d) \
3447   (int64x2_t)__builtin_neon_vqdmlal_lanev2si (__a, __b, __c, __d, 1)
3448
3449 #define vmls_lane_s16(__a, __b, __c, __d) \
3450   (int16x4_t)__builtin_neon_vmls_lanev4hi (__a, __b, __c, __d, 1)
3451
3452 #define vmls_lane_s32(__a, __b, __c, __d) \
3453   (int32x2_t)__builtin_neon_vmls_lanev2si (__a, __b, __c, __d, 1)
3454
3455 #define vmls_lane_f32(__a, __b, __c, __d) \
3456   (float32x2_t)__builtin_neon_vmls_lanev2sf (__a, __b, __c, __d, 5)
3457
3458 #define vmls_lane_u16(__a, __b, __c, __d) \
3459   (uint16x4_t)__builtin_neon_vmls_lanev4hi (__a, __b, __c, __d, 0)
3460
3461 #define vmls_lane_u32(__a, __b, __c, __d) \
3462   (uint32x2_t)__builtin_neon_vmls_lanev2si (__a, __b, __c, __d, 0)
3463
3464 #define vmlsq_lane_s16(__a, __b, __c, __d) \
3465   (int16x8_t)__builtin_neon_vmls_lanev8hi (__a, __b, __c, __d, 1)
3466
3467 #define vmlsq_lane_s32(__a, __b, __c, __d) \
3468   (int32x4_t)__builtin_neon_vmls_lanev4si (__a, __b, __c, __d, 1)
3469
3470 #define vmlsq_lane_f32(__a, __b, __c, __d) \
3471   (float32x4_t)__builtin_neon_vmls_lanev4sf (__a, __b, __c, __d, 5)
3472
3473 #define vmlsq_lane_u16(__a, __b, __c, __d) \
3474   (uint16x8_t)__builtin_neon_vmls_lanev8hi (__a, __b, __c, __d, 0)
3475
3476 #define vmlsq_lane_u32(__a, __b, __c, __d) \
3477   (uint32x4_t)__builtin_neon_vmls_lanev4si (__a, __b, __c, __d, 0)
3478
3479 #define vmlsl_lane_s16(__a, __b, __c, __d) \
3480   (int32x4_t)__builtin_neon_vmlsl_lanev4hi (__a, __b, __c, __d, 1)
3481
3482 #define vmlsl_lane_s32(__a, __b, __c, __d) \
3483   (int64x2_t)__builtin_neon_vmlsl_lanev2si (__a, __b, __c, __d, 1)
3484
3485 #define vmlsl_lane_u16(__a, __b, __c, __d) \
3486   (uint32x4_t)__builtin_neon_vmlsl_lanev4hi (__a, __b, __c, __d, 0)
3487
3488 #define vmlsl_lane_u32(__a, __b, __c, __d) \
3489   (uint64x2_t)__builtin_neon_vmlsl_lanev2si (__a, __b, __c, __d, 0)
3490
3491 #define vqdmlsl_lane_s16(__a, __b, __c, __d) \
3492   (int32x4_t)__builtin_neon_vqdmlsl_lanev4hi (__a, __b, __c, __d, 1)
3493
3494 #define vqdmlsl_lane_s32(__a, __b, __c, __d) \
3495   (int64x2_t)__builtin_neon_vqdmlsl_lanev2si (__a, __b, __c, __d, 1)
3496
3497 #define vmull_lane_s16(__a, __b, __c) \
3498   (int32x4_t)__builtin_neon_vmull_lanev4hi (__a, __b, __c, 1)
3499
3500 #define vmull_lane_s32(__a, __b, __c) \
3501   (int64x2_t)__builtin_neon_vmull_lanev2si (__a, __b, __c, 1)
3502
3503 #define vmull_lane_u16(__a, __b, __c) \
3504   (uint32x4_t)__builtin_neon_vmull_lanev4hi (__a, __b, __c, 0)
3505
3506 #define vmull_lane_u32(__a, __b, __c) \
3507   (uint64x2_t)__builtin_neon_vmull_lanev2si (__a, __b, __c, 0)
3508
3509 #define vqdmull_lane_s16(__a, __b, __c) \
3510   (int32x4_t)__builtin_neon_vqdmull_lanev4hi (__a, __b, __c, 1)
3511
3512 #define vqdmull_lane_s32(__a, __b, __c) \
3513   (int64x2_t)__builtin_neon_vqdmull_lanev2si (__a, __b, __c, 1)
3514
3515 #define vqdmulhq_lane_s16(__a, __b, __c) \
3516   (int16x8_t)__builtin_neon_vqdmulh_lanev8hi (__a, __b, __c, 1)
3517
3518 #define vqdmulhq_lane_s32(__a, __b, __c) \
3519   (int32x4_t)__builtin_neon_vqdmulh_lanev4si (__a, __b, __c, 1)
3520
3521 #define vqdmulh_lane_s16(__a, __b, __c) \
3522   (int16x4_t)__builtin_neon_vqdmulh_lanev4hi (__a, __b, __c, 1)
3523
3524 #define vqdmulh_lane_s32(__a, __b, __c) \
3525   (int32x2_t)__builtin_neon_vqdmulh_lanev2si (__a, __b, __c, 1)
3526
3527 #define vqrdmulhq_lane_s16(__a, __b, __c) \
3528   (int16x8_t)__builtin_neon_vqdmulh_lanev8hi (__a, __b, __c, 3)
3529
3530 #define vqrdmulhq_lane_s32(__a, __b, __c) \
3531   (int32x4_t)__builtin_neon_vqdmulh_lanev4si (__a, __b, __c, 3)
3532
3533 #define vqrdmulh_lane_s16(__a, __b, __c) \
3534   (int16x4_t)__builtin_neon_vqdmulh_lanev4hi (__a, __b, __c, 3)
3535
3536 #define vqrdmulh_lane_s32(__a, __b, __c) \
3537   (int32x2_t)__builtin_neon_vqdmulh_lanev2si (__a, __b, __c, 3)
3538
3539 #define vmul_n_s16(__a, __b) \
3540   (int16x4_t)__builtin_neon_vmul_nv4hi (__a, __b, 1)
3541
3542 #define vmul_n_s32(__a, __b) \
3543   (int32x2_t)__builtin_neon_vmul_nv2si (__a, __b, 1)
3544
3545 #define vmul_n_f32(__a, __b) \
3546   (float32x2_t)__builtin_neon_vmul_nv2sf (__a, __b, 5)
3547
3548 #define vmul_n_u16(__a, __b) \
3549   (uint16x4_t)__builtin_neon_vmul_nv4hi (__a, __b, 0)
3550
3551 #define vmul_n_u32(__a, __b) \
3552   (uint32x2_t)__builtin_neon_vmul_nv2si (__a, __b, 0)
3553
3554 #define vmulq_n_s16(__a, __b) \
3555   (int16x8_t)__builtin_neon_vmul_nv8hi (__a, __b, 1)
3556
3557 #define vmulq_n_s32(__a, __b) \
3558   (int32x4_t)__builtin_neon_vmul_nv4si (__a, __b, 1)
3559
3560 #define vmulq_n_f32(__a, __b) \
3561   (float32x4_t)__builtin_neon_vmul_nv4sf (__a, __b, 5)
3562
3563 #define vmulq_n_u16(__a, __b) \
3564   (uint16x8_t)__builtin_neon_vmul_nv8hi (__a, __b, 0)
3565
3566 #define vmulq_n_u32(__a, __b) \
3567   (uint32x4_t)__builtin_neon_vmul_nv4si (__a, __b, 0)
3568
3569 #define vmull_n_s16(__a, __b) \
3570   (int32x4_t)__builtin_neon_vmull_nv4hi (__a, __b, 1)
3571
3572 #define vmull_n_s32(__a, __b) \
3573   (int64x2_t)__builtin_neon_vmull_nv2si (__a, __b, 1)
3574
3575 #define vmull_n_u16(__a, __b) \
3576   (uint32x4_t)__builtin_neon_vmull_nv4hi (__a, __b, 0)
3577
3578 #define vmull_n_u32(__a, __b) \
3579   (uint64x2_t)__builtin_neon_vmull_nv2si (__a, __b, 0)
3580
3581 #define vqdmull_n_s16(__a, __b) \
3582   (int32x4_t)__builtin_neon_vqdmull_nv4hi (__a, __b, 1)
3583
3584 #define vqdmull_n_s32(__a, __b) \
3585   (int64x2_t)__builtin_neon_vqdmull_nv2si (__a, __b, 1)
3586
3587 #define vqdmulhq_n_s16(__a, __b) \
3588   (int16x8_t)__builtin_neon_vqdmulh_nv8hi (__a, __b, 1)
3589
3590 #define vqdmulhq_n_s32(__a, __b) \
3591   (int32x4_t)__builtin_neon_vqdmulh_nv4si (__a, __b, 1)
3592
3593 #define vqdmulh_n_s16(__a, __b) \
3594   (int16x4_t)__builtin_neon_vqdmulh_nv4hi (__a, __b, 1)
3595
3596 #define vqdmulh_n_s32(__a, __b) \
3597   (int32x2_t)__builtin_neon_vqdmulh_nv2si (__a, __b, 1)
3598
3599 #define vqrdmulhq_n_s16(__a, __b) \
3600   (int16x8_t)__builtin_neon_vqdmulh_nv8hi (__a, __b, 3)
3601
3602 #define vqrdmulhq_n_s32(__a, __b) \
3603   (int32x4_t)__builtin_neon_vqdmulh_nv4si (__a, __b, 3)
3604
3605 #define vqrdmulh_n_s16(__a, __b) \
3606   (int16x4_t)__builtin_neon_vqdmulh_nv4hi (__a, __b, 3)
3607
3608 #define vqrdmulh_n_s32(__a, __b) \
3609   (int32x2_t)__builtin_neon_vqdmulh_nv2si (__a, __b, 3)
3610
3611 #define vmla_n_s16(__a, __b, __c) \
3612   (int16x4_t)__builtin_neon_vmla_nv4hi (__a, __b, __c, 1)
3613
3614 #define vmla_n_s32(__a, __b, __c) \
3615   (int32x2_t)__builtin_neon_vmla_nv2si (__a, __b, __c, 1)
3616
3617 #define vmla_n_f32(__a, __b, __c) \
3618   (float32x2_t)__builtin_neon_vmla_nv2sf (__a, __b, __c, 5)
3619
3620 #define vmla_n_u16(__a, __b, __c) \
3621   (uint16x4_t)__builtin_neon_vmla_nv4hi (__a, __b, __c, 0)
3622
3623 #define vmla_n_u32(__a, __b, __c) \
3624   (uint32x2_t)__builtin_neon_vmla_nv2si (__a, __b, __c, 0)
3625
3626 #define vmlaq_n_s16(__a, __b, __c) \
3627   (int16x8_t)__builtin_neon_vmla_nv8hi (__a, __b, __c, 1)
3628
3629 #define vmlaq_n_s32(__a, __b, __c) \
3630   (int32x4_t)__builtin_neon_vmla_nv4si (__a, __b, __c, 1)
3631
3632 #define vmlaq_n_f32(__a, __b, __c) \
3633   (float32x4_t)__builtin_neon_vmla_nv4sf (__a, __b, __c, 5)
3634
3635 #define vmlaq_n_u16(__a, __b, __c) \
3636   (uint16x8_t)__builtin_neon_vmla_nv8hi (__a, __b, __c, 0)
3637
3638 #define vmlaq_n_u32(__a, __b, __c) \
3639   (uint32x4_t)__builtin_neon_vmla_nv4si (__a, __b, __c, 0)
3640
3641 #define vmlal_n_s16(__a, __b, __c) \
3642   (int32x4_t)__builtin_neon_vmlal_nv4hi (__a, __b, __c, 1)
3643
3644 #define vmlal_n_s32(__a, __b, __c) \
3645   (int64x2_t)__builtin_neon_vmlal_nv2si (__a, __b, __c, 1)
3646
3647 #define vmlal_n_u16(__a, __b, __c) \
3648   (uint32x4_t)__builtin_neon_vmlal_nv4hi (__a, __b, __c, 0)
3649
3650 #define vmlal_n_u32(__a, __b, __c) \
3651   (uint64x2_t)__builtin_neon_vmlal_nv2si (__a, __b, __c, 0)
3652
3653 #define vqdmlal_n_s16(__a, __b, __c) \
3654   (int32x4_t)__builtin_neon_vqdmlal_nv4hi (__a, __b, __c, 1)
3655
3656 #define vqdmlal_n_s32(__a, __b, __c) \
3657   (int64x2_t)__builtin_neon_vqdmlal_nv2si (__a, __b, __c, 1)
3658
3659 #define vmls_n_s16(__a, __b, __c) \
3660   (int16x4_t)__builtin_neon_vmls_nv4hi (__a, __b, __c, 1)
3661
3662 #define vmls_n_s32(__a, __b, __c) \
3663   (int32x2_t)__builtin_neon_vmls_nv2si (__a, __b, __c, 1)
3664
3665 #define vmls_n_f32(__a, __b, __c) \
3666   (float32x2_t)__builtin_neon_vmls_nv2sf (__a, __b, __c, 5)
3667
3668 #define vmls_n_u16(__a, __b, __c) \
3669   (uint16x4_t)__builtin_neon_vmls_nv4hi (__a, __b, __c, 0)
3670
3671 #define vmls_n_u32(__a, __b, __c) \
3672   (uint32x2_t)__builtin_neon_vmls_nv2si (__a, __b, __c, 0)
3673
3674 #define vmlsq_n_s16(__a, __b, __c) \
3675   (int16x8_t)__builtin_neon_vmls_nv8hi (__a, __b, __c, 1)
3676
3677 #define vmlsq_n_s32(__a, __b, __c) \
3678   (int32x4_t)__builtin_neon_vmls_nv4si (__a, __b, __c, 1)
3679
3680 #define vmlsq_n_f32(__a, __b, __c) \
3681   (float32x4_t)__builtin_neon_vmls_nv4sf (__a, __b, __c, 5)
3682
3683 #define vmlsq_n_u16(__a, __b, __c) \
3684   (uint16x8_t)__builtin_neon_vmls_nv8hi (__a, __b, __c, 0)
3685
3686 #define vmlsq_n_u32(__a, __b, __c) \
3687   (uint32x4_t)__builtin_neon_vmls_nv4si (__a, __b, __c, 0)
3688
3689 #define vmlsl_n_s16(__a, __b, __c) \
3690   (int32x4_t)__builtin_neon_vmlsl_nv4hi (__a, __b, __c, 1)
3691
3692 #define vmlsl_n_s32(__a, __b, __c) \
3693   (int64x2_t)__builtin_neon_vmlsl_nv2si (__a, __b, __c, 1)
3694
3695 #define vmlsl_n_u16(__a, __b, __c) \
3696   (uint32x4_t)__builtin_neon_vmlsl_nv4hi (__a, __b, __c, 0)
3697
3698 #define vmlsl_n_u32(__a, __b, __c) \
3699   (uint64x2_t)__builtin_neon_vmlsl_nv2si (__a, __b, __c, 0)
3700
3701 #define vqdmlsl_n_s16(__a, __b, __c) \
3702   (int32x4_t)__builtin_neon_vqdmlsl_nv4hi (__a, __b, __c, 1)
3703
3704 #define vqdmlsl_n_s32(__a, __b, __c) \
3705   (int64x2_t)__builtin_neon_vqdmlsl_nv2si (__a, __b, __c, 1)
3706
3707 #define vext_s8(__a, __b, __c) \
3708   (int8x8_t)__builtin_neon_vextv8qi (__a, __b, __c)
3709
3710 #define vext_s16(__a, __b, __c) \
3711   (int16x4_t)__builtin_neon_vextv4hi (__a, __b, __c)
3712
3713 #define vext_s32(__a, __b, __c) \
3714   (int32x2_t)__builtin_neon_vextv2si (__a, __b, __c)
3715
3716 #define vext_s64(__a, __b, __c) \
3717   (int64x1_t)__builtin_neon_vextv1di (__a, __b, __c)
3718
3719 #define vext_f32(__a, __b, __c) \
3720   (float32x2_t)__builtin_neon_vextv2sf (__a, __b, __c)
3721
3722 #define vext_u8(__a, __b, __c) \
3723   (uint8x8_t)__builtin_neon_vextv8qi (__a, __b, __c)
3724
3725 #define vext_u16(__a, __b, __c) \
3726   (uint16x4_t)__builtin_neon_vextv4hi (__a, __b, __c)
3727
3728 #define vext_u32(__a, __b, __c) \
3729   (uint32x2_t)__builtin_neon_vextv2si (__a, __b, __c)
3730
3731 #define vext_u64(__a, __b, __c) \
3732   (uint64x1_t)__builtin_neon_vextv1di (__a, __b, __c)
3733
3734 #define vext_p8(__a, __b, __c) \
3735   (poly8x8_t)__builtin_neon_vextv8qi (__a, __b, __c)
3736
3737 #define vext_p16(__a, __b, __c) \
3738   (poly16x4_t)__builtin_neon_vextv4hi (__a, __b, __c)
3739
3740 #define vextq_s8(__a, __b, __c) \
3741   (int8x16_t)__builtin_neon_vextv16qi (__a, __b, __c)
3742
3743 #define vextq_s16(__a, __b, __c) \
3744   (int16x8_t)__builtin_neon_vextv8hi (__a, __b, __c)
3745
3746 #define vextq_s32(__a, __b, __c) \
3747   (int32x4_t)__builtin_neon_vextv4si (__a, __b, __c)
3748
3749 #define vextq_s64(__a, __b, __c) \
3750   (int64x2_t)__builtin_neon_vextv2di (__a, __b, __c)
3751
3752 #define vextq_f32(__a, __b, __c) \
3753   (float32x4_t)__builtin_neon_vextv4sf (__a, __b, __c)
3754
3755 #define vextq_u8(__a, __b, __c) \
3756   (uint8x16_t)__builtin_neon_vextv16qi (__a, __b, __c)
3757
3758 #define vextq_u16(__a, __b, __c) \
3759   (uint16x8_t)__builtin_neon_vextv8hi (__a, __b, __c)
3760
3761 #define vextq_u32(__a, __b, __c) \
3762   (uint32x4_t)__builtin_neon_vextv4si (__a, __b, __c)
3763
3764 #define vextq_u64(__a, __b, __c) \
3765   (uint64x2_t)__builtin_neon_vextv2di (__a, __b, __c)
3766
3767 #define vextq_p8(__a, __b, __c) \
3768   (poly8x16_t)__builtin_neon_vextv16qi (__a, __b, __c)
3769
3770 #define vextq_p16(__a, __b, __c) \
3771   (poly16x8_t)__builtin_neon_vextv8hi (__a, __b, __c)
3772
3773 #define vrev64_s8(__a) \
3774   (int8x8_t)__builtin_neon_vrev64v8qi (__a, 1)
3775
3776 #define vrev64_s16(__a) \
3777   (int16x4_t)__builtin_neon_vrev64v4hi (__a, 1)
3778
3779 #define vrev64_s32(__a) \
3780   (int32x2_t)__builtin_neon_vrev64v2si (__a, 1)
3781
3782 #define vrev64_f32(__a) \
3783   (float32x2_t)__builtin_neon_vrev64v2sf (__a, 5)
3784
3785 #define vrev64_u8(__a) \
3786   (uint8x8_t)__builtin_neon_vrev64v8qi (__a, 0)
3787
3788 #define vrev64_u16(__a) \
3789   (uint16x4_t)__builtin_neon_vrev64v4hi (__a, 0)
3790
3791 #define vrev64_u32(__a) \
3792   (uint32x2_t)__builtin_neon_vrev64v2si (__a, 0)
3793
3794 #define vrev64_p8(__a) \
3795   (poly8x8_t)__builtin_neon_vrev64v8qi (__a, 4)
3796
3797 #define vrev64_p16(__a) \
3798   (poly16x4_t)__builtin_neon_vrev64v4hi (__a, 4)
3799
3800 #define vrev64q_s8(__a) \
3801   (int8x16_t)__builtin_neon_vrev64v16qi (__a, 1)
3802
3803 #define vrev64q_s16(__a) \
3804   (int16x8_t)__builtin_neon_vrev64v8hi (__a, 1)
3805
3806 #define vrev64q_s32(__a) \
3807   (int32x4_t)__builtin_neon_vrev64v4si (__a, 1)
3808
3809 #define vrev64q_f32(__a) \
3810   (float32x4_t)__builtin_neon_vrev64v4sf (__a, 5)
3811
3812 #define vrev64q_u8(__a) \
3813   (uint8x16_t)__builtin_neon_vrev64v16qi (__a, 0)
3814
3815 #define vrev64q_u16(__a) \
3816   (uint16x8_t)__builtin_neon_vrev64v8hi (__a, 0)
3817
3818 #define vrev64q_u32(__a) \
3819   (uint32x4_t)__builtin_neon_vrev64v4si (__a, 0)
3820
3821 #define vrev64q_p8(__a) \
3822   (poly8x16_t)__builtin_neon_vrev64v16qi (__a, 4)
3823
3824 #define vrev64q_p16(__a) \
3825   (poly16x8_t)__builtin_neon_vrev64v8hi (__a, 4)
3826
3827 #define vrev32_s8(__a) \
3828   (int8x8_t)__builtin_neon_vrev32v8qi (__a, 1)
3829
3830 #define vrev32_s16(__a) \
3831   (int16x4_t)__builtin_neon_vrev32v4hi (__a, 1)
3832
3833 #define vrev32_u8(__a) \
3834   (uint8x8_t)__builtin_neon_vrev32v8qi (__a, 0)
3835
3836 #define vrev32_u16(__a) \
3837   (uint16x4_t)__builtin_neon_vrev32v4hi (__a, 0)
3838
3839 #define vrev32_p8(__a) \
3840   (poly8x8_t)__builtin_neon_vrev32v8qi (__a, 4)
3841
3842 #define vrev32_p16(__a) \
3843   (poly16x4_t)__builtin_neon_vrev32v4hi (__a, 4)
3844
3845 #define vrev32q_s8(__a) \
3846   (int8x16_t)__builtin_neon_vrev32v16qi (__a, 1)
3847
3848 #define vrev32q_s16(__a) \
3849   (int16x8_t)__builtin_neon_vrev32v8hi (__a, 1)
3850
3851 #define vrev32q_u8(__a) \
3852   (uint8x16_t)__builtin_neon_vrev32v16qi (__a, 0)
3853
3854 #define vrev32q_u16(__a) \
3855   (uint16x8_t)__builtin_neon_vrev32v8hi (__a, 0)
3856
3857 #define vrev32q_p8(__a) \
3858   (poly8x16_t)__builtin_neon_vrev32v16qi (__a, 4)
3859
3860 #define vrev32q_p16(__a) \
3861   (poly16x8_t)__builtin_neon_vrev32v8hi (__a, 4)
3862
3863 #define vrev16_s8(__a) \
3864   (int8x8_t)__builtin_neon_vrev16v8qi (__a, 1)
3865
3866 #define vrev16_u8(__a) \
3867   (uint8x8_t)__builtin_neon_vrev16v8qi (__a, 0)
3868
3869 #define vrev16_p8(__a) \
3870   (poly8x8_t)__builtin_neon_vrev16v8qi (__a, 4)
3871
3872 #define vrev16q_s8(__a) \
3873   (int8x16_t)__builtin_neon_vrev16v16qi (__a, 1)
3874
3875 #define vrev16q_u8(__a) \
3876   (uint8x16_t)__builtin_neon_vrev16v16qi (__a, 0)
3877
3878 #define vrev16q_p8(__a) \
3879   (poly8x16_t)__builtin_neon_vrev16v16qi (__a, 4)
3880
3881 #define vbsl_s8(__a, __b, __c) \
3882   (int8x8_t)__builtin_neon_vbslv8qi (__a, __b, __c)
3883
3884 #define vbsl_s16(__a, __b, __c) \
3885   (int16x4_t)__builtin_neon_vbslv4hi (__a, __b, __c)
3886
3887 #define vbsl_s32(__a, __b, __c) \
3888   (int32x2_t)__builtin_neon_vbslv2si (__a, __b, __c)
3889
3890 #define vbsl_s64(__a, __b, __c) \
3891   (int64x1_t)__builtin_neon_vbslv1di (__a, __b, __c)
3892
3893 #define vbsl_f32(__a, __b, __c) \
3894   (float32x2_t)__builtin_neon_vbslv2sf (__a, __b, __c)
3895
3896 #define vbsl_u8(__a, __b, __c) \
3897   (uint8x8_t)__builtin_neon_vbslv8qi (__a, __b, __c)
3898
3899 #define vbsl_u16(__a, __b, __c) \
3900   (uint16x4_t)__builtin_neon_vbslv4hi (__a, __b, __c)
3901
3902 #define vbsl_u32(__a, __b, __c) \
3903   (uint32x2_t)__builtin_neon_vbslv2si (__a, __b, __c)
3904
3905 #define vbsl_u64(__a, __b, __c) \
3906   (uint64x1_t)__builtin_neon_vbslv1di (__a, __b, __c)
3907
3908 #define vbsl_p8(__a, __b, __c) \
3909   (poly8x8_t)__builtin_neon_vbslv8qi (__a, __b, __c)
3910
3911 #define vbsl_p16(__a, __b, __c) \
3912   (poly16x4_t)__builtin_neon_vbslv4hi (__a, __b, __c)
3913
3914 #define vbslq_s8(__a, __b, __c) \
3915   (int8x16_t)__builtin_neon_vbslv16qi (__a, __b, __c)
3916
3917 #define vbslq_s16(__a, __b, __c) \
3918   (int16x8_t)__builtin_neon_vbslv8hi (__a, __b, __c)
3919
3920 #define vbslq_s32(__a, __b, __c) \
3921   (int32x4_t)__builtin_neon_vbslv4si (__a, __b, __c)
3922
3923 #define vbslq_s64(__a, __b, __c) \
3924   (int64x2_t)__builtin_neon_vbslv2di (__a, __b, __c)
3925
3926 #define vbslq_f32(__a, __b, __c) \
3927   (float32x4_t)__builtin_neon_vbslv4sf (__a, __b, __c)
3928
3929 #define vbslq_u8(__a, __b, __c) \
3930   (uint8x16_t)__builtin_neon_vbslv16qi (__a, __b, __c)
3931
3932 #define vbslq_u16(__a, __b, __c) \
3933   (uint16x8_t)__builtin_neon_vbslv8hi (__a, __b, __c)
3934
3935 #define vbslq_u32(__a, __b, __c) \
3936   (uint32x4_t)__builtin_neon_vbslv4si (__a, __b, __c)
3937
3938 #define vbslq_u64(__a, __b, __c) \
3939   (uint64x2_t)__builtin_neon_vbslv2di (__a, __b, __c)
3940
3941 #define vbslq_p8(__a, __b, __c) \
3942   (poly8x16_t)__builtin_neon_vbslv16qi (__a, __b, __c)
3943
3944 #define vbslq_p16(__a, __b, __c) \
3945   (poly16x8_t)__builtin_neon_vbslv8hi (__a, __b, __c)
3946
3947 #define vtrn_s8(__a, __b) \
3948   ({ \
3949      union { int8x8x2_t __i; __builtin_neon_v8qi2 __o; } __rv; \
3950      __rv.__o = __builtin_neon_vtrnv8qi (__a, __b); \
3951      __rv.__i; \
3952    })
3953
3954 #define vtrn_s16(__a, __b) \
3955   ({ \
3956      union { int16x4x2_t __i; __builtin_neon_v4hi2 __o; } __rv; \
3957      __rv.__o = __builtin_neon_vtrnv4hi (__a, __b); \
3958      __rv.__i; \
3959    })
3960
3961 #define vtrn_s32(__a, __b) \
3962   ({ \
3963      union { int32x2x2_t __i; __builtin_neon_v2si2 __o; } __rv; \
3964      __rv.__o = __builtin_neon_vtrnv2si (__a, __b); \
3965      __rv.__i; \
3966    })
3967
3968 #define vtrn_f32(__a, __b) \
3969   ({ \
3970      union { float32x2x2_t __i; __builtin_neon_v2sf2 __o; } __rv; \
3971      __rv.__o = __builtin_neon_vtrnv2sf (__a, __b); \
3972      __rv.__i; \
3973    })
3974
3975 #define vtrn_u8(__a, __b) \
3976   ({ \
3977      union { uint8x8x2_t __i; __builtin_neon_v8qi2 __o; } __rv; \
3978      __rv.__o = __builtin_neon_vtrnv8qi (__a, __b); \
3979      __rv.__i; \
3980    })
3981
3982 #define vtrn_u16(__a, __b) \
3983   ({ \
3984      union { uint16x4x2_t __i; __builtin_neon_v4hi2 __o; } __rv; \
3985      __rv.__o = __builtin_neon_vtrnv4hi (__a, __b); \
3986      __rv.__i; \
3987    })
3988
3989 #define vtrn_u32(__a, __b) \
3990   ({ \
3991      union { uint32x2x2_t __i; __builtin_neon_v2si2 __o; } __rv; \
3992      __rv.__o = __builtin_neon_vtrnv2si (__a, __b); \
3993      __rv.__i; \
3994    })
3995
3996 #define vtrn_p8(__a, __b) \
3997   ({ \
3998      union { poly8x8x2_t __i; __builtin_neon_v8qi2 __o; } __rv; \
3999      __rv.__o = __builtin_neon_vtrnv8qi (__a, __b); \
4000      __rv.__i; \
4001    })
4002
4003 #define vtrn_p16(__a, __b) \
4004   ({ \
4005      union { poly16x4x2_t __i; __builtin_neon_v4hi2 __o; } __rv; \
4006      __rv.__o = __builtin_neon_vtrnv4hi (__a, __b); \
4007      __rv.__i; \
4008    })
4009
4010 #define vtrnq_s8(__a, __b) \
4011   ({ \
4012      union { int8x16x2_t __i; __builtin_neon_v16qi2 __o; } __rv; \
4013      __rv.__o = __builtin_neon_vtrnv16qi (__a, __b); \
4014      __rv.__i; \
4015    })
4016
4017 #define vtrnq_s16(__a, __b) \
4018   ({ \
4019      union { int16x8x2_t __i; __builtin_neon_v8hi2 __o; } __rv; \
4020      __rv.__o = __builtin_neon_vtrnv8hi (__a, __b); \
4021      __rv.__i; \
4022    })
4023
4024 #define vtrnq_s32(__a, __b) \
4025   ({ \
4026      union { int32x4x2_t __i; __builtin_neon_v4si2 __o; } __rv; \
4027      __rv.__o = __builtin_neon_vtrnv4si (__a, __b); \
4028      __rv.__i; \
4029    })
4030
4031 #define vtrnq_f32(__a, __b) \
4032   ({ \
4033      union { float32x4x2_t __i; __builtin_neon_v4sf2 __o; } __rv; \
4034      __rv.__o = __builtin_neon_vtrnv4sf (__a, __b); \
4035      __rv.__i; \
4036    })
4037
4038 #define vtrnq_u8(__a, __b) \
4039   ({ \
4040      union { uint8x16x2_t __i; __builtin_neon_v16qi2 __o; } __rv; \
4041      __rv.__o = __builtin_neon_vtrnv16qi (__a, __b); \
4042      __rv.__i; \
4043    })
4044
4045 #define vtrnq_u16(__a, __b) \
4046   ({ \
4047      union { uint16x8x2_t __i; __builtin_neon_v8hi2 __o; } __rv; \
4048      __rv.__o = __builtin_neon_vtrnv8hi (__a, __b); \
4049      __rv.__i; \
4050    })
4051
4052 #define vtrnq_u32(__a, __b) \
4053   ({ \
4054      union { uint32x4x2_t __i; __builtin_neon_v4si2 __o; } __rv; \
4055      __rv.__o = __builtin_neon_vtrnv4si (__a, __b); \
4056      __rv.__i; \
4057    })
4058
4059 #define vtrnq_p8(__a, __b) \
4060   ({ \
4061      union { poly8x16x2_t __i; __builtin_neon_v16qi2 __o; } __rv; \
4062      __rv.__o = __builtin_neon_vtrnv16qi (__a, __b); \
4063      __rv.__i; \
4064    })
4065
4066 #define vtrnq_p16(__a, __b) \
4067   ({ \
4068      union { poly16x8x2_t __i; __builtin_neon_v8hi2 __o; } __rv; \
4069      __rv.__o = __builtin_neon_vtrnv8hi (__a, __b); \
4070      __rv.__i; \
4071    })
4072
4073 #define vzip_s8(__a, __b) \
4074   ({ \
4075      union { int8x8x2_t __i; __builtin_neon_v8qi2 __o; } __rv; \
4076      __rv.__o = __builtin_neon_vzipv8qi (__a, __b); \
4077      __rv.__i; \
4078    })
4079
4080 #define vzip_s16(__a, __b) \
4081   ({ \
4082      union { int16x4x2_t __i; __builtin_neon_v4hi2 __o; } __rv; \
4083      __rv.__o = __builtin_neon_vzipv4hi (__a, __b); \
4084      __rv.__i; \
4085    })
4086
4087 #define vzip_s32(__a, __b) \
4088   ({ \
4089      union { int32x2x2_t __i; __builtin_neon_v2si2 __o; } __rv; \
4090      __rv.__o = __builtin_neon_vzipv2si (__a, __b); \
4091      __rv.__i; \
4092    })
4093
4094 #define vzip_f32(__a, __b) \
4095   ({ \
4096      union { float32x2x2_t __i; __builtin_neon_v2sf2 __o; } __rv; \
4097      __rv.__o = __builtin_neon_vzipv2sf (__a, __b); \
4098      __rv.__i; \
4099    })
4100
4101 #define vzip_u8(__a, __b) \
4102   ({ \
4103      union { uint8x8x2_t __i; __builtin_neon_v8qi2 __o; } __rv; \
4104      __rv.__o = __builtin_neon_vzipv8qi (__a, __b); \
4105      __rv.__i; \
4106    })
4107
4108 #define vzip_u16(__a, __b) \
4109   ({ \
4110      union { uint16x4x2_t __i; __builtin_neon_v4hi2 __o; } __rv; \
4111      __rv.__o = __builtin_neon_vzipv4hi (__a, __b); \
4112      __rv.__i; \
4113    })
4114
4115 #define vzip_u32(__a, __b) \
4116   ({ \
4117      union { uint32x2x2_t __i; __builtin_neon_v2si2 __o; } __rv; \
4118      __rv.__o = __builtin_neon_vzipv2si (__a, __b); \
4119      __rv.__i; \
4120    })
4121
4122 #define vzip_p8(__a, __b) \
4123   ({ \
4124      union { poly8x8x2_t __i; __builtin_neon_v8qi2 __o; } __rv; \
4125      __rv.__o = __builtin_neon_vzipv8qi (__a, __b); \
4126      __rv.__i; \
4127    })
4128
4129 #define vzip_p16(__a, __b) \
4130   ({ \
4131      union { poly16x4x2_t __i; __builtin_neon_v4hi2 __o; } __rv; \
4132      __rv.__o = __builtin_neon_vzipv4hi (__a, __b); \
4133      __rv.__i; \
4134    })
4135
4136 #define vzipq_s8(__a, __b) \
4137   ({ \
4138      union { int8x16x2_t __i; __builtin_neon_v16qi2 __o; } __rv; \
4139      __rv.__o = __builtin_neon_vzipv16qi (__a, __b); \
4140      __rv.__i; \
4141    })
4142
4143 #define vzipq_s16(__a, __b) \
4144   ({ \
4145      union { int16x8x2_t __i; __builtin_neon_v8hi2 __o; } __rv; \
4146      __rv.__o = __builtin_neon_vzipv8hi (__a, __b); \
4147      __rv.__i; \
4148    })
4149
4150 #define vzipq_s32(__a, __b) \
4151   ({ \
4152      union { int32x4x2_t __i; __builtin_neon_v4si2 __o; } __rv; \
4153      __rv.__o = __builtin_neon_vzipv4si (__a, __b); \
4154      __rv.__i; \
4155    })
4156
4157 #define vzipq_f32(__a, __b) \
4158   ({ \
4159      union { float32x4x2_t __i; __builtin_neon_v4sf2 __o; } __rv; \
4160      __rv.__o = __builtin_neon_vzipv4sf (__a, __b); \
4161      __rv.__i; \
4162    })
4163
4164 #define vzipq_u8(__a, __b) \
4165   ({ \
4166      union { uint8x16x2_t __i; __builtin_neon_v16qi2 __o; } __rv; \
4167      __rv.__o = __builtin_neon_vzipv16qi (__a, __b); \
4168      __rv.__i; \
4169    })
4170
4171 #define vzipq_u16(__a, __b) \
4172   ({ \
4173      union { uint16x8x2_t __i; __builtin_neon_v8hi2 __o; } __rv; \
4174      __rv.__o = __builtin_neon_vzipv8hi (__a, __b); \
4175      __rv.__i; \
4176    })
4177
4178 #define vzipq_u32(__a, __b) \
4179   ({ \
4180      union { uint32x4x2_t __i; __builtin_neon_v4si2 __o; } __rv; \
4181      __rv.__o = __builtin_neon_vzipv4si (__a, __b); \
4182      __rv.__i; \
4183    })
4184
4185 #define vzipq_p8(__a, __b) \
4186   ({ \
4187      union { poly8x16x2_t __i; __builtin_neon_v16qi2 __o; } __rv; \
4188      __rv.__o = __builtin_neon_vzipv16qi (__a, __b); \
4189      __rv.__i; \
4190    })
4191
4192 #define vzipq_p16(__a, __b) \
4193   ({ \
4194      union { poly16x8x2_t __i; __builtin_neon_v8hi2 __o; } __rv; \
4195      __rv.__o = __builtin_neon_vzipv8hi (__a, __b); \
4196      __rv.__i; \
4197    })
4198
4199 #define vuzp_s8(__a, __b) \
4200   ({ \
4201      union { int8x8x2_t __i; __builtin_neon_v8qi2 __o; } __rv; \
4202      __rv.__o = __builtin_neon_vuzpv8qi (__a, __b); \
4203      __rv.__i; \
4204    })
4205
4206 #define vuzp_s16(__a, __b) \
4207   ({ \
4208      union { int16x4x2_t __i; __builtin_neon_v4hi2 __o; } __rv; \
4209      __rv.__o = __builtin_neon_vuzpv4hi (__a, __b); \
4210      __rv.__i; \
4211    })
4212
4213 #define vuzp_s32(__a, __b) \
4214   ({ \
4215      union { int32x2x2_t __i; __builtin_neon_v2si2 __o; } __rv; \
4216      __rv.__o = __builtin_neon_vuzpv2si (__a, __b); \
4217      __rv.__i; \
4218    })
4219
4220 #define vuzp_f32(__a, __b) \
4221   ({ \
4222      union { float32x2x2_t __i; __builtin_neon_v2sf2 __o; } __rv; \
4223      __rv.__o = __builtin_neon_vuzpv2sf (__a, __b); \
4224      __rv.__i; \
4225    })
4226
4227 #define vuzp_u8(__a, __b) \
4228   ({ \
4229      union { uint8x8x2_t __i; __builtin_neon_v8qi2 __o; } __rv; \
4230      __rv.__o = __builtin_neon_vuzpv8qi (__a, __b); \
4231      __rv.__i; \
4232    })
4233
4234 #define vuzp_u16(__a, __b) \
4235   ({ \
4236      union { uint16x4x2_t __i; __builtin_neon_v4hi2 __o; } __rv; \
4237      __rv.__o = __builtin_neon_vuzpv4hi (__a, __b); \
4238      __rv.__i; \
4239    })
4240
4241 #define vuzp_u32(__a, __b) \
4242   ({ \
4243      union { uint32x2x2_t __i; __builtin_neon_v2si2 __o; } __rv; \
4244      __rv.__o = __builtin_neon_vuzpv2si (__a, __b); \
4245      __rv.__i; \
4246    })
4247
4248 #define vuzp_p8(__a, __b) \
4249   ({ \
4250      union { poly8x8x2_t __i; __builtin_neon_v8qi2 __o; } __rv; \
4251      __rv.__o = __builtin_neon_vuzpv8qi (__a, __b); \
4252      __rv.__i; \
4253    })
4254
4255 #define vuzp_p16(__a, __b) \
4256   ({ \
4257      union { poly16x4x2_t __i; __builtin_neon_v4hi2 __o; } __rv; \
4258      __rv.__o = __builtin_neon_vuzpv4hi (__a, __b); \
4259      __rv.__i; \
4260    })
4261
4262 #define vuzpq_s8(__a, __b) \
4263   ({ \
4264      union { int8x16x2_t __i; __builtin_neon_v16qi2 __o; } __rv; \
4265      __rv.__o = __builtin_neon_vuzpv16qi (__a, __b); \
4266      __rv.__i; \
4267    })
4268
4269 #define vuzpq_s16(__a, __b) \
4270   ({ \
4271      union { int16x8x2_t __i; __builtin_neon_v8hi2 __o; } __rv; \
4272      __rv.__o = __builtin_neon_vuzpv8hi (__a, __b); \
4273      __rv.__i; \
4274    })
4275
4276 #define vuzpq_s32(__a, __b) \
4277   ({ \
4278      union { int32x4x2_t __i; __builtin_neon_v4si2 __o; } __rv; \
4279      __rv.__o = __builtin_neon_vuzpv4si (__a, __b); \
4280      __rv.__i; \
4281    })
4282
4283 #define vuzpq_f32(__a, __b) \
4284   ({ \
4285      union { float32x4x2_t __i; __builtin_neon_v4sf2 __o; } __rv; \
4286      __rv.__o = __builtin_neon_vuzpv4sf (__a, __b); \
4287      __rv.__i; \
4288    })
4289
4290 #define vuzpq_u8(__a, __b) \
4291   ({ \
4292      union { uint8x16x2_t __i; __builtin_neon_v16qi2 __o; } __rv; \
4293      __rv.__o = __builtin_neon_vuzpv16qi (__a, __b); \
4294      __rv.__i; \
4295    })
4296
4297 #define vuzpq_u16(__a, __b) \
4298   ({ \
4299      union { uint16x8x2_t __i; __builtin_neon_v8hi2 __o; } __rv; \
4300      __rv.__o = __builtin_neon_vuzpv8hi (__a, __b); \
4301      __rv.__i; \
4302    })
4303
4304 #define vuzpq_u32(__a, __b) \
4305   ({ \
4306      union { uint32x4x2_t __i; __builtin_neon_v4si2 __o; } __rv; \
4307      __rv.__o = __builtin_neon_vuzpv4si (__a, __b); \
4308      __rv.__i; \
4309    })
4310
4311 #define vuzpq_p8(__a, __b) \
4312   ({ \
4313      union { poly8x16x2_t __i; __builtin_neon_v16qi2 __o; } __rv; \
4314      __rv.__o = __builtin_neon_vuzpv16qi (__a, __b); \
4315      __rv.__i; \
4316    })
4317
4318 #define vuzpq_p16(__a, __b) \
4319   ({ \
4320      union { poly16x8x2_t __i; __builtin_neon_v8hi2 __o; } __rv; \
4321      __rv.__o = __builtin_neon_vuzpv8hi (__a, __b); \
4322      __rv.__i; \
4323    })
4324
4325 #define vld1_s8(__a) \
4326   (int8x8_t)__builtin_neon_vld1v8qi (__a)
4327
4328 #define vld1_s16(__a) \
4329   (int16x4_t)__builtin_neon_vld1v4hi (__a)
4330
4331 #define vld1_s32(__a) \
4332   (int32x2_t)__builtin_neon_vld1v2si (__a)
4333
4334 #define vld1_s64(__a) \
4335   (int64x1_t)__builtin_neon_vld1v1di (__a)
4336
4337 #define vld1_f32(__a) \
4338   (float32x2_t)__builtin_neon_vld1v2sf (__a)
4339
4340 #define vld1_u8(__a) \
4341   (uint8x8_t)__builtin_neon_vld1v8qi (__a)
4342
4343 #define vld1_u16(__a) \
4344   (uint16x4_t)__builtin_neon_vld1v4hi (__a)
4345
4346 #define vld1_u32(__a) \
4347   (uint32x2_t)__builtin_neon_vld1v2si (__a)
4348
4349 #define vld1_u64(__a) \
4350   (uint64x1_t)__builtin_neon_vld1v1di (__a)
4351
4352 #define vld1_p8(__a) \
4353   (poly8x8_t)__builtin_neon_vld1v8qi (__a)
4354
4355 #define vld1_p16(__a) \
4356   (poly16x4_t)__builtin_neon_vld1v4hi (__a)
4357
4358 #define vld1q_s8(__a) \
4359   (int8x16_t)__builtin_neon_vld1v16qi (__a)
4360
4361 #define vld1q_s16(__a) \
4362   (int16x8_t)__builtin_neon_vld1v8hi (__a)
4363
4364 #define vld1q_s32(__a) \
4365   (int32x4_t)__builtin_neon_vld1v4si (__a)
4366
4367 #define vld1q_s64(__a) \
4368   (int64x2_t)__builtin_neon_vld1v2di (__a)
4369
4370 #define vld1q_f32(__a) \
4371   (float32x4_t)__builtin_neon_vld1v4sf (__a)
4372
4373 #define vld1q_u8(__a) \
4374   (uint8x16_t)__builtin_neon_vld1v16qi (__a)
4375
4376 #define vld1q_u16(__a) \
4377   (uint16x8_t)__builtin_neon_vld1v8hi (__a)
4378
4379 #define vld1q_u32(__a) \
4380   (uint32x4_t)__builtin_neon_vld1v4si (__a)
4381
4382 #define vld1q_u64(__a) \
4383   (uint64x2_t)__builtin_neon_vld1v2di (__a)
4384
4385 #define vld1q_p8(__a) \
4386   (poly8x16_t)__builtin_neon_vld1v16qi (__a)
4387
4388 #define vld1q_p16(__a) \
4389   (poly16x8_t)__builtin_neon_vld1v8hi (__a)
4390
4391 #define vld1_lane_s8(__a, __b, __c) \
4392   (int8x8_t)__builtin_neon_vld1_lanev8qi (__a, __b, __c)
4393
4394 #define vld1_lane_s16(__a, __b, __c) \
4395   (int16x4_t)__builtin_neon_vld1_lanev4hi (__a, __b, __c)
4396
4397 #define vld1_lane_s32(__a, __b, __c) \
4398   (int32x2_t)__builtin_neon_vld1_lanev2si (__a, __b, __c)
4399
4400 #define vld1_lane_f32(__a, __b, __c) \
4401   (float32x2_t)__builtin_neon_vld1_lanev2sf (__a, __b, __c)
4402
4403 #define vld1_lane_u8(__a, __b, __c) \
4404   (uint8x8_t)__builtin_neon_vld1_lanev8qi (__a, __b, __c)
4405
4406 #define vld1_lane_u16(__a, __b, __c) \
4407   (uint16x4_t)__builtin_neon_vld1_lanev4hi (__a, __b, __c)
4408
4409 #define vld1_lane_u32(__a, __b, __c) \
4410   (uint32x2_t)__builtin_neon_vld1_lanev2si (__a, __b, __c)
4411
4412 #define vld1_lane_p8(__a, __b, __c) \
4413   (poly8x8_t)__builtin_neon_vld1_lanev8qi (__a, __b, __c)
4414
4415 #define vld1_lane_p16(__a, __b, __c) \
4416   (poly16x4_t)__builtin_neon_vld1_lanev4hi (__a, __b, __c)
4417
4418 #define vld1_lane_s64(__a, __b, __c) \
4419   (int64x1_t)__builtin_neon_vld1_lanev1di (__a, __b, __c)
4420
4421 #define vld1_lane_u64(__a, __b, __c) \
4422   (uint64x1_t)__builtin_neon_vld1_lanev1di (__a, __b, __c)
4423
4424 #define vld1q_lane_s8(__a, __b, __c) \
4425   (int8x16_t)__builtin_neon_vld1_lanev16qi (__a, __b, __c)
4426
4427 #define vld1q_lane_s16(__a, __b, __c) \
4428   (int16x8_t)__builtin_neon_vld1_lanev8hi (__a, __b, __c)
4429
4430 #define vld1q_lane_s32(__a, __b, __c) \
4431   (int32x4_t)__builtin_neon_vld1_lanev4si (__a, __b, __c)
4432
4433 #define vld1q_lane_f32(__a, __b, __c) \
4434   (float32x4_t)__builtin_neon_vld1_lanev4sf (__a, __b, __c)
4435
4436 #define vld1q_lane_u8(__a, __b, __c) \
4437   (uint8x16_t)__builtin_neon_vld1_lanev16qi (__a, __b, __c)
4438
4439 #define vld1q_lane_u16(__a, __b, __c) \
4440   (uint16x8_t)__builtin_neon_vld1_lanev8hi (__a, __b, __c)
4441
4442 #define vld1q_lane_u32(__a, __b, __c) \
4443   (uint32x4_t)__builtin_neon_vld1_lanev4si (__a, __b, __c)
4444
4445 #define vld1q_lane_p8(__a, __b, __c) \
4446   (poly8x16_t)__builtin_neon_vld1_lanev16qi (__a, __b, __c)
4447
4448 #define vld1q_lane_p16(__a, __b, __c) \
4449   (poly16x8_t)__builtin_neon_vld1_lanev8hi (__a, __b, __c)
4450
4451 #define vld1q_lane_s64(__a, __b, __c) \
4452   (int64x2_t)__builtin_neon_vld1_lanev2di (__a, __b, __c)
4453
4454 #define vld1q_lane_u64(__a, __b, __c) \
4455   (uint64x2_t)__builtin_neon_vld1_lanev2di (__a, __b, __c)
4456
4457 #define vld1_dup_s8(__a) \
4458   (int8x8_t)__builtin_neon_vld1_dupv8qi (__a)
4459
4460 #define vld1_dup_s16(__a) \
4461   (int16x4_t)__builtin_neon_vld1_dupv4hi (__a)
4462
4463 #define vld1_dup_s32(__a) \
4464   (int32x2_t)__builtin_neon_vld1_dupv2si (__a)
4465
4466 #define vld1_dup_f32(__a) \
4467   (float32x2_t)__builtin_neon_vld1_dupv2sf (__a)
4468
4469 #define vld1_dup_u8(__a) \
4470   (uint8x8_t)__builtin_neon_vld1_dupv8qi (__a)
4471
4472 #define vld1_dup_u16(__a) \
4473   (uint16x4_t)__builtin_neon_vld1_dupv4hi (__a)
4474
4475 #define vld1_dup_u32(__a) \
4476   (uint32x2_t)__builtin_neon_vld1_dupv2si (__a)
4477
4478 #define vld1_dup_p8(__a) \
4479   (poly8x8_t)__builtin_neon_vld1_dupv8qi (__a)
4480
4481 #define vld1_dup_p16(__a) \
4482   (poly16x4_t)__builtin_neon_vld1_dupv4hi (__a)
4483
4484 #define vld1_dup_s64(__a) \
4485   (int64x1_t)__builtin_neon_vld1_dupv1di (__a)
4486
4487 #define vld1_dup_u64(__a) \
4488   (uint64x1_t)__builtin_neon_vld1_dupv1di (__a)
4489
4490 #define vld1q_dup_s8(__a) \
4491   (int8x16_t)__builtin_neon_vld1_dupv16qi (__a)
4492
4493 #define vld1q_dup_s16(__a) \
4494   (int16x8_t)__builtin_neon_vld1_dupv8hi (__a)
4495
4496 #define vld1q_dup_s32(__a) \
4497   (int32x4_t)__builtin_neon_vld1_dupv4si (__a)
4498
4499 #define vld1q_dup_f32(__a) \
4500   (float32x4_t)__builtin_neon_vld1_dupv4sf (__a)
4501
4502 #define vld1q_dup_u8(__a) \
4503   (uint8x16_t)__builtin_neon_vld1_dupv16qi (__a)
4504
4505 #define vld1q_dup_u16(__a) \
4506   (uint16x8_t)__builtin_neon_vld1_dupv8hi (__a)
4507
4508 #define vld1q_dup_u32(__a) \
4509   (uint32x4_t)__builtin_neon_vld1_dupv4si (__a)
4510
4511 #define vld1q_dup_p8(__a) \
4512   (poly8x16_t)__builtin_neon_vld1_dupv16qi (__a)
4513
4514 #define vld1q_dup_p16(__a) \
4515   (poly16x8_t)__builtin_neon_vld1_dupv8hi (__a)
4516
4517 #define vld1q_dup_s64(__a) \
4518   (int64x2_t)__builtin_neon_vld1_dupv2di (__a)
4519
4520 #define vld1q_dup_u64(__a) \
4521   (uint64x2_t)__builtin_neon_vld1_dupv2di (__a)
4522
4523 #define vst1_s8(__a, __b) \
4524   __builtin_neon_vst1v8qi (__a, __b)
4525
4526 #define vst1_s16(__a, __b) \
4527   __builtin_neon_vst1v4hi (__a, __b)
4528
4529 #define vst1_s32(__a, __b) \
4530   __builtin_neon_vst1v2si (__a, __b)
4531
4532 #define vst1_s64(__a, __b) \
4533   __builtin_neon_vst1v1di (__a, __b)
4534
4535 #define vst1_f32(__a, __b) \
4536   __builtin_neon_vst1v2sf (__a, __b)
4537
4538 #define vst1_u8(__a, __b) \
4539   __builtin_neon_vst1v8qi (__a, __b)
4540
4541 #define vst1_u16(__a, __b) \
4542   __builtin_neon_vst1v4hi (__a, __b)
4543
4544 #define vst1_u32(__a, __b) \
4545   __builtin_neon_vst1v2si (__a, __b)
4546
4547 #define vst1_u64(__a, __b) \
4548   __builtin_neon_vst1v1di (__a, __b)
4549
4550 #define vst1_p8(__a, __b) \
4551   __builtin_neon_vst1v8qi (__a, __b)
4552
4553 #define vst1_p16(__a, __b) \
4554   __builtin_neon_vst1v4hi (__a, __b)
4555
4556 #define vst1q_s8(__a, __b) \
4557   __builtin_neon_vst1v16qi (__a, __b)
4558
4559 #define vst1q_s16(__a, __b) \
4560   __builtin_neon_vst1v8hi (__a, __b)
4561
4562 #define vst1q_s32(__a, __b) \
4563   __builtin_neon_vst1v4si (__a, __b)
4564
4565 #define vst1q_s64(__a, __b) \
4566   __builtin_neon_vst1v2di (__a, __b)
4567
4568 #define vst1q_f32(__a, __b) \
4569   __builtin_neon_vst1v4sf (__a, __b)
4570
4571 #define vst1q_u8(__a, __b) \
4572   __builtin_neon_vst1v16qi (__a, __b)
4573
4574 #define vst1q_u16(__a, __b) \
4575   __builtin_neon_vst1v8hi (__a, __b)
4576
4577 #define vst1q_u32(__a, __b) \
4578   __builtin_neon_vst1v4si (__a, __b)
4579
4580 #define vst1q_u64(__a, __b) \
4581   __builtin_neon_vst1v2di (__a, __b)
4582
4583 #define vst1q_p8(__a, __b) \
4584   __builtin_neon_vst1v16qi (__a, __b)
4585
4586 #define vst1q_p16(__a, __b) \
4587   __builtin_neon_vst1v8hi (__a, __b)
4588
4589 #define vst1_lane_s8(__a, __b, __c) \
4590   __builtin_neon_vst1_lanev8qi (__a, __b, __c)
4591
4592 #define vst1_lane_s16(__a, __b, __c) \
4593   __builtin_neon_vst1_lanev4hi (__a, __b, __c)
4594
4595 #define vst1_lane_s32(__a, __b, __c) \
4596   __builtin_neon_vst1_lanev2si (__a, __b, __c)
4597
4598 #define vst1_lane_f32(__a, __b, __c) \
4599   __builtin_neon_vst1_lanev2sf (__a, __b, __c)
4600
4601 #define vst1_lane_u8(__a, __b, __c) \
4602   __builtin_neon_vst1_lanev8qi (__a, __b, __c)
4603
4604 #define vst1_lane_u16(__a, __b, __c) \
4605   __builtin_neon_vst1_lanev4hi (__a, __b, __c)
4606
4607 #define vst1_lane_u32(__a, __b, __c) \
4608   __builtin_neon_vst1_lanev2si (__a, __b, __c)
4609
4610 #define vst1_lane_p8(__a, __b, __c) \
4611   __builtin_neon_vst1_lanev8qi (__a, __b, __c)
4612
4613 #define vst1_lane_p16(__a, __b, __c) \
4614   __builtin_neon_vst1_lanev4hi (__a, __b, __c)
4615
4616 #define vst1_lane_s64(__a, __b, __c) \
4617   __builtin_neon_vst1_lanev1di (__a, __b, __c)
4618
4619 #define vst1_lane_u64(__a, __b, __c) \
4620   __builtin_neon_vst1_lanev1di (__a, __b, __c)
4621
4622 #define vst1q_lane_s8(__a, __b, __c) \
4623   __builtin_neon_vst1_lanev16qi (__a, __b, __c)
4624
4625 #define vst1q_lane_s16(__a, __b, __c) \
4626   __builtin_neon_vst1_lanev8hi (__a, __b, __c)
4627
4628 #define vst1q_lane_s32(__a, __b, __c) \
4629   __builtin_neon_vst1_lanev4si (__a, __b, __c)
4630
4631 #define vst1q_lane_f32(__a, __b, __c) \
4632   __builtin_neon_vst1_lanev4sf (__a, __b, __c)
4633
4634 #define vst1q_lane_u8(__a, __b, __c) \
4635   __builtin_neon_vst1_lanev16qi (__a, __b, __c)
4636
4637 #define vst1q_lane_u16(__a, __b, __c) \
4638   __builtin_neon_vst1_lanev8hi (__a, __b, __c)
4639
4640 #define vst1q_lane_u32(__a, __b, __c) \
4641   __builtin_neon_vst1_lanev4si (__a, __b, __c)
4642
4643 #define vst1q_lane_p8(__a, __b, __c) \
4644   __builtin_neon_vst1_lanev16qi (__a, __b, __c)
4645
4646 #define vst1q_lane_p16(__a, __b, __c) \
4647   __builtin_neon_vst1_lanev8hi (__a, __b, __c)
4648
4649 #define vst1q_lane_s64(__a, __b, __c) \
4650   __builtin_neon_vst1_lanev2di (__a, __b, __c)
4651
4652 #define vst1q_lane_u64(__a, __b, __c) \
4653   __builtin_neon_vst1_lanev2di (__a, __b, __c)
4654
4655 #define vld2_s8(__a) \
4656   ({ \
4657      union { int8x8x2_t __i; __builtin_neon_v8qi2 __o; } __rv; \
4658      __rv.__o = __builtin_neon_vld2v8qi (__a); \
4659      __rv.__i; \
4660    })
4661
4662 #define vld2_s16(__a) \
4663   ({ \
4664      union { int16x4x2_t __i; __builtin_neon_v4hi2 __o; } __rv; \
4665      __rv.__o = __builtin_neon_vld2v4hi (__a); \
4666      __rv.__i; \
4667    })
4668
4669 #define vld2_s32(__a) \
4670   ({ \
4671      union { int32x2x2_t __i; __builtin_neon_v2si2 __o; } __rv; \
4672      __rv.__o = __builtin_neon_vld2v2si (__a); \
4673      __rv.__i; \
4674    })
4675
4676 #define vld2_f32(__a) \
4677   ({ \
4678      union { float32x2x2_t __i; __builtin_neon_v2sf2 __o; } __rv; \
4679      __rv.__o = __builtin_neon_vld2v2sf (__a); \
4680      __rv.__i; \
4681    })
4682
4683 #define vld2_u8(__a) \
4684   ({ \
4685      union { uint8x8x2_t __i; __builtin_neon_v8qi2 __o; } __rv; \
4686      __rv.__o = __builtin_neon_vld2v8qi (__a); \
4687      __rv.__i; \
4688    })
4689
4690 #define vld2_u16(__a) \
4691   ({ \
4692      union { uint16x4x2_t __i; __builtin_neon_v4hi2 __o; } __rv; \
4693      __rv.__o = __builtin_neon_vld2v4hi (__a); \
4694      __rv.__i; \
4695    })
4696
4697 #define vld2_u32(__a) \
4698   ({ \
4699      union { uint32x2x2_t __i; __builtin_neon_v2si2 __o; } __rv; \
4700      __rv.__o = __builtin_neon_vld2v2si (__a); \
4701      __rv.__i; \
4702    })
4703
4704 #define vld2_p8(__a) \
4705   ({ \
4706      union { poly8x8x2_t __i; __builtin_neon_v8qi2 __o; } __rv; \
4707      __rv.__o = __builtin_neon_vld2v8qi (__a); \
4708      __rv.__i; \
4709    })
4710
4711 #define vld2_p16(__a) \
4712   ({ \
4713      union { poly16x4x2_t __i; __builtin_neon_v4hi2 __o; } __rv; \
4714      __rv.__o = __builtin_neon_vld2v4hi (__a); \
4715      __rv.__i; \
4716    })
4717
4718 #define vld2_s64(__a) \
4719   ({ \
4720      union { int64x1x2_t __i; __builtin_neon_v1di2 __o; } __rv; \
4721      __rv.__o = __builtin_neon_vld2v1di (__a); \
4722      __rv.__i; \
4723    })
4724
4725 #define vld2_u64(__a) \
4726   ({ \
4727      union { uint64x1x2_t __i; __builtin_neon_v1di2 __o; } __rv; \
4728      __rv.__o = __builtin_neon_vld2v1di (__a); \
4729      __rv.__i; \
4730    })
4731
4732 #define vld2q_s8(__a) \
4733   ({ \
4734      union { int8x16x2_t __i; __builtin_neon_v16qi2 __o; } __rv; \
4735      __rv.__o = __builtin_neon_vld2v16qi (__a); \
4736      __rv.__i; \
4737    })
4738
4739 #define vld2q_s16(__a) \
4740   ({ \
4741      union { int16x8x2_t __i; __builtin_neon_v8hi2 __o; } __rv; \
4742      __rv.__o = __builtin_neon_vld2v8hi (__a); \
4743      __rv.__i; \
4744    })
4745
4746 #define vld2q_s32(__a) \
4747   ({ \
4748      union { int32x4x2_t __i; __builtin_neon_v4si2 __o; } __rv; \
4749      __rv.__o = __builtin_neon_vld2v4si (__a); \
4750      __rv.__i; \
4751    })
4752
4753 #define vld2q_f32(__a) \
4754   ({ \
4755      union { float32x4x2_t __i; __builtin_neon_v4sf2 __o; } __rv; \
4756      __rv.__o = __builtin_neon_vld2v4sf (__a); \
4757      __rv.__i; \
4758    })
4759
4760 #define vld2q_u8(__a) \
4761   ({ \
4762      union { uint8x16x2_t __i; __builtin_neon_v16qi2 __o; } __rv; \
4763      __rv.__o = __builtin_neon_vld2v16qi (__a); \
4764      __rv.__i; \
4765    })
4766
4767 #define vld2q_u16(__a) \
4768   ({ \
4769      union { uint16x8x2_t __i; __builtin_neon_v8hi2 __o; } __rv; \
4770      __rv.__o = __builtin_neon_vld2v8hi (__a); \
4771      __rv.__i; \
4772    })
4773
4774 #define vld2q_u32(__a) \
4775   ({ \
4776      union { uint32x4x2_t __i; __builtin_neon_v4si2 __o; } __rv; \
4777      __rv.__o = __builtin_neon_vld2v4si (__a); \
4778      __rv.__i; \
4779    })
4780
4781 #define vld2q_p8(__a) \
4782   ({ \
4783      union { poly8x16x2_t __i; __builtin_neon_v16qi2 __o; } __rv; \
4784      __rv.__o = __builtin_neon_vld2v16qi (__a); \
4785      __rv.__i; \
4786    })
4787
4788 #define vld2q_p16(__a) \
4789   ({ \
4790      union { poly16x8x2_t __i; __builtin_neon_v8hi2 __o; } __rv; \
4791      __rv.__o = __builtin_neon_vld2v8hi (__a); \
4792      __rv.__i; \
4793    })
4794
4795 #define vld2_lane_s8(__a, __b, __c) \
4796   ({ \
4797      union { int8x8x2_t __i; __builtin_neon_v8qi2 __o; } __bu = { __b }; \
4798      union { int8x8x2_t __i; __builtin_neon_v8qi2 __o; } __rv; \
4799      __rv.__o = __builtin_neon_vld2_lanev8qi (__a, __bu.__o, __c); \
4800      __rv.__i; \
4801    })
4802
4803 #define vld2_lane_s16(__a, __b, __c) \
4804   ({ \
4805      union { int16x4x2_t __i; __builtin_neon_v4hi2 __o; } __bu = { __b }; \
4806      union { int16x4x2_t __i; __builtin_neon_v4hi2 __o; } __rv; \
4807      __rv.__o = __builtin_neon_vld2_lanev4hi (__a, __bu.__o, __c); \
4808      __rv.__i; \
4809    })
4810
4811 #define vld2_lane_s32(__a, __b, __c) \
4812   ({ \
4813      union { int32x2x2_t __i; __builtin_neon_v2si2 __o; } __bu = { __b }; \
4814      union { int32x2x2_t __i; __builtin_neon_v2si2 __o; } __rv; \
4815      __rv.__o = __builtin_neon_vld2_lanev2si (__a, __bu.__o, __c); \
4816      __rv.__i; \
4817    })
4818
4819 #define vld2_lane_f32(__a, __b, __c) \
4820   ({ \
4821      union { float32x2x2_t __i; __builtin_neon_v2sf2 __o; } __bu = { __b }; \
4822      union { float32x2x2_t __i; __builtin_neon_v2sf2 __o; } __rv; \
4823      __rv.__o = __builtin_neon_vld2_lanev2sf (__a, __bu.__o, __c); \
4824      __rv.__i; \
4825    })
4826
4827 #define vld2_lane_u8(__a, __b, __c) \
4828   ({ \
4829      union { uint8x8x2_t __i; __builtin_neon_v8qi2 __o; } __bu = { __b }; \
4830      union { uint8x8x2_t __i; __builtin_neon_v8qi2 __o; } __rv; \
4831      __rv.__o = __builtin_neon_vld2_lanev8qi (__a, __bu.__o, __c); \
4832      __rv.__i; \
4833    })
4834
4835 #define vld2_lane_u16(__a, __b, __c) \
4836   ({ \
4837      union { uint16x4x2_t __i; __builtin_neon_v4hi2 __o; } __bu = { __b }; \
4838      union { uint16x4x2_t __i; __builtin_neon_v4hi2 __o; } __rv; \
4839      __rv.__o = __builtin_neon_vld2_lanev4hi (__a, __bu.__o, __c); \
4840      __rv.__i; \
4841    })
4842
4843 #define vld2_lane_u32(__a, __b, __c) \
4844   ({ \
4845      union { uint32x2x2_t __i; __builtin_neon_v2si2 __o; } __bu = { __b }; \
4846      union { uint32x2x2_t __i; __builtin_neon_v2si2 __o; } __rv; \
4847      __rv.__o = __builtin_neon_vld2_lanev2si (__a, __bu.__o, __c); \
4848      __rv.__i; \
4849    })
4850
4851 #define vld2_lane_p8(__a, __b, __c) \
4852   ({ \
4853      union { poly8x8x2_t __i; __builtin_neon_v8qi2 __o; } __bu = { __b }; \
4854      union { poly8x8x2_t __i; __builtin_neon_v8qi2 __o; } __rv; \
4855      __rv.__o = __builtin_neon_vld2_lanev8qi (__a, __bu.__o, __c); \
4856      __rv.__i; \
4857    })
4858
4859 #define vld2_lane_p16(__a, __b, __c) \
4860   ({ \
4861      union { poly16x4x2_t __i; __builtin_neon_v4hi2 __o; } __bu = { __b }; \
4862      union { poly16x4x2_t __i; __builtin_neon_v4hi2 __o; } __rv; \
4863      __rv.__o = __builtin_neon_vld2_lanev4hi (__a, __bu.__o, __c); \
4864      __rv.__i; \
4865    })
4866
4867 #define vld2q_lane_s16(__a, __b, __c) \
4868   ({ \
4869      union { int16x8x2_t __i; __builtin_neon_v8hi2 __o; } __bu = { __b }; \
4870      union { int16x8x2_t __i; __builtin_neon_v8hi2 __o; } __rv; \
4871      __rv.__o = __builtin_neon_vld2_lanev8hi (__a, __bu.__o, __c); \
4872      __rv.__i; \
4873    })
4874
4875 #define vld2q_lane_s32(__a, __b, __c) \
4876   ({ \
4877      union { int32x4x2_t __i; __builtin_neon_v4si2 __o; } __bu = { __b }; \
4878      union { int32x4x2_t __i; __builtin_neon_v4si2 __o; } __rv; \
4879      __rv.__o = __builtin_neon_vld2_lanev4si (__a, __bu.__o, __c); \
4880      __rv.__i; \
4881    })
4882
4883 #define vld2q_lane_f32(__a, __b, __c) \
4884   ({ \
4885      union { float32x4x2_t __i; __builtin_neon_v4sf2 __o; } __bu = { __b }; \
4886      union { float32x4x2_t __i; __builtin_neon_v4sf2 __o; } __rv; \
4887      __rv.__o = __builtin_neon_vld2_lanev4sf (__a, __bu.__o, __c); \
4888      __rv.__i; \
4889    })
4890
4891 #define vld2q_lane_u16(__a, __b, __c) \
4892   ({ \
4893      union { uint16x8x2_t __i; __builtin_neon_v8hi2 __o; } __bu = { __b }; \
4894      union { uint16x8x2_t __i; __builtin_neon_v8hi2 __o; } __rv; \
4895      __rv.__o = __builtin_neon_vld2_lanev8hi (__a, __bu.__o, __c); \
4896      __rv.__i; \
4897    })
4898
4899 #define vld2q_lane_u32(__a, __b, __c) \
4900   ({ \
4901      union { uint32x4x2_t __i; __builtin_neon_v4si2 __o; } __bu = { __b }; \
4902      union { uint32x4x2_t __i; __builtin_neon_v4si2 __o; } __rv; \
4903      __rv.__o = __builtin_neon_vld2_lanev4si (__a, __bu.__o, __c); \
4904      __rv.__i; \
4905    })
4906
4907 #define vld2q_lane_p16(__a, __b, __c) \
4908   ({ \
4909      union { poly16x8x2_t __i; __builtin_neon_v8hi2 __o; } __bu = { __b }; \
4910      union { poly16x8x2_t __i; __builtin_neon_v8hi2 __o; } __rv; \
4911      __rv.__o = __builtin_neon_vld2_lanev8hi (__a, __bu.__o, __c); \
4912      __rv.__i; \
4913    })
4914
4915 #define vld2_dup_s8(__a) \
4916   ({ \
4917      union { int8x8x2_t __i; __builtin_neon_v8qi2 __o; } __rv; \
4918      __rv.__o = __builtin_neon_vld2_dupv8qi (__a); \
4919      __rv.__i; \
4920    })
4921
4922 #define vld2_dup_s16(__a) \
4923   ({ \
4924      union { int16x4x2_t __i; __builtin_neon_v4hi2 __o; } __rv; \
4925      __rv.__o = __builtin_neon_vld2_dupv4hi (__a); \
4926      __rv.__i; \
4927    })
4928
4929 #define vld2_dup_s32(__a) \
4930   ({ \
4931      union { int32x2x2_t __i; __builtin_neon_v2si2 __o; } __rv; \
4932      __rv.__o = __builtin_neon_vld2_dupv2si (__a); \
4933      __rv.__i; \
4934    })
4935
4936 #define vld2_dup_f32(__a) \
4937   ({ \
4938      union { float32x2x2_t __i; __builtin_neon_v2sf2 __o; } __rv; \
4939      __rv.__o = __builtin_neon_vld2_dupv2sf (__a); \
4940      __rv.__i; \
4941    })
4942
4943 #define vld2_dup_u8(__a) \
4944   ({ \
4945      union { uint8x8x2_t __i; __builtin_neon_v8qi2 __o; } __rv; \
4946      __rv.__o = __builtin_neon_vld2_dupv8qi (__a); \
4947      __rv.__i; \
4948    })
4949
4950 #define vld2_dup_u16(__a) \
4951   ({ \
4952      union { uint16x4x2_t __i; __builtin_neon_v4hi2 __o; } __rv; \
4953      __rv.__o = __builtin_neon_vld2_dupv4hi (__a); \
4954      __rv.__i; \
4955    })
4956
4957 #define vld2_dup_u32(__a) \
4958   ({ \
4959      union { uint32x2x2_t __i; __builtin_neon_v2si2 __o; } __rv; \
4960      __rv.__o = __builtin_neon_vld2_dupv2si (__a); \
4961      __rv.__i; \
4962    })
4963
4964 #define vld2_dup_p8(__a) \
4965   ({ \
4966      union { poly8x8x2_t __i; __builtin_neon_v8qi2 __o; } __rv; \
4967      __rv.__o = __builtin_neon_vld2_dupv8qi (__a); \
4968      __rv.__i; \
4969    })
4970
4971 #define vld2_dup_p16(__a) \
4972   ({ \
4973      union { poly16x4x2_t __i; __builtin_neon_v4hi2 __o; } __rv; \
4974      __rv.__o = __builtin_neon_vld2_dupv4hi (__a); \
4975      __rv.__i; \
4976    })
4977
4978 #define vld2_dup_s64(__a) \
4979   ({ \
4980      union { int64x1x2_t __i; __builtin_neon_v1di2 __o; } __rv; \
4981      __rv.__o = __builtin_neon_vld2_dupv1di (__a); \
4982      __rv.__i; \
4983    })
4984
4985 #define vld2_dup_u64(__a) \
4986   ({ \
4987      union { uint64x1x2_t __i; __builtin_neon_v1di2 __o; } __rv; \
4988      __rv.__o = __builtin_neon_vld2_dupv1di (__a); \
4989      __rv.__i; \
4990    })
4991
4992 #define vst2_s8(__a, __b) \
4993   ({ \
4994      union { int8x8x2_t __i; __builtin_neon_v8qi2 __o; } __bu = { __b }; \
4995      __builtin_neon_vst2v8qi (__a, __bu.__o); \
4996    })
4997
4998 #define vst2_s16(__a, __b) \
4999   ({ \
5000      union { int16x4x2_t __i; __builtin_neon_v4hi2 __o; } __bu = { __b }; \
5001      __builtin_neon_vst2v4hi (__a, __bu.__o); \
5002    })
5003
5004 #define vst2_s32(__a, __b) \
5005   ({ \
5006      union { int32x2x2_t __i; __builtin_neon_v2si2 __o; } __bu = { __b }; \
5007      __builtin_neon_vst2v2si (__a, __bu.__o); \
5008    })
5009
5010 #define vst2_f32(__a, __b) \
5011   ({ \
5012      union { float32x2x2_t __i; __builtin_neon_v2sf2 __o; } __bu = { __b }; \
5013      __builtin_neon_vst2v2sf (__a, __bu.__o); \
5014    })
5015
5016 #define vst2_u8(__a, __b) \
5017   ({ \
5018      union { uint8x8x2_t __i; __builtin_neon_v8qi2 __o; } __bu = { __b }; \
5019      __builtin_neon_vst2v8qi (__a, __bu.__o); \
5020    })
5021
5022 #define vst2_u16(__a, __b) \
5023   ({ \
5024      union { uint16x4x2_t __i; __builtin_neon_v4hi2 __o; } __bu = { __b }; \
5025      __builtin_neon_vst2v4hi (__a, __bu.__o); \
5026    })
5027
5028 #define vst2_u32(__a, __b) \
5029   ({ \
5030      union { uint32x2x2_t __i; __builtin_neon_v2si2 __o; } __bu = { __b }; \
5031      __builtin_neon_vst2v2si (__a, __bu.__o); \
5032    })
5033
5034 #define vst2_p8(__a, __b) \
5035   ({ \
5036      union { poly8x8x2_t __i; __builtin_neon_v8qi2 __o; } __bu = { __b }; \
5037      __builtin_neon_vst2v8qi (__a, __bu.__o); \
5038    })
5039
5040 #define vst2_p16(__a, __b) \
5041   ({ \
5042      union { poly16x4x2_t __i; __builtin_neon_v4hi2 __o; } __bu = { __b }; \
5043      __builtin_neon_vst2v4hi (__a, __bu.__o); \
5044    })
5045
5046 #define vst2_s64(__a, __b) \
5047   ({ \
5048      union { int64x1x2_t __i; __builtin_neon_v1di2 __o; } __bu = { __b }; \
5049      __builtin_neon_vst2v1di (__a, __bu.__o); \
5050    })
5051
5052 #define vst2_u64(__a, __b) \
5053   ({ \
5054      union { uint64x1x2_t __i; __builtin_neon_v1di2 __o; } __bu = { __b }; \
5055      __builtin_neon_vst2v1di (__a, __bu.__o); \
5056    })
5057
5058 #define vst2q_s8(__a, __b) \
5059   ({ \
5060      union { int8x16x2_t __i; __builtin_neon_v16qi2 __o; } __bu = { __b }; \
5061      __builtin_neon_vst2v16qi (__a, __bu.__o); \
5062    })
5063
5064 #define vst2q_s16(__a, __b) \
5065   ({ \
5066      union { int16x8x2_t __i; __builtin_neon_v8hi2 __o; } __bu = { __b }; \
5067      __builtin_neon_vst2v8hi (__a, __bu.__o); \
5068    })
5069
5070 #define vst2q_s32(__a, __b) \
5071   ({ \
5072      union { int32x4x2_t __i; __builtin_neon_v4si2 __o; } __bu = { __b }; \
5073      __builtin_neon_vst2v4si (__a, __bu.__o); \
5074    })
5075
5076 #define vst2q_f32(__a, __b) \
5077   ({ \
5078      union { float32x4x2_t __i; __builtin_neon_v4sf2 __o; } __bu = { __b }; \
5079      __builtin_neon_vst2v4sf (__a, __bu.__o); \
5080    })
5081
5082 #define vst2q_u8(__a, __b) \
5083   ({ \
5084      union { uint8x16x2_t __i; __builtin_neon_v16qi2 __o; } __bu = { __b }; \
5085      __builtin_neon_vst2v16qi (__a, __bu.__o); \
5086    })
5087
5088 #define vst2q_u16(__a, __b) \
5089   ({ \
5090      union { uint16x8x2_t __i; __builtin_neon_v8hi2 __o; } __bu = { __b }; \
5091      __builtin_neon_vst2v8hi (__a, __bu.__o); \
5092    })
5093
5094 #define vst2q_u32(__a, __b) \
5095   ({ \
5096      union { uint32x4x2_t __i; __builtin_neon_v4si2 __o; } __bu = { __b }; \
5097      __builtin_neon_vst2v4si (__a, __bu.__o); \
5098    })
5099
5100 #define vst2q_p8(__a, __b) \
5101   ({ \
5102      union { poly8x16x2_t __i; __builtin_neon_v16qi2 __o; } __bu = { __b }; \
5103      __builtin_neon_vst2v16qi (__a, __bu.__o); \
5104    })
5105
5106 #define vst2q_p16(__a, __b) \
5107   ({ \
5108      union { poly16x8x2_t __i; __builtin_neon_v8hi2 __o; } __bu = { __b }; \
5109      __builtin_neon_vst2v8hi (__a, __bu.__o); \
5110    })
5111
5112 #define vst2_lane_s8(__a, __b, __c) \
5113   ({ \
5114      union { int8x8x2_t __i; __builtin_neon_v8qi2 __o; } __bu = { __b }; \
5115      __builtin_neon_vst2_lanev8qi (__a, __bu.__o, __c); \
5116    })
5117
5118 #define vst2_lane_s16(__a, __b, __c) \
5119   ({ \
5120      union { int16x4x2_t __i; __builtin_neon_v4hi2 __o; } __bu = { __b }; \
5121      __builtin_neon_vst2_lanev4hi (__a, __bu.__o, __c); \
5122    })
5123
5124 #define vst2_lane_s32(__a, __b, __c) \
5125   ({ \
5126      union { int32x2x2_t __i; __builtin_neon_v2si2 __o; } __bu = { __b }; \
5127      __builtin_neon_vst2_lanev2si (__a, __bu.__o, __c); \
5128    })
5129
5130 #define vst2_lane_f32(__a, __b, __c) \
5131   ({ \
5132      union { float32x2x2_t __i; __builtin_neon_v2sf2 __o; } __bu = { __b }; \
5133      __builtin_neon_vst2_lanev2sf (__a, __bu.__o, __c); \
5134    })
5135
5136 #define vst2_lane_u8(__a, __b, __c) \
5137   ({ \
5138      union { uint8x8x2_t __i; __builtin_neon_v8qi2 __o; } __bu = { __b }; \
5139      __builtin_neon_vst2_lanev8qi (__a, __bu.__o, __c); \
5140    })
5141
5142 #define vst2_lane_u16(__a, __b, __c) \
5143   ({ \
5144      union { uint16x4x2_t __i; __builtin_neon_v4hi2 __o; } __bu = { __b }; \
5145      __builtin_neon_vst2_lanev4hi (__a, __bu.__o, __c); \
5146    })
5147
5148 #define vst2_lane_u32(__a, __b, __c) \
5149   ({ \
5150      union { uint32x2x2_t __i; __builtin_neon_v2si2 __o; } __bu = { __b }; \
5151      __builtin_neon_vst2_lanev2si (__a, __bu.__o, __c); \
5152    })
5153
5154 #define vst2_lane_p8(__a, __b, __c) \
5155   ({ \
5156      union { poly8x8x2_t __i; __builtin_neon_v8qi2 __o; } __bu = { __b }; \
5157      __builtin_neon_vst2_lanev8qi (__a, __bu.__o, __c); \
5158    })
5159
5160 #define vst2_lane_p16(__a, __b, __c) \
5161   ({ \
5162      union { poly16x4x2_t __i; __builtin_neon_v4hi2 __o; } __bu = { __b }; \
5163      __builtin_neon_vst2_lanev4hi (__a, __bu.__o, __c); \
5164    })
5165
5166 #define vst2q_lane_s16(__a, __b, __c) \
5167   ({ \
5168      union { int16x8x2_t __i; __builtin_neon_v8hi2 __o; } __bu = { __b }; \
5169      __builtin_neon_vst2_lanev8hi (__a, __bu.__o, __c); \
5170    })
5171
5172 #define vst2q_lane_s32(__a, __b, __c) \
5173   ({ \
5174      union { int32x4x2_t __i; __builtin_neon_v4si2 __o; } __bu = { __b }; \
5175      __builtin_neon_vst2_lanev4si (__a, __bu.__o, __c); \
5176    })
5177
5178 #define vst2q_lane_f32(__a, __b, __c) \
5179   ({ \
5180      union { float32x4x2_t __i; __builtin_neon_v4sf2 __o; } __bu = { __b }; \
5181      __builtin_neon_vst2_lanev4sf (__a, __bu.__o, __c); \
5182    })
5183
5184 #define vst2q_lane_u16(__a, __b, __c) \
5185   ({ \
5186      union { uint16x8x2_t __i; __builtin_neon_v8hi2 __o; } __bu = { __b }; \
5187      __builtin_neon_vst2_lanev8hi (__a, __bu.__o, __c); \
5188    })
5189
5190 #define vst2q_lane_u32(__a, __b, __c) \
5191   ({ \
5192      union { uint32x4x2_t __i; __builtin_neon_v4si2 __o; } __bu = { __b }; \
5193      __builtin_neon_vst2_lanev4si (__a, __bu.__o, __c); \
5194    })
5195
5196 #define vst2q_lane_p16(__a, __b, __c) \
5197   ({ \
5198      union { poly16x8x2_t __i; __builtin_neon_v8hi2 __o; } __bu = { __b }; \
5199      __builtin_neon_vst2_lanev8hi (__a, __bu.__o, __c); \
5200    })
5201
5202 #define vld3_s8(__a) \
5203   ({ \
5204      union { int8x8x3_t __i; __builtin_neon_v8qi3 __o; } __rv; \
5205      __rv.__o = __builtin_neon_vld3v8qi (__a); \
5206      __rv.__i; \
5207    })
5208
5209 #define vld3_s16(__a) \
5210   ({ \
5211      union { int16x4x3_t __i; __builtin_neon_v4hi3 __o; } __rv; \
5212      __rv.__o = __builtin_neon_vld3v4hi (__a); \
5213      __rv.__i; \
5214    })
5215
5216 #define vld3_s32(__a) \
5217   ({ \
5218      union { int32x2x3_t __i; __builtin_neon_v2si3 __o; } __rv; \
5219      __rv.__o = __builtin_neon_vld3v2si (__a); \
5220      __rv.__i; \
5221    })
5222
5223 #define vld3_f32(__a) \
5224   ({ \
5225      union { float32x2x3_t __i; __builtin_neon_v2sf3 __o; } __rv; \
5226      __rv.__o = __builtin_neon_vld3v2sf (__a); \
5227      __rv.__i; \
5228    })
5229
5230 #define vld3_u8(__a) \
5231   ({ \
5232      union { uint8x8x3_t __i; __builtin_neon_v8qi3 __o; } __rv; \
5233      __rv.__o = __builtin_neon_vld3v8qi (__a); \
5234      __rv.__i; \
5235    })
5236
5237 #define vld3_u16(__a) \
5238   ({ \
5239      union { uint16x4x3_t __i; __builtin_neon_v4hi3 __o; } __rv; \
5240      __rv.__o = __builtin_neon_vld3v4hi (__a); \
5241      __rv.__i; \
5242    })
5243
5244 #define vld3_u32(__a) \
5245   ({ \
5246      union { uint32x2x3_t __i; __builtin_neon_v2si3 __o; } __rv; \
5247      __rv.__o = __builtin_neon_vld3v2si (__a); \
5248      __rv.__i; \
5249    })
5250
5251 #define vld3_p8(__a) \
5252   ({ \
5253      union { poly8x8x3_t __i; __builtin_neon_v8qi3 __o; } __rv; \
5254      __rv.__o = __builtin_neon_vld3v8qi (__a); \
5255      __rv.__i; \
5256    })
5257
5258 #define vld3_p16(__a) \
5259   ({ \
5260      union { poly16x4x3_t __i; __builtin_neon_v4hi3 __o; } __rv; \
5261      __rv.__o = __builtin_neon_vld3v4hi (__a); \
5262      __rv.__i; \
5263    })
5264
5265 #define vld3_s64(__a) \
5266   ({ \
5267      union { int64x1x3_t __i; __builtin_neon_v1di3 __o; } __rv; \
5268      __rv.__o = __builtin_neon_vld3v1di (__a); \
5269      __rv.__i; \
5270    })
5271
5272 #define vld3_u64(__a) \
5273   ({ \
5274      union { uint64x1x3_t __i; __builtin_neon_v1di3 __o; } __rv; \
5275      __rv.__o = __builtin_neon_vld3v1di (__a); \
5276      __rv.__i; \
5277    })
5278
5279 #define vld3q_s8(__a) \
5280   ({ \
5281      union { int8x16x3_t __i; __builtin_neon_v16qi3 __o; } __rv; \
5282      __rv.__o = __builtin_neon_vld3v16qi (__a); \
5283      __rv.__i; \
5284    })
5285
5286 #define vld3q_s16(__a) \
5287   ({ \
5288      union { int16x8x3_t __i; __builtin_neon_v8hi3 __o; } __rv; \
5289      __rv.__o = __builtin_neon_vld3v8hi (__a); \
5290      __rv.__i; \
5291    })
5292
5293 #define vld3q_s32(__a) \
5294   ({ \
5295      union { int32x4x3_t __i; __builtin_neon_v4si3 __o; } __rv; \
5296      __rv.__o = __builtin_neon_vld3v4si (__a); \
5297      __rv.__i; \
5298    })
5299
5300 #define vld3q_f32(__a) \
5301   ({ \
5302      union { float32x4x3_t __i; __builtin_neon_v4sf3 __o; } __rv; \
5303      __rv.__o = __builtin_neon_vld3v4sf (__a); \
5304      __rv.__i; \
5305    })
5306
5307 #define vld3q_u8(__a) \
5308   ({ \
5309      union { uint8x16x3_t __i; __builtin_neon_v16qi3 __o; } __rv; \
5310      __rv.__o = __builtin_neon_vld3v16qi (__a); \
5311      __rv.__i; \
5312    })
5313
5314 #define vld3q_u16(__a) \
5315   ({ \
5316      union { uint16x8x3_t __i; __builtin_neon_v8hi3 __o; } __rv; \
5317      __rv.__o = __builtin_neon_vld3v8hi (__a); \
5318      __rv.__i; \
5319    })
5320
5321 #define vld3q_u32(__a) \
5322   ({ \
5323      union { uint32x4x3_t __i; __builtin_neon_v4si3 __o; } __rv; \
5324      __rv.__o = __builtin_neon_vld3v4si (__a); \
5325      __rv.__i; \
5326    })
5327
5328 #define vld3q_p8(__a) \
5329   ({ \
5330      union { poly8x16x3_t __i; __builtin_neon_v16qi3 __o; } __rv; \
5331      __rv.__o = __builtin_neon_vld3v16qi (__a); \
5332      __rv.__i; \
5333    })
5334
5335 #define vld3q_p16(__a) \
5336   ({ \
5337      union { poly16x8x3_t __i; __builtin_neon_v8hi3 __o; } __rv; \
5338      __rv.__o = __builtin_neon_vld3v8hi (__a); \
5339      __rv.__i; \
5340    })
5341
5342 #define vld3_lane_s8(__a, __b, __c) \
5343   ({ \
5344      union { int8x8x3_t __i; __builtin_neon_v8qi3 __o; } __bu = { __b }; \
5345      union { int8x8x3_t __i; __builtin_neon_v8qi3 __o; } __rv; \
5346      __rv.__o = __builtin_neon_vld3_lanev8qi (__a, __bu.__o, __c); \
5347      __rv.__i; \
5348    })
5349
5350 #define vld3_lane_s16(__a, __b, __c) \
5351   ({ \
5352      union { int16x4x3_t __i; __builtin_neon_v4hi3 __o; } __bu = { __b }; \
5353      union { int16x4x3_t __i; __builtin_neon_v4hi3 __o; } __rv; \
5354      __rv.__o = __builtin_neon_vld3_lanev4hi (__a, __bu.__o, __c); \
5355      __rv.__i; \
5356    })
5357
5358 #define vld3_lane_s32(__a, __b, __c) \
5359   ({ \
5360      union { int32x2x3_t __i; __builtin_neon_v2si3 __o; } __bu = { __b }; \
5361      union { int32x2x3_t __i; __builtin_neon_v2si3 __o; } __rv; \
5362      __rv.__o = __builtin_neon_vld3_lanev2si (__a, __bu.__o, __c); \
5363      __rv.__i; \
5364    })
5365
5366 #define vld3_lane_f32(__a, __b, __c) \
5367   ({ \
5368      union { float32x2x3_t __i; __builtin_neon_v2sf3 __o; } __bu = { __b }; \
5369      union { float32x2x3_t __i; __builtin_neon_v2sf3 __o; } __rv; \
5370      __rv.__o = __builtin_neon_vld3_lanev2sf (__a, __bu.__o, __c); \
5371      __rv.__i; \
5372    })
5373
5374 #define vld3_lane_u8(__a, __b, __c) \
5375   ({ \
5376      union { uint8x8x3_t __i; __builtin_neon_v8qi3 __o; } __bu = { __b }; \
5377      union { uint8x8x3_t __i; __builtin_neon_v8qi3 __o; } __rv; \
5378      __rv.__o = __builtin_neon_vld3_lanev8qi (__a, __bu.__o, __c); \
5379      __rv.__i; \
5380    })
5381
5382 #define vld3_lane_u16(__a, __b, __c) \
5383   ({ \
5384      union { uint16x4x3_t __i; __builtin_neon_v4hi3 __o; } __bu = { __b }; \
5385      union { uint16x4x3_t __i; __builtin_neon_v4hi3 __o; } __rv; \
5386      __rv.__o = __builtin_neon_vld3_lanev4hi (__a, __bu.__o, __c); \
5387      __rv.__i; \
5388    })
5389
5390 #define vld3_lane_u32(__a, __b, __c) \
5391   ({ \
5392      union { uint32x2x3_t __i; __builtin_neon_v2si3 __o; } __bu = { __b }; \
5393      union { uint32x2x3_t __i; __builtin_neon_v2si3 __o; } __rv; \
5394      __rv.__o = __builtin_neon_vld3_lanev2si (__a, __bu.__o, __c); \
5395      __rv.__i; \
5396    })
5397
5398 #define vld3_lane_p8(__a, __b, __c) \
5399   ({ \
5400      union { poly8x8x3_t __i; __builtin_neon_v8qi3 __o; } __bu = { __b }; \
5401      union { poly8x8x3_t __i; __builtin_neon_v8qi3 __o; } __rv; \
5402      __rv.__o = __builtin_neon_vld3_lanev8qi (__a, __bu.__o, __c); \
5403      __rv.__i; \
5404    })
5405
5406 #define vld3_lane_p16(__a, __b, __c) \
5407   ({ \
5408      union { poly16x4x3_t __i; __builtin_neon_v4hi3 __o; } __bu = { __b }; \
5409      union { poly16x4x3_t __i; __builtin_neon_v4hi3 __o; } __rv; \
5410      __rv.__o = __builtin_neon_vld3_lanev4hi (__a, __bu.__o, __c); \
5411      __rv.__i; \
5412    })
5413
5414 #define vld3q_lane_s16(__a, __b, __c) \
5415   ({ \
5416      union { int16x8x3_t __i; __builtin_neon_v8hi3 __o; } __bu = { __b }; \
5417      union { int16x8x3_t __i; __builtin_neon_v8hi3 __o; } __rv; \
5418      __rv.__o = __builtin_neon_vld3_lanev8hi (__a, __bu.__o, __c); \
5419      __rv.__i; \
5420    })
5421
5422 #define vld3q_lane_s32(__a, __b, __c) \
5423   ({ \
5424      union { int32x4x3_t __i; __builtin_neon_v4si3 __o; } __bu = { __b }; \
5425      union { int32x4x3_t __i; __builtin_neon_v4si3 __o; } __rv; \
5426      __rv.__o = __builtin_neon_vld3_lanev4si (__a, __bu.__o, __c); \
5427      __rv.__i; \
5428    })
5429
5430 #define vld3q_lane_f32(__a, __b, __c) \
5431   ({ \
5432      union { float32x4x3_t __i; __builtin_neon_v4sf3 __o; } __bu = { __b }; \
5433      union { float32x4x3_t __i; __builtin_neon_v4sf3 __o; } __rv; \
5434      __rv.__o = __builtin_neon_vld3_lanev4sf (__a, __bu.__o, __c); \
5435      __rv.__i; \
5436    })
5437
5438 #define vld3q_lane_u16(__a, __b, __c) \
5439   ({ \
5440      union { uint16x8x3_t __i; __builtin_neon_v8hi3 __o; } __bu = { __b }; \
5441      union { uint16x8x3_t __i; __builtin_neon_v8hi3 __o; } __rv; \
5442      __rv.__o = __builtin_neon_vld3_lanev8hi (__a, __bu.__o, __c); \
5443      __rv.__i; \
5444    })
5445
5446 #define vld3q_lane_u32(__a, __b, __c) \
5447   ({ \
5448      union { uint32x4x3_t __i; __builtin_neon_v4si3 __o; } __bu = { __b }; \
5449      union { uint32x4x3_t __i; __builtin_neon_v4si3 __o; } __rv; \
5450      __rv.__o = __builtin_neon_vld3_lanev4si (__a, __bu.__o, __c); \
5451      __rv.__i; \
5452    })
5453
5454 #define vld3q_lane_p16(__a, __b, __c) \
5455   ({ \
5456      union { poly16x8x3_t __i; __builtin_neon_v8hi3 __o; } __bu = { __b }; \
5457      union { poly16x8x3_t __i; __builtin_neon_v8hi3 __o; } __rv; \
5458      __rv.__o = __builtin_neon_vld3_lanev8hi (__a, __bu.__o, __c); \
5459      __rv.__i; \
5460    })
5461
5462 #define vld3_dup_s8(__a) \
5463   ({ \
5464      union { int8x8x3_t __i; __builtin_neon_v8qi3 __o; } __rv; \
5465      __rv.__o = __builtin_neon_vld3_dupv8qi (__a); \
5466      __rv.__i; \
5467    })
5468
5469 #define vld3_dup_s16(__a) \
5470   ({ \
5471      union { int16x4x3_t __i; __builtin_neon_v4hi3 __o; } __rv; \
5472      __rv.__o = __builtin_neon_vld3_dupv4hi (__a); \
5473      __rv.__i; \
5474    })
5475
5476 #define vld3_dup_s32(__a) \
5477   ({ \
5478      union { int32x2x3_t __i; __builtin_neon_v2si3 __o; } __rv; \
5479      __rv.__o = __builtin_neon_vld3_dupv2si (__a); \
5480      __rv.__i; \
5481    })
5482
5483 #define vld3_dup_f32(__a) \
5484   ({ \
5485      union { float32x2x3_t __i; __builtin_neon_v2sf3 __o; } __rv; \
5486      __rv.__o = __builtin_neon_vld3_dupv2sf (__a); \
5487      __rv.__i; \
5488    })
5489
5490 #define vld3_dup_u8(__a) \
5491   ({ \
5492      union { uint8x8x3_t __i; __builtin_neon_v8qi3 __o; } __rv; \
5493      __rv.__o = __builtin_neon_vld3_dupv8qi (__a); \
5494      __rv.__i; \
5495    })
5496
5497 #define vld3_dup_u16(__a) \
5498   ({ \
5499      union { uint16x4x3_t __i; __builtin_neon_v4hi3 __o; } __rv; \
5500      __rv.__o = __builtin_neon_vld3_dupv4hi (__a); \
5501      __rv.__i; \
5502    })
5503
5504 #define vld3_dup_u32(__a) \
5505   ({ \
5506      union { uint32x2x3_t __i; __builtin_neon_v2si3 __o; } __rv; \
5507      __rv.__o = __builtin_neon_vld3_dupv2si (__a); \
5508      __rv.__i; \
5509    })
5510
5511 #define vld3_dup_p8(__a) \
5512   ({ \
5513      union { poly8x8x3_t __i; __builtin_neon_v8qi3 __o; } __rv; \
5514      __rv.__o = __builtin_neon_vld3_dupv8qi (__a); \
5515      __rv.__i; \
5516    })
5517
5518 #define vld3_dup_p16(__a) \
5519   ({ \
5520      union { poly16x4x3_t __i; __builtin_neon_v4hi3 __o; } __rv; \
5521      __rv.__o = __builtin_neon_vld3_dupv4hi (__a); \
5522      __rv.__i; \
5523    })
5524
5525 #define vld3_dup_s64(__a) \
5526   ({ \
5527      union { int64x1x3_t __i; __builtin_neon_v1di3 __o; } __rv; \
5528      __rv.__o = __builtin_neon_vld3_dupv1di (__a); \
5529      __rv.__i; \
5530    })
5531
5532 #define vld3_dup_u64(__a) \
5533   ({ \
5534      union { uint64x1x3_t __i; __builtin_neon_v1di3 __o; } __rv; \
5535      __rv.__o = __builtin_neon_vld3_dupv1di (__a); \
5536      __rv.__i; \
5537    })
5538
5539 #define vst3_s8(__a, __b) \
5540   ({ \
5541      union { int8x8x3_t __i; __builtin_neon_v8qi3 __o; } __bu = { __b }; \
5542      __builtin_neon_vst3v8qi (__a, __bu.__o); \
5543    })
5544
5545 #define vst3_s16(__a, __b) \
5546   ({ \
5547      union { int16x4x3_t __i; __builtin_neon_v4hi3 __o; } __bu = { __b }; \
5548      __builtin_neon_vst3v4hi (__a, __bu.__o); \
5549    })
5550
5551 #define vst3_s32(__a, __b) \
5552   ({ \
5553      union { int32x2x3_t __i; __builtin_neon_v2si3 __o; } __bu = { __b }; \
5554      __builtin_neon_vst3v2si (__a, __bu.__o); \
5555    })
5556
5557 #define vst3_f32(__a, __b) \
5558   ({ \
5559      union { float32x2x3_t __i; __builtin_neon_v2sf3 __o; } __bu = { __b }; \
5560      __builtin_neon_vst3v2sf (__a, __bu.__o); \
5561    })
5562
5563 #define vst3_u8(__a, __b) \
5564   ({ \
5565      union { uint8x8x3_t __i; __builtin_neon_v8qi3 __o; } __bu = { __b }; \
5566      __builtin_neon_vst3v8qi (__a, __bu.__o); \
5567    })
5568
5569 #define vst3_u16(__a, __b) \
5570   ({ \
5571      union { uint16x4x3_t __i; __builtin_neon_v4hi3 __o; } __bu = { __b }; \
5572      __builtin_neon_vst3v4hi (__a, __bu.__o); \
5573    })
5574
5575 #define vst3_u32(__a, __b) \
5576   ({ \
5577      union { uint32x2x3_t __i; __builtin_neon_v2si3 __o; } __bu = { __b }; \
5578      __builtin_neon_vst3v2si (__a, __bu.__o); \
5579    })
5580
5581 #define vst3_p8(__a, __b) \
5582   ({ \
5583      union { poly8x8x3_t __i; __builtin_neon_v8qi3 __o; } __bu = { __b }; \
5584      __builtin_neon_vst3v8qi (__a, __bu.__o); \
5585    })
5586
5587 #define vst3_p16(__a, __b) \
5588   ({ \
5589      union { poly16x4x3_t __i; __builtin_neon_v4hi3 __o; } __bu = { __b }; \
5590      __builtin_neon_vst3v4hi (__a, __bu.__o); \
5591    })
5592
5593 #define vst3_s64(__a, __b) \
5594   ({ \
5595      union { int64x1x3_t __i; __builtin_neon_v1di3 __o; } __bu = { __b }; \
5596      __builtin_neon_vst3v1di (__a, __bu.__o); \
5597    })
5598
5599 #define vst3_u64(__a, __b) \
5600   ({ \
5601      union { uint64x1x3_t __i; __builtin_neon_v1di3 __o; } __bu = { __b }; \
5602      __builtin_neon_vst3v1di (__a, __bu.__o); \
5603    })
5604
5605 #define vst3q_s8(__a, __b) \
5606   ({ \
5607      union { int8x16x3_t __i; __builtin_neon_v16qi3 __o; } __bu = { __b }; \
5608      __builtin_neon_vst3v16qi (__a, __bu.__o); \
5609    })
5610
5611 #define vst3q_s16(__a, __b) \
5612   ({ \
5613      union { int16x8x3_t __i; __builtin_neon_v8hi3 __o; } __bu = { __b }; \
5614      __builtin_neon_vst3v8hi (__a, __bu.__o); \
5615    })
5616
5617 #define vst3q_s32(__a, __b) \
5618   ({ \
5619      union { int32x4x3_t __i; __builtin_neon_v4si3 __o; } __bu = { __b }; \
5620      __builtin_neon_vst3v4si (__a, __bu.__o); \
5621    })
5622
5623 #define vst3q_f32(__a, __b) \
5624   ({ \
5625      union { float32x4x3_t __i; __builtin_neon_v4sf3 __o; } __bu = { __b }; \
5626      __builtin_neon_vst3v4sf (__a, __bu.__o); \
5627    })
5628
5629 #define vst3q_u8(__a, __b) \
5630   ({ \
5631      union { uint8x16x3_t __i; __builtin_neon_v16qi3 __o; } __bu = { __b }; \
5632      __builtin_neon_vst3v16qi (__a, __bu.__o); \
5633    })
5634
5635 #define vst3q_u16(__a, __b) \
5636   ({ \
5637      union { uint16x8x3_t __i; __builtin_neon_v8hi3 __o; } __bu = { __b }; \
5638      __builtin_neon_vst3v8hi (__a, __bu.__o); \
5639    })
5640
5641 #define vst3q_u32(__a, __b) \
5642   ({ \
5643      union { uint32x4x3_t __i; __builtin_neon_v4si3 __o; } __bu = { __b }; \
5644      __builtin_neon_vst3v4si (__a, __bu.__o); \
5645    })
5646
5647 #define vst3q_p8(__a, __b) \
5648   ({ \
5649      union { poly8x16x3_t __i; __builtin_neon_v16qi3 __o; } __bu = { __b }; \
5650      __builtin_neon_vst3v16qi (__a, __bu.__o); \
5651    })
5652
5653 #define vst3q_p16(__a, __b) \
5654   ({ \
5655      union { poly16x8x3_t __i; __builtin_neon_v8hi3 __o; } __bu = { __b }; \
5656      __builtin_neon_vst3v8hi (__a, __bu.__o); \
5657    })
5658
5659 #define vst3_lane_s8(__a, __b, __c) \
5660   ({ \
5661      union { int8x8x3_t __i; __builtin_neon_v8qi3 __o; } __bu = { __b }; \
5662      __builtin_neon_vst3_lanev8qi (__a, __bu.__o, __c); \
5663    })
5664
5665 #define vst3_lane_s16(__a, __b, __c) \
5666   ({ \
5667      union { int16x4x3_t __i; __builtin_neon_v4hi3 __o; } __bu = { __b }; \
5668      __builtin_neon_vst3_lanev4hi (__a, __bu.__o, __c); \
5669    })
5670
5671 #define vst3_lane_s32(__a, __b, __c) \
5672   ({ \
5673      union { int32x2x3_t __i; __builtin_neon_v2si3 __o; } __bu = { __b }; \
5674      __builtin_neon_vst3_lanev2si (__a, __bu.__o, __c); \
5675    })
5676
5677 #define vst3_lane_f32(__a, __b, __c) \
5678   ({ \
5679      union { float32x2x3_t __i; __builtin_neon_v2sf3 __o; } __bu = { __b }; \
5680      __builtin_neon_vst3_lanev2sf (__a, __bu.__o, __c); \
5681    })
5682
5683 #define vst3_lane_u8(__a, __b, __c) \
5684   ({ \
5685      union { uint8x8x3_t __i; __builtin_neon_v8qi3 __o; } __bu = { __b }; \
5686      __builtin_neon_vst3_lanev8qi (__a, __bu.__o, __c); \
5687    })
5688
5689 #define vst3_lane_u16(__a, __b, __c) \
5690   ({ \
5691      union { uint16x4x3_t __i; __builtin_neon_v4hi3 __o; } __bu = { __b }; \
5692      __builtin_neon_vst3_lanev4hi (__a, __bu.__o, __c); \
5693    })
5694
5695 #define vst3_lane_u32(__a, __b, __c) \
5696   ({ \
5697      union { uint32x2x3_t __i; __builtin_neon_v2si3 __o; } __bu = { __b }; \
5698      __builtin_neon_vst3_lanev2si (__a, __bu.__o, __c); \
5699    })
5700
5701 #define vst3_lane_p8(__a, __b, __c) \
5702   ({ \
5703      union { poly8x8x3_t __i; __builtin_neon_v8qi3 __o; } __bu = { __b }; \
5704      __builtin_neon_vst3_lanev8qi (__a, __bu.__o, __c); \
5705    })
5706
5707 #define vst3_lane_p16(__a, __b, __c) \
5708   ({ \
5709      union { poly16x4x3_t __i; __builtin_neon_v4hi3 __o; } __bu = { __b }; \
5710      __builtin_neon_vst3_lanev4hi (__a, __bu.__o, __c); \
5711    })
5712
5713 #define vst3q_lane_s16(__a, __b, __c) \
5714   ({ \
5715      union { int16x8x3_t __i; __builtin_neon_v8hi3 __o; } __bu = { __b }; \
5716      __builtin_neon_vst3_lanev8hi (__a, __bu.__o, __c); \
5717    })
5718
5719 #define vst3q_lane_s32(__a, __b, __c) \
5720   ({ \
5721      union { int32x4x3_t __i; __builtin_neon_v4si3 __o; } __bu = { __b }; \
5722      __builtin_neon_vst3_lanev4si (__a, __bu.__o, __c); \
5723    })
5724
5725 #define vst3q_lane_f32(__a, __b, __c) \
5726   ({ \
5727      union { float32x4x3_t __i; __builtin_neon_v4sf3 __o; } __bu = { __b }; \
5728      __builtin_neon_vst3_lanev4sf (__a, __bu.__o, __c); \
5729    })
5730
5731 #define vst3q_lane_u16(__a, __b, __c) \
5732   ({ \
5733      union { uint16x8x3_t __i; __builtin_neon_v8hi3 __o; } __bu = { __b }; \
5734      __builtin_neon_vst3_lanev8hi (__a, __bu.__o, __c); \
5735    })
5736
5737 #define vst3q_lane_u32(__a, __b, __c) \
5738   ({ \
5739      union { uint32x4x3_t __i; __builtin_neon_v4si3 __o; } __bu = { __b }; \
5740      __builtin_neon_vst3_lanev4si (__a, __bu.__o, __c); \
5741    })
5742
5743 #define vst3q_lane_p16(__a, __b, __c) \
5744   ({ \
5745      union { poly16x8x3_t __i; __builtin_neon_v8hi3 __o; } __bu = { __b }; \
5746      __builtin_neon_vst3_lanev8hi (__a, __bu.__o, __c); \
5747    })
5748
5749 #define vld4_s8(__a) \
5750   ({ \
5751      union { int8x8x4_t __i; __builtin_neon_v8qi4 __o; } __rv; \
5752      __rv.__o = __builtin_neon_vld4v8qi (__a); \
5753      __rv.__i; \
5754    })
5755
5756 #define vld4_s16(__a) \
5757   ({ \
5758      union { int16x4x4_t __i; __builtin_neon_v4hi4 __o; } __rv; \
5759      __rv.__o = __builtin_neon_vld4v4hi (__a); \
5760      __rv.__i; \
5761    })
5762
5763 #define vld4_s32(__a) \
5764   ({ \
5765      union { int32x2x4_t __i; __builtin_neon_v2si4 __o; } __rv; \
5766      __rv.__o = __builtin_neon_vld4v2si (__a); \
5767      __rv.__i; \
5768    })
5769
5770 #define vld4_f32(__a) \
5771   ({ \
5772      union { float32x2x4_t __i; __builtin_neon_v2sf4 __o; } __rv; \
5773      __rv.__o = __builtin_neon_vld4v2sf (__a); \
5774      __rv.__i; \
5775    })
5776
5777 #define vld4_u8(__a) \
5778   ({ \
5779      union { uint8x8x4_t __i; __builtin_neon_v8qi4 __o; } __rv; \
5780      __rv.__o = __builtin_neon_vld4v8qi (__a); \
5781      __rv.__i; \
5782    })
5783
5784 #define vld4_u16(__a) \
5785   ({ \
5786      union { uint16x4x4_t __i; __builtin_neon_v4hi4 __o; } __rv; \
5787      __rv.__o = __builtin_neon_vld4v4hi (__a); \
5788      __rv.__i; \
5789    })
5790
5791 #define vld4_u32(__a) \
5792   ({ \
5793      union { uint32x2x4_t __i; __builtin_neon_v2si4 __o; } __rv; \
5794      __rv.__o = __builtin_neon_vld4v2si (__a); \
5795      __rv.__i; \
5796    })
5797
5798 #define vld4_p8(__a) \
5799   ({ \
5800      union { poly8x8x4_t __i; __builtin_neon_v8qi4 __o; } __rv; \
5801      __rv.__o = __builtin_neon_vld4v8qi (__a); \
5802      __rv.__i; \
5803    })
5804
5805 #define vld4_p16(__a) \
5806   ({ \
5807      union { poly16x4x4_t __i; __builtin_neon_v4hi4 __o; } __rv; \
5808      __rv.__o = __builtin_neon_vld4v4hi (__a); \
5809      __rv.__i; \
5810    })
5811
5812 #define vld4_s64(__a) \
5813   ({ \
5814      union { int64x1x4_t __i; __builtin_neon_v1di4 __o; } __rv; \
5815      __rv.__o = __builtin_neon_vld4v1di (__a); \
5816      __rv.__i; \
5817    })
5818
5819 #define vld4_u64(__a) \
5820   ({ \
5821      union { uint64x1x4_t __i; __builtin_neon_v1di4 __o; } __rv; \
5822      __rv.__o = __builtin_neon_vld4v1di (__a); \
5823      __rv.__i; \
5824    })
5825
5826 #define vld4q_s8(__a) \
5827   ({ \
5828      union { int8x16x4_t __i; __builtin_neon_v16qi4 __o; } __rv; \
5829      __rv.__o = __builtin_neon_vld4v16qi (__a); \
5830      __rv.__i; \
5831    })
5832
5833 #define vld4q_s16(__a) \
5834   ({ \
5835      union { int16x8x4_t __i; __builtin_neon_v8hi4 __o; } __rv; \
5836      __rv.__o = __builtin_neon_vld4v8hi (__a); \
5837      __rv.__i; \
5838    })
5839
5840 #define vld4q_s32(__a) \
5841   ({ \
5842      union { int32x4x4_t __i; __builtin_neon_v4si4 __o; } __rv; \
5843      __rv.__o = __builtin_neon_vld4v4si (__a); \
5844      __rv.__i; \
5845    })
5846
5847 #define vld4q_f32(__a) \
5848   ({ \
5849      union { float32x4x4_t __i; __builtin_neon_v4sf4 __o; } __rv; \
5850      __rv.__o = __builtin_neon_vld4v4sf (__a); \
5851      __rv.__i; \
5852    })
5853
5854 #define vld4q_u8(__a) \
5855   ({ \
5856      union { uint8x16x4_t __i; __builtin_neon_v16qi4 __o; } __rv; \
5857      __rv.__o = __builtin_neon_vld4v16qi (__a); \
5858      __rv.__i; \
5859    })
5860
5861 #define vld4q_u16(__a) \
5862   ({ \
5863      union { uint16x8x4_t __i; __builtin_neon_v8hi4 __o; } __rv; \
5864      __rv.__o = __builtin_neon_vld4v8hi (__a); \
5865      __rv.__i; \
5866    })
5867
5868 #define vld4q_u32(__a) \
5869   ({ \
5870      union { uint32x4x4_t __i; __builtin_neon_v4si4 __o; } __rv; \
5871      __rv.__o = __builtin_neon_vld4v4si (__a); \
5872      __rv.__i; \
5873    })
5874
5875 #define vld4q_p8(__a) \
5876   ({ \
5877      union { poly8x16x4_t __i; __builtin_neon_v16qi4 __o; } __rv; \
5878      __rv.__o = __builtin_neon_vld4v16qi (__a); \
5879      __rv.__i; \
5880    })
5881
5882 #define vld4q_p16(__a) \
5883   ({ \
5884      union { poly16x8x4_t __i; __builtin_neon_v8hi4 __o; } __rv; \
5885      __rv.__o = __builtin_neon_vld4v8hi (__a); \
5886      __rv.__i; \
5887    })
5888
5889 #define vld4_lane_s8(__a, __b, __c) \
5890   ({ \
5891      union { int8x8x4_t __i; __builtin_neon_v8qi4 __o; } __bu = { __b }; \
5892      union { int8x8x4_t __i; __builtin_neon_v8qi4 __o; } __rv; \
5893      __rv.__o = __builtin_neon_vld4_lanev8qi (__a, __bu.__o, __c); \
5894      __rv.__i; \
5895    })
5896
5897 #define vld4_lane_s16(__a, __b, __c) \
5898   ({ \
5899      union { int16x4x4_t __i; __builtin_neon_v4hi4 __o; } __bu = { __b }; \
5900      union { int16x4x4_t __i; __builtin_neon_v4hi4 __o; } __rv; \
5901      __rv.__o = __builtin_neon_vld4_lanev4hi (__a, __bu.__o, __c); \
5902      __rv.__i; \
5903    })
5904
5905 #define vld4_lane_s32(__a, __b, __c) \
5906   ({ \
5907      union { int32x2x4_t __i; __builtin_neon_v2si4 __o; } __bu = { __b }; \
5908      union { int32x2x4_t __i; __builtin_neon_v2si4 __o; } __rv; \
5909      __rv.__o = __builtin_neon_vld4_lanev2si (__a, __bu.__o, __c); \
5910      __rv.__i; \
5911    })
5912
5913 #define vld4_lane_f32(__a, __b, __c) \
5914   ({ \
5915      union { float32x2x4_t __i; __builtin_neon_v2sf4 __o; } __bu = { __b }; \
5916      union { float32x2x4_t __i; __builtin_neon_v2sf4 __o; } __rv; \
5917      __rv.__o = __builtin_neon_vld4_lanev2sf (__a, __bu.__o, __c); \
5918      __rv.__i; \
5919    })
5920
5921 #define vld4_lane_u8(__a, __b, __c) \
5922   ({ \
5923      union { uint8x8x4_t __i; __builtin_neon_v8qi4 __o; } __bu = { __b }; \
5924      union { uint8x8x4_t __i; __builtin_neon_v8qi4 __o; } __rv; \
5925      __rv.__o = __builtin_neon_vld4_lanev8qi (__a, __bu.__o, __c); \
5926      __rv.__i; \
5927    })
5928
5929 #define vld4_lane_u16(__a, __b, __c) \
5930   ({ \
5931      union { uint16x4x4_t __i; __builtin_neon_v4hi4 __o; } __bu = { __b }; \
5932      union { uint16x4x4_t __i; __builtin_neon_v4hi4 __o; } __rv; \
5933      __rv.__o = __builtin_neon_vld4_lanev4hi (__a, __bu.__o, __c); \
5934      __rv.__i; \
5935    })
5936
5937 #define vld4_lane_u32(__a, __b, __c) \
5938   ({ \
5939      union { uint32x2x4_t __i; __builtin_neon_v2si4 __o; } __bu = { __b }; \
5940      union { uint32x2x4_t __i; __builtin_neon_v2si4 __o; } __rv; \
5941      __rv.__o = __builtin_neon_vld4_lanev2si (__a, __bu.__o, __c); \
5942      __rv.__i; \
5943    })
5944
5945 #define vld4_lane_p8(__a, __b, __c) \
5946   ({ \
5947      union { poly8x8x4_t __i; __builtin_neon_v8qi4 __o; } __bu = { __b }; \
5948      union { poly8x8x4_t __i; __builtin_neon_v8qi4 __o; } __rv; \
5949      __rv.__o = __builtin_neon_vld4_lanev8qi (__a, __bu.__o, __c); \
5950      __rv.__i; \
5951    })
5952
5953 #define vld4_lane_p16(__a, __b, __c) \
5954   ({ \
5955      union { poly16x4x4_t __i; __builtin_neon_v4hi4 __o; } __bu = { __b }; \
5956      union { poly16x4x4_t __i; __builtin_neon_v4hi4 __o; } __rv; \
5957      __rv.__o = __builtin_neon_vld4_lanev4hi (__a, __bu.__o, __c); \
5958      __rv.__i; \
5959    })
5960
5961 #define vld4q_lane_s16(__a, __b, __c) \
5962   ({ \
5963      union { int16x8x4_t __i; __builtin_neon_v8hi4 __o; } __bu = { __b }; \
5964      union { int16x8x4_t __i; __builtin_neon_v8hi4 __o; } __rv; \
5965      __rv.__o = __builtin_neon_vld4_lanev8hi (__a, __bu.__o, __c); \
5966      __rv.__i; \
5967    })
5968
5969 #define vld4q_lane_s32(__a, __b, __c) \
5970   ({ \
5971      union { int32x4x4_t __i; __builtin_neon_v4si4 __o; } __bu = { __b }; \
5972      union { int32x4x4_t __i; __builtin_neon_v4si4 __o; } __rv; \
5973      __rv.__o = __builtin_neon_vld4_lanev4si (__a, __bu.__o, __c); \
5974      __rv.__i; \
5975    })
5976
5977 #define vld4q_lane_f32(__a, __b, __c) \
5978   ({ \
5979      union { float32x4x4_t __i; __builtin_neon_v4sf4 __o; } __bu = { __b }; \
5980      union { float32x4x4_t __i; __builtin_neon_v4sf4 __o; } __rv; \
5981      __rv.__o = __builtin_neon_vld4_lanev4sf (__a, __bu.__o, __c); \
5982      __rv.__i; \
5983    })
5984
5985 #define vld4q_lane_u16(__a, __b, __c) \
5986   ({ \
5987      union { uint16x8x4_t __i; __builtin_neon_v8hi4 __o; } __bu = { __b }; \
5988      union { uint16x8x4_t __i; __builtin_neon_v8hi4 __o; } __rv; \
5989      __rv.__o = __builtin_neon_vld4_lanev8hi (__a, __bu.__o, __c); \
5990      __rv.__i; \
5991    })
5992
5993 #define vld4q_lane_u32(__a, __b, __c) \
5994   ({ \
5995      union { uint32x4x4_t __i; __builtin_neon_v4si4 __o; } __bu = { __b }; \
5996      union { uint32x4x4_t __i; __builtin_neon_v4si4 __o; } __rv; \
5997      __rv.__o = __builtin_neon_vld4_lanev4si (__a, __bu.__o, __c); \
5998      __rv.__i; \
5999    })
6000
6001 #define vld4q_lane_p16(__a, __b, __c) \
6002   ({ \
6003      union { poly16x8x4_t __i; __builtin_neon_v8hi4 __o; } __bu = { __b }; \
6004      union { poly16x8x4_t __i; __builtin_neon_v8hi4 __o; } __rv; \
6005      __rv.__o = __builtin_neon_vld4_lanev8hi (__a, __bu.__o, __c); \
6006      __rv.__i; \
6007    })
6008
6009 #define vld4_dup_s8(__a) \
6010   ({ \
6011      union { int8x8x4_t __i; __builtin_neon_v8qi4 __o; } __rv; \
6012      __rv.__o = __builtin_neon_vld4_dupv8qi (__a); \
6013      __rv.__i; \
6014    })
6015
6016 #define vld4_dup_s16(__a) \
6017   ({ \
6018      union { int16x4x4_t __i; __builtin_neon_v4hi4 __o; } __rv; \
6019      __rv.__o = __builtin_neon_vld4_dupv4hi (__a); \
6020      __rv.__i; \
6021    })
6022
6023 #define vld4_dup_s32(__a) \
6024   ({ \
6025      union { int32x2x4_t __i; __builtin_neon_v2si4 __o; } __rv; \
6026      __rv.__o = __builtin_neon_vld4_dupv2si (__a); \
6027      __rv.__i; \
6028    })
6029
6030 #define vld4_dup_f32(__a) \
6031   ({ \
6032      union { float32x2x4_t __i; __builtin_neon_v2sf4 __o; } __rv; \
6033      __rv.__o = __builtin_neon_vld4_dupv2sf (__a); \
6034      __rv.__i; \
6035    })
6036
6037 #define vld4_dup_u8(__a) \
6038   ({ \
6039      union { uint8x8x4_t __i; __builtin_neon_v8qi4 __o; } __rv; \
6040      __rv.__o = __builtin_neon_vld4_dupv8qi (__a); \
6041      __rv.__i; \
6042    })
6043
6044 #define vld4_dup_u16(__a) \
6045   ({ \
6046      union { uint16x4x4_t __i; __builtin_neon_v4hi4 __o; } __rv; \
6047      __rv.__o = __builtin_neon_vld4_dupv4hi (__a); \
6048      __rv.__i; \
6049    })
6050
6051 #define vld4_dup_u32(__a) \
6052   ({ \
6053      union { uint32x2x4_t __i; __builtin_neon_v2si4 __o; } __rv; \
6054      __rv.__o = __builtin_neon_vld4_dupv2si (__a); \
6055      __rv.__i; \
6056    })
6057
6058 #define vld4_dup_p8(__a) \
6059   ({ \
6060      union { poly8x8x4_t __i; __builtin_neon_v8qi4 __o; } __rv; \
6061      __rv.__o = __builtin_neon_vld4_dupv8qi (__a); \
6062      __rv.__i; \
6063    })
6064
6065 #define vld4_dup_p16(__a) \
6066   ({ \
6067      union { poly16x4x4_t __i; __builtin_neon_v4hi4 __o; } __rv; \
6068      __rv.__o = __builtin_neon_vld4_dupv4hi (__a); \
6069      __rv.__i; \
6070    })
6071
6072 #define vld4_dup_s64(__a) \
6073   ({ \
6074      union { int64x1x4_t __i; __builtin_neon_v1di4 __o; } __rv; \
6075      __rv.__o = __builtin_neon_vld4_dupv1di (__a); \
6076      __rv.__i; \
6077    })
6078
6079 #define vld4_dup_u64(__a) \
6080   ({ \
6081      union { uint64x1x4_t __i; __builtin_neon_v1di4 __o; } __rv; \
6082      __rv.__o = __builtin_neon_vld4_dupv1di (__a); \
6083      __rv.__i; \
6084    })
6085
6086 #define vst4_s8(__a, __b) \
6087   ({ \
6088      union { int8x8x4_t __i; __builtin_neon_v8qi4 __o; } __bu = { __b }; \
6089      __builtin_neon_vst4v8qi (__a, __bu.__o); \
6090    })
6091
6092 #define vst4_s16(__a, __b) \
6093   ({ \
6094      union { int16x4x4_t __i; __builtin_neon_v4hi4 __o; } __bu = { __b }; \
6095      __builtin_neon_vst4v4hi (__a, __bu.__o); \
6096    })
6097
6098 #define vst4_s32(__a, __b) \
6099   ({ \
6100      union { int32x2x4_t __i; __builtin_neon_v2si4 __o; } __bu = { __b }; \
6101      __builtin_neon_vst4v2si (__a, __bu.__o); \
6102    })
6103
6104 #define vst4_f32(__a, __b) \
6105   ({ \
6106      union { float32x2x4_t __i; __builtin_neon_v2sf4 __o; } __bu = { __b }; \
6107      __builtin_neon_vst4v2sf (__a, __bu.__o); \
6108    })
6109
6110 #define vst4_u8(__a, __b) \
6111   ({ \
6112      union { uint8x8x4_t __i; __builtin_neon_v8qi4 __o; } __bu = { __b }; \
6113      __builtin_neon_vst4v8qi (__a, __bu.__o); \
6114    })
6115
6116 #define vst4_u16(__a, __b) \
6117   ({ \
6118      union { uint16x4x4_t __i; __builtin_neon_v4hi4 __o; } __bu = { __b }; \
6119      __builtin_neon_vst4v4hi (__a, __bu.__o); \
6120    })
6121
6122 #define vst4_u32(__a, __b) \
6123   ({ \
6124      union { uint32x2x4_t __i; __builtin_neon_v2si4 __o; } __bu = { __b }; \
6125      __builtin_neon_vst4v2si (__a, __bu.__o); \
6126    })
6127
6128 #define vst4_p8(__a, __b) \
6129   ({ \
6130      union { poly8x8x4_t __i; __builtin_neon_v8qi4 __o; } __bu = { __b }; \
6131      __builtin_neon_vst4v8qi (__a, __bu.__o); \
6132    })
6133
6134 #define vst4_p16(__a, __b) \
6135   ({ \
6136      union { poly16x4x4_t __i; __builtin_neon_v4hi4 __o; } __bu = { __b }; \
6137      __builtin_neon_vst4v4hi (__a, __bu.__o); \
6138    })
6139
6140 #define vst4_s64(__a, __b) \
6141   ({ \
6142      union { int64x1x4_t __i; __builtin_neon_v1di4 __o; } __bu = { __b }; \
6143      __builtin_neon_vst4v1di (__a, __bu.__o); \
6144    })
6145
6146 #define vst4_u64(__a, __b) \
6147   ({ \
6148      union { uint64x1x4_t __i; __builtin_neon_v1di4 __o; } __bu = { __b }; \
6149      __builtin_neon_vst4v1di (__a, __bu.__o); \
6150    })
6151
6152 #define vst4q_s8(__a, __b) \
6153   ({ \
6154      union { int8x16x4_t __i; __builtin_neon_v16qi4 __o; } __bu = { __b }; \
6155      __builtin_neon_vst4v16qi (__a, __bu.__o); \
6156    })
6157
6158 #define vst4q_s16(__a, __b) \
6159   ({ \
6160      union { int16x8x4_t __i; __builtin_neon_v8hi4 __o; } __bu = { __b }; \
6161      __builtin_neon_vst4v8hi (__a, __bu.__o); \
6162    })
6163
6164 #define vst4q_s32(__a, __b) \
6165   ({ \
6166      union { int32x4x4_t __i; __builtin_neon_v4si4 __o; } __bu = { __b }; \
6167      __builtin_neon_vst4v4si (__a, __bu.__o); \
6168    })
6169
6170 #define vst4q_f32(__a, __b) \
6171   ({ \
6172      union { float32x4x4_t __i; __builtin_neon_v4sf4 __o; } __bu = { __b }; \
6173      __builtin_neon_vst4v4sf (__a, __bu.__o); \
6174    })
6175
6176 #define vst4q_u8(__a, __b) \
6177   ({ \
6178      union { uint8x16x4_t __i; __builtin_neon_v16qi4 __o; } __bu = { __b }; \
6179      __builtin_neon_vst4v16qi (__a, __bu.__o); \
6180    })
6181
6182 #define vst4q_u16(__a, __b) \
6183   ({ \
6184      union { uint16x8x4_t __i; __builtin_neon_v8hi4 __o; } __bu = { __b }; \
6185      __builtin_neon_vst4v8hi (__a, __bu.__o); \
6186    })
6187
6188 #define vst4q_u32(__a, __b) \
6189   ({ \
6190      union { uint32x4x4_t __i; __builtin_neon_v4si4 __o; } __bu = { __b }; \
6191      __builtin_neon_vst4v4si (__a, __bu.__o); \
6192    })
6193
6194 #define vst4q_p8(__a, __b) \
6195   ({ \
6196      union { poly8x16x4_t __i; __builtin_neon_v16qi4 __o; } __bu = { __b }; \
6197      __builtin_neon_vst4v16qi (__a, __bu.__o); \
6198    })
6199
6200 #define vst4q_p16(__a, __b) \
6201   ({ \
6202      union { poly16x8x4_t __i; __builtin_neon_v8hi4 __o; } __bu = { __b }; \
6203      __builtin_neon_vst4v8hi (__a, __bu.__o); \
6204    })
6205
6206 #define vst4_lane_s8(__a, __b, __c) \
6207   ({ \
6208      union { int8x8x4_t __i; __builtin_neon_v8qi4 __o; } __bu = { __b }; \
6209      __builtin_neon_vst4_lanev8qi (__a, __bu.__o, __c); \
6210    })
6211
6212 #define vst4_lane_s16(__a, __b, __c) \
6213   ({ \
6214      union { int16x4x4_t __i; __builtin_neon_v4hi4 __o; } __bu = { __b }; \
6215      __builtin_neon_vst4_lanev4hi (__a, __bu.__o, __c); \
6216    })
6217
6218 #define vst4_lane_s32(__a, __b, __c) \
6219   ({ \
6220      union { int32x2x4_t __i; __builtin_neon_v2si4 __o; } __bu = { __b }; \
6221      __builtin_neon_vst4_lanev2si (__a, __bu.__o, __c); \
6222    })
6223
6224 #define vst4_lane_f32(__a, __b, __c) \
6225   ({ \
6226      union { float32x2x4_t __i; __builtin_neon_v2sf4 __o; } __bu = { __b }; \
6227      __builtin_neon_vst4_lanev2sf (__a, __bu.__o, __c); \
6228    })
6229
6230 #define vst4_lane_u8(__a, __b, __c) \
6231   ({ \
6232      union { uint8x8x4_t __i; __builtin_neon_v8qi4 __o; } __bu = { __b }; \
6233      __builtin_neon_vst4_lanev8qi (__a, __bu.__o, __c); \
6234    })
6235
6236 #define vst4_lane_u16(__a, __b, __c) \
6237   ({ \
6238      union { uint16x4x4_t __i; __builtin_neon_v4hi4 __o; } __bu = { __b }; \
6239      __builtin_neon_vst4_lanev4hi (__a, __bu.__o, __c); \
6240    })
6241
6242 #define vst4_lane_u32(__a, __b, __c) \
6243   ({ \
6244      union { uint32x2x4_t __i; __builtin_neon_v2si4 __o; } __bu = { __b }; \
6245      __builtin_neon_vst4_lanev2si (__a, __bu.__o, __c); \
6246    })
6247
6248 #define vst4_lane_p8(__a, __b, __c) \
6249   ({ \
6250      union { poly8x8x4_t __i; __builtin_neon_v8qi4 __o; } __bu = { __b }; \
6251      __builtin_neon_vst4_lanev8qi (__a, __bu.__o, __c); \
6252    })
6253
6254 #define vst4_lane_p16(__a, __b, __c) \
6255   ({ \
6256      union { poly16x4x4_t __i; __builtin_neon_v4hi4 __o; } __bu = { __b }; \
6257      __builtin_neon_vst4_lanev4hi (__a, __bu.__o, __c); \
6258    })
6259
6260 #define vst4q_lane_s16(__a, __b, __c) \
6261   ({ \
6262      union { int16x8x4_t __i; __builtin_neon_v8hi4 __o; } __bu = { __b }; \
6263      __builtin_neon_vst4_lanev8hi (__a, __bu.__o, __c); \
6264    })
6265
6266 #define vst4q_lane_s32(__a, __b, __c) \
6267   ({ \
6268      union { int32x4x4_t __i; __builtin_neon_v4si4 __o; } __bu = { __b }; \
6269      __builtin_neon_vst4_lanev4si (__a, __bu.__o, __c); \
6270    })
6271
6272 #define vst4q_lane_f32(__a, __b, __c) \
6273   ({ \
6274      union { float32x4x4_t __i; __builtin_neon_v4sf4 __o; } __bu = { __b }; \
6275      __builtin_neon_vst4_lanev4sf (__a, __bu.__o, __c); \
6276    })
6277
6278 #define vst4q_lane_u16(__a, __b, __c) \
6279   ({ \
6280      union { uint16x8x4_t __i; __builtin_neon_v8hi4 __o; } __bu = { __b }; \
6281      __builtin_neon_vst4_lanev8hi (__a, __bu.__o, __c); \
6282    })
6283
6284 #define vst4q_lane_u32(__a, __b, __c) \
6285   ({ \
6286      union { uint32x4x4_t __i; __builtin_neon_v4si4 __o; } __bu = { __b }; \
6287      __builtin_neon_vst4_lanev4si (__a, __bu.__o, __c); \
6288    })
6289
6290 #define vst4q_lane_p16(__a, __b, __c) \
6291   ({ \
6292      union { poly16x8x4_t __i; __builtin_neon_v8hi4 __o; } __bu = { __b }; \
6293      __builtin_neon_vst4_lanev8hi (__a, __bu.__o, __c); \
6294    })
6295
6296 #define vand_s8(__a, __b) \
6297   (int8x8_t)__builtin_neon_vandv8qi (__a, __b, 1)
6298
6299 #define vand_s16(__a, __b) \
6300   (int16x4_t)__builtin_neon_vandv4hi (__a, __b, 1)
6301
6302 #define vand_s32(__a, __b) \
6303   (int32x2_t)__builtin_neon_vandv2si (__a, __b, 1)
6304
6305 #define vand_s64(__a, __b) \
6306   (int64x1_t)__builtin_neon_vandv1di (__a, __b, 1)
6307
6308 #define vand_u8(__a, __b) \
6309   (uint8x8_t)__builtin_neon_vandv8qi (__a, __b, 0)
6310
6311 #define vand_u16(__a, __b) \
6312   (uint16x4_t)__builtin_neon_vandv4hi (__a, __b, 0)
6313
6314 #define vand_u32(__a, __b) \
6315   (uint32x2_t)__builtin_neon_vandv2si (__a, __b, 0)
6316
6317 #define vand_u64(__a, __b) \
6318   (uint64x1_t)__builtin_neon_vandv1di (__a, __b, 0)
6319
6320 #define vandq_s8(__a, __b) \
6321   (int8x16_t)__builtin_neon_vandv16qi (__a, __b, 1)
6322
6323 #define vandq_s16(__a, __b) \
6324   (int16x8_t)__builtin_neon_vandv8hi (__a, __b, 1)
6325
6326 #define vandq_s32(__a, __b) \
6327   (int32x4_t)__builtin_neon_vandv4si (__a, __b, 1)
6328
6329 #define vandq_s64(__a, __b) \
6330   (int64x2_t)__builtin_neon_vandv2di (__a, __b, 1)
6331
6332 #define vandq_u8(__a, __b) \
6333   (uint8x16_t)__builtin_neon_vandv16qi (__a, __b, 0)
6334
6335 #define vandq_u16(__a, __b) \
6336   (uint16x8_t)__builtin_neon_vandv8hi (__a, __b, 0)
6337
6338 #define vandq_u32(__a, __b) \
6339   (uint32x4_t)__builtin_neon_vandv4si (__a, __b, 0)
6340
6341 #define vandq_u64(__a, __b) \
6342   (uint64x2_t)__builtin_neon_vandv2di (__a, __b, 0)
6343
6344 #define vorr_s8(__a, __b) \
6345   (int8x8_t)__builtin_neon_vorrv8qi (__a, __b, 1)
6346
6347 #define vorr_s16(__a, __b) \
6348   (int16x4_t)__builtin_neon_vorrv4hi (__a, __b, 1)
6349
6350 #define vorr_s32(__a, __b) \
6351   (int32x2_t)__builtin_neon_vorrv2si (__a, __b, 1)
6352
6353 #define vorr_s64(__a, __b) \
6354   (int64x1_t)__builtin_neon_vorrv1di (__a, __b, 1)
6355
6356 #define vorr_u8(__a, __b) \
6357   (uint8x8_t)__builtin_neon_vorrv8qi (__a, __b, 0)
6358
6359 #define vorr_u16(__a, __b) \
6360   (uint16x4_t)__builtin_neon_vorrv4hi (__a, __b, 0)
6361
6362 #define vorr_u32(__a, __b) \
6363   (uint32x2_t)__builtin_neon_vorrv2si (__a, __b, 0)
6364
6365 #define vorr_u64(__a, __b) \
6366   (uint64x1_t)__builtin_neon_vorrv1di (__a, __b, 0)
6367
6368 #define vorrq_s8(__a, __b) \
6369   (int8x16_t)__builtin_neon_vorrv16qi (__a, __b, 1)
6370
6371 #define vorrq_s16(__a, __b) \
6372   (int16x8_t)__builtin_neon_vorrv8hi (__a, __b, 1)
6373
6374 #define vorrq_s32(__a, __b) \
6375   (int32x4_t)__builtin_neon_vorrv4si (__a, __b, 1)
6376
6377 #define vorrq_s64(__a, __b) \
6378   (int64x2_t)__builtin_neon_vorrv2di (__a, __b, 1)
6379
6380 #define vorrq_u8(__a, __b) \
6381   (uint8x16_t)__builtin_neon_vorrv16qi (__a, __b, 0)
6382
6383 #define vorrq_u16(__a, __b) \
6384   (uint16x8_t)__builtin_neon_vorrv8hi (__a, __b, 0)
6385
6386 #define vorrq_u32(__a, __b) \
6387   (uint32x4_t)__builtin_neon_vorrv4si (__a, __b, 0)
6388
6389 #define vorrq_u64(__a, __b) \
6390   (uint64x2_t)__builtin_neon_vorrv2di (__a, __b, 0)
6391
6392 #define veor_s8(__a, __b) \
6393   (int8x8_t)__builtin_neon_veorv8qi (__a, __b, 1)
6394
6395 #define veor_s16(__a, __b) \
6396   (int16x4_t)__builtin_neon_veorv4hi (__a, __b, 1)
6397
6398 #define veor_s32(__a, __b) \
6399   (int32x2_t)__builtin_neon_veorv2si (__a, __b, 1)
6400
6401 #define veor_s64(__a, __b) \
6402   (int64x1_t)__builtin_neon_veorv1di (__a, __b, 1)
6403
6404 #define veor_u8(__a, __b) \
6405   (uint8x8_t)__builtin_neon_veorv8qi (__a, __b, 0)
6406
6407 #define veor_u16(__a, __b) \
6408   (uint16x4_t)__builtin_neon_veorv4hi (__a, __b, 0)
6409
6410 #define veor_u32(__a, __b) \
6411   (uint32x2_t)__builtin_neon_veorv2si (__a, __b, 0)
6412
6413 #define veor_u64(__a, __b) \
6414   (uint64x1_t)__builtin_neon_veorv1di (__a, __b, 0)
6415
6416 #define veorq_s8(__a, __b) \
6417   (int8x16_t)__builtin_neon_veorv16qi (__a, __b, 1)
6418
6419 #define veorq_s16(__a, __b) \
6420   (int16x8_t)__builtin_neon_veorv8hi (__a, __b, 1)
6421
6422 #define veorq_s32(__a, __b) \
6423   (int32x4_t)__builtin_neon_veorv4si (__a, __b, 1)
6424
6425 #define veorq_s64(__a, __b) \
6426   (int64x2_t)__builtin_neon_veorv2di (__a, __b, 1)
6427
6428 #define veorq_u8(__a, __b) \
6429   (uint8x16_t)__builtin_neon_veorv16qi (__a, __b, 0)
6430
6431 #define veorq_u16(__a, __b) \
6432   (uint16x8_t)__builtin_neon_veorv8hi (__a, __b, 0)
6433
6434 #define veorq_u32(__a, __b) \
6435   (uint32x4_t)__builtin_neon_veorv4si (__a, __b, 0)
6436
6437 #define veorq_u64(__a, __b) \
6438   (uint64x2_t)__builtin_neon_veorv2di (__a, __b, 0)
6439
6440 #define vbic_s8(__a, __b) \
6441   (int8x8_t)__builtin_neon_vbicv8qi (__a, __b, 1)
6442
6443 #define vbic_s16(__a, __b) \
6444   (int16x4_t)__builtin_neon_vbicv4hi (__a, __b, 1)
6445
6446 #define vbic_s32(__a, __b) \
6447   (int32x2_t)__builtin_neon_vbicv2si (__a, __b, 1)
6448
6449 #define vbic_s64(__a, __b) \
6450   (int64x1_t)__builtin_neon_vbicv1di (__a, __b, 1)
6451
6452 #define vbic_u8(__a, __b) \
6453   (uint8x8_t)__builtin_neon_vbicv8qi (__a, __b, 0)
6454
6455 #define vbic_u16(__a, __b) \
6456   (uint16x4_t)__builtin_neon_vbicv4hi (__a, __b, 0)
6457
6458 #define vbic_u32(__a, __b) \
6459   (uint32x2_t)__builtin_neon_vbicv2si (__a, __b, 0)
6460
6461 #define vbic_u64(__a, __b) \
6462   (uint64x1_t)__builtin_neon_vbicv1di (__a, __b, 0)
6463
6464 #define vbicq_s8(__a, __b) \
6465   (int8x16_t)__builtin_neon_vbicv16qi (__a, __b, 1)
6466
6467 #define vbicq_s16(__a, __b) \
6468   (int16x8_t)__builtin_neon_vbicv8hi (__a, __b, 1)
6469
6470 #define vbicq_s32(__a, __b) \
6471   (int32x4_t)__builtin_neon_vbicv4si (__a, __b, 1)
6472
6473 #define vbicq_s64(__a, __b) \
6474   (int64x2_t)__builtin_neon_vbicv2di (__a, __b, 1)
6475
6476 #define vbicq_u8(__a, __b) \
6477   (uint8x16_t)__builtin_neon_vbicv16qi (__a, __b, 0)
6478
6479 #define vbicq_u16(__a, __b) \
6480   (uint16x8_t)__builtin_neon_vbicv8hi (__a, __b, 0)
6481
6482 #define vbicq_u32(__a, __b) \
6483   (uint32x4_t)__builtin_neon_vbicv4si (__a, __b, 0)
6484
6485 #define vbicq_u64(__a, __b) \
6486   (uint64x2_t)__builtin_neon_vbicv2di (__a, __b, 0)
6487
6488 #define vorn_s8(__a, __b) \
6489   (int8x8_t)__builtin_neon_vornv8qi (__a, __b, 1)
6490
6491 #define vorn_s16(__a, __b) \
6492   (int16x4_t)__builtin_neon_vornv4hi (__a, __b, 1)
6493
6494 #define vorn_s32(__a, __b) \
6495   (int32x2_t)__builtin_neon_vornv2si (__a, __b, 1)
6496
6497 #define vorn_s64(__a, __b) \
6498   (int64x1_t)__builtin_neon_vornv1di (__a, __b, 1)
6499
6500 #define vorn_u8(__a, __b) \
6501   (uint8x8_t)__builtin_neon_vornv8qi (__a, __b, 0)
6502
6503 #define vorn_u16(__a, __b) \
6504   (uint16x4_t)__builtin_neon_vornv4hi (__a, __b, 0)
6505
6506 #define vorn_u32(__a, __b) \
6507   (uint32x2_t)__builtin_neon_vornv2si (__a, __b, 0)
6508
6509 #define vorn_u64(__a, __b) \
6510   (uint64x1_t)__builtin_neon_vornv1di (__a, __b, 0)
6511
6512 #define vornq_s8(__a, __b) \
6513   (int8x16_t)__builtin_neon_vornv16qi (__a, __b, 1)
6514
6515 #define vornq_s16(__a, __b) \
6516   (int16x8_t)__builtin_neon_vornv8hi (__a, __b, 1)
6517
6518 #define vornq_s32(__a, __b) \
6519   (int32x4_t)__builtin_neon_vornv4si (__a, __b, 1)
6520
6521 #define vornq_s64(__a, __b) \
6522   (int64x2_t)__builtin_neon_vornv2di (__a, __b, 1)
6523
6524 #define vornq_u8(__a, __b) \
6525   (uint8x16_t)__builtin_neon_vornv16qi (__a, __b, 0)
6526
6527 #define vornq_u16(__a, __b) \
6528   (uint16x8_t)__builtin_neon_vornv8hi (__a, __b, 0)
6529
6530 #define vornq_u32(__a, __b) \
6531   (uint32x4_t)__builtin_neon_vornv4si (__a, __b, 0)
6532
6533 #define vornq_u64(__a, __b) \
6534   (uint64x2_t)__builtin_neon_vornv2di (__a, __b, 0)
6535
6536
6537 #define vreinterpret_p8_s8(__a) \
6538   (poly8x8_t)__builtin_neon_vreinterpretv8qiv8qi (__a)
6539
6540 #define vreinterpret_p8_s16(__a) \
6541   (poly8x8_t)__builtin_neon_vreinterpretv8qiv4hi (__a)
6542
6543 #define vreinterpret_p8_s32(__a) \
6544   (poly8x8_t)__builtin_neon_vreinterpretv8qiv2si (__a)
6545
6546 #define vreinterpret_p8_s64(__a) \
6547   (poly8x8_t)__builtin_neon_vreinterpretv8qiv1di (__a)
6548
6549 #define vreinterpret_p8_f32(__a) \
6550   (poly8x8_t)__builtin_neon_vreinterpretv8qiv2sf (__a)
6551
6552 #define vreinterpret_p8_u8(__a) \
6553   (poly8x8_t)__builtin_neon_vreinterpretv8qiv8qi (__a)
6554
6555 #define vreinterpret_p8_u16(__a) \
6556   (poly8x8_t)__builtin_neon_vreinterpretv8qiv4hi (__a)
6557
6558 #define vreinterpret_p8_u32(__a) \
6559   (poly8x8_t)__builtin_neon_vreinterpretv8qiv2si (__a)
6560
6561 #define vreinterpret_p8_u64(__a) \
6562   (poly8x8_t)__builtin_neon_vreinterpretv8qiv1di (__a)
6563
6564 #define vreinterpret_p8_p16(__a) \
6565   (poly8x8_t)__builtin_neon_vreinterpretv8qiv4hi (__a)
6566
6567 #define vreinterpretq_p8_s8(__a) \
6568   (poly8x16_t)__builtin_neon_vreinterpretv16qiv16qi (__a)
6569
6570 #define vreinterpretq_p8_s16(__a) \
6571   (poly8x16_t)__builtin_neon_vreinterpretv16qiv8hi (__a)
6572
6573 #define vreinterpretq_p8_s32(__a) \
6574   (poly8x16_t)__builtin_neon_vreinterpretv16qiv4si (__a)
6575
6576 #define vreinterpretq_p8_s64(__a) \
6577   (poly8x16_t)__builtin_neon_vreinterpretv16qiv2di (__a)
6578
6579 #define vreinterpretq_p8_f32(__a) \
6580   (poly8x16_t)__builtin_neon_vreinterpretv16qiv4sf (__a)
6581
6582 #define vreinterpretq_p8_u8(__a) \
6583   (poly8x16_t)__builtin_neon_vreinterpretv16qiv16qi (__a)
6584
6585 #define vreinterpretq_p8_u16(__a) \
6586   (poly8x16_t)__builtin_neon_vreinterpretv16qiv8hi (__a)
6587
6588 #define vreinterpretq_p8_u32(__a) \
6589   (poly8x16_t)__builtin_neon_vreinterpretv16qiv4si (__a)
6590
6591 #define vreinterpretq_p8_u64(__a) \
6592   (poly8x16_t)__builtin_neon_vreinterpretv16qiv2di (__a)
6593
6594 #define vreinterpretq_p8_p16(__a) \
6595   (poly8x16_t)__builtin_neon_vreinterpretv16qiv8hi (__a)
6596
6597 #define vreinterpret_p16_s8(__a) \
6598   (poly16x4_t)__builtin_neon_vreinterpretv4hiv8qi (__a)
6599
6600 #define vreinterpret_p16_s16(__a) \
6601   (poly16x4_t)__builtin_neon_vreinterpretv4hiv4hi (__a)
6602
6603 #define vreinterpret_p16_s32(__a) \
6604   (poly16x4_t)__builtin_neon_vreinterpretv4hiv2si (__a)
6605
6606 #define vreinterpret_p16_s64(__a) \
6607   (poly16x4_t)__builtin_neon_vreinterpretv4hiv1di (__a)
6608
6609 #define vreinterpret_p16_f32(__a) \
6610   (poly16x4_t)__builtin_neon_vreinterpretv4hiv2sf (__a)
6611
6612 #define vreinterpret_p16_u8(__a) \
6613   (poly16x4_t)__builtin_neon_vreinterpretv4hiv8qi (__a)
6614
6615 #define vreinterpret_p16_u16(__a) \
6616   (poly16x4_t)__builtin_neon_vreinterpretv4hiv4hi (__a)
6617
6618 #define vreinterpret_p16_u32(__a) \
6619   (poly16x4_t)__builtin_neon_vreinterpretv4hiv2si (__a)
6620
6621 #define vreinterpret_p16_u64(__a) \
6622   (poly16x4_t)__builtin_neon_vreinterpretv4hiv1di (__a)
6623
6624 #define vreinterpret_p16_p8(__a) \
6625   (poly16x4_t)__builtin_neon_vreinterpretv4hiv8qi (__a)
6626
6627 #define vreinterpretq_p16_s8(__a) \
6628   (poly16x8_t)__builtin_neon_vreinterpretv8hiv16qi (__a)
6629
6630 #define vreinterpretq_p16_s16(__a) \
6631   (poly16x8_t)__builtin_neon_vreinterpretv8hiv8hi (__a)
6632
6633 #define vreinterpretq_p16_s32(__a) \
6634   (poly16x8_t)__builtin_neon_vreinterpretv8hiv4si (__a)
6635
6636 #define vreinterpretq_p16_s64(__a) \
6637   (poly16x8_t)__builtin_neon_vreinterpretv8hiv2di (__a)
6638
6639 #define vreinterpretq_p16_f32(__a) \
6640   (poly16x8_t)__builtin_neon_vreinterpretv8hiv4sf (__a)
6641
6642 #define vreinterpretq_p16_u8(__a) \
6643   (poly16x8_t)__builtin_neon_vreinterpretv8hiv16qi (__a)
6644
6645 #define vreinterpretq_p16_u16(__a) \
6646   (poly16x8_t)__builtin_neon_vreinterpretv8hiv8hi (__a)
6647
6648 #define vreinterpretq_p16_u32(__a) \
6649   (poly16x8_t)__builtin_neon_vreinterpretv8hiv4si (__a)
6650
6651 #define vreinterpretq_p16_u64(__a) \
6652   (poly16x8_t)__builtin_neon_vreinterpretv8hiv2di (__a)
6653
6654 #define vreinterpretq_p16_p8(__a) \
6655   (poly16x8_t)__builtin_neon_vreinterpretv8hiv16qi (__a)
6656
6657 #define vreinterpret_f32_s8(__a) \
6658   (float32x2_t)__builtin_neon_vreinterpretv2sfv8qi (__a)
6659
6660 #define vreinterpret_f32_s16(__a) \
6661   (float32x2_t)__builtin_neon_vreinterpretv2sfv4hi (__a)
6662
6663 #define vreinterpret_f32_s32(__a) \
6664   (float32x2_t)__builtin_neon_vreinterpretv2sfv2si (__a)
6665
6666 #define vreinterpret_f32_s64(__a) \
6667   (float32x2_t)__builtin_neon_vreinterpretv2sfv1di (__a)
6668
6669 #define vreinterpret_f32_u8(__a) \
6670   (float32x2_t)__builtin_neon_vreinterpretv2sfv8qi (__a)
6671
6672 #define vreinterpret_f32_u16(__a) \
6673   (float32x2_t)__builtin_neon_vreinterpretv2sfv4hi (__a)
6674
6675 #define vreinterpret_f32_u32(__a) \
6676   (float32x2_t)__builtin_neon_vreinterpretv2sfv2si (__a)
6677
6678 #define vreinterpret_f32_u64(__a) \
6679   (float32x2_t)__builtin_neon_vreinterpretv2sfv1di (__a)
6680
6681 #define vreinterpret_f32_p8(__a) \
6682   (float32x2_t)__builtin_neon_vreinterpretv2sfv8qi (__a)
6683
6684 #define vreinterpret_f32_p16(__a) \
6685   (float32x2_t)__builtin_neon_vreinterpretv2sfv4hi (__a)
6686
6687 #define vreinterpretq_f32_s8(__a) \
6688   (float32x4_t)__builtin_neon_vreinterpretv4sfv16qi (__a)
6689
6690 #define vreinterpretq_f32_s16(__a) \
6691   (float32x4_t)__builtin_neon_vreinterpretv4sfv8hi (__a)
6692
6693 #define vreinterpretq_f32_s32(__a) \
6694   (float32x4_t)__builtin_neon_vreinterpretv4sfv4si (__a)
6695
6696 #define vreinterpretq_f32_s64(__a) \
6697   (float32x4_t)__builtin_neon_vreinterpretv4sfv2di (__a)
6698
6699 #define vreinterpretq_f32_u8(__a) \
6700   (float32x4_t)__builtin_neon_vreinterpretv4sfv16qi (__a)
6701
6702 #define vreinterpretq_f32_u16(__a) \
6703   (float32x4_t)__builtin_neon_vreinterpretv4sfv8hi (__a)
6704
6705 #define vreinterpretq_f32_u32(__a) \
6706   (float32x4_t)__builtin_neon_vreinterpretv4sfv4si (__a)
6707
6708 #define vreinterpretq_f32_u64(__a) \
6709   (float32x4_t)__builtin_neon_vreinterpretv4sfv2di (__a)
6710
6711 #define vreinterpretq_f32_p8(__a) \
6712   (float32x4_t)__builtin_neon_vreinterpretv4sfv16qi (__a)
6713
6714 #define vreinterpretq_f32_p16(__a) \
6715   (float32x4_t)__builtin_neon_vreinterpretv4sfv8hi (__a)
6716
6717 #define vreinterpret_s64_s8(__a) \
6718   (int64x1_t)__builtin_neon_vreinterpretv1div8qi (__a)
6719
6720 #define vreinterpret_s64_s16(__a) \
6721   (int64x1_t)__builtin_neon_vreinterpretv1div4hi (__a)
6722
6723 #define vreinterpret_s64_s32(__a) \
6724   (int64x1_t)__builtin_neon_vreinterpretv1div2si (__a)
6725
6726 #define vreinterpret_s64_f32(__a) \
6727   (int64x1_t)__builtin_neon_vreinterpretv1div2sf (__a)
6728
6729 #define vreinterpret_s64_u8(__a) \
6730   (int64x1_t)__builtin_neon_vreinterpretv1div8qi (__a)
6731
6732 #define vreinterpret_s64_u16(__a) \
6733   (int64x1_t)__builtin_neon_vreinterpretv1div4hi (__a)
6734
6735 #define vreinterpret_s64_u32(__a) \
6736   (int64x1_t)__builtin_neon_vreinterpretv1div2si (__a)
6737
6738 #define vreinterpret_s64_u64(__a) \
6739   (int64x1_t)__builtin_neon_vreinterpretv1div1di (__a)
6740
6741 #define vreinterpret_s64_p8(__a) \
6742   (int64x1_t)__builtin_neon_vreinterpretv1div8qi (__a)
6743
6744 #define vreinterpret_s64_p16(__a) \
6745   (int64x1_t)__builtin_neon_vreinterpretv1div4hi (__a)
6746
6747 #define vreinterpretq_s64_s8(__a) \
6748   (int64x2_t)__builtin_neon_vreinterpretv2div16qi (__a)
6749
6750 #define vreinterpretq_s64_s16(__a) \
6751   (int64x2_t)__builtin_neon_vreinterpretv2div8hi (__a)
6752
6753 #define vreinterpretq_s64_s32(__a) \
6754   (int64x2_t)__builtin_neon_vreinterpretv2div4si (__a)
6755
6756 #define vreinterpretq_s64_f32(__a) \
6757   (int64x2_t)__builtin_neon_vreinterpretv2div4sf (__a)
6758
6759 #define vreinterpretq_s64_u8(__a) \
6760   (int64x2_t)__builtin_neon_vreinterpretv2div16qi (__a)
6761
6762 #define vreinterpretq_s64_u16(__a) \
6763   (int64x2_t)__builtin_neon_vreinterpretv2div8hi (__a)
6764
6765 #define vreinterpretq_s64_u32(__a) \
6766   (int64x2_t)__builtin_neon_vreinterpretv2div4si (__a)
6767
6768 #define vreinterpretq_s64_u64(__a) \
6769   (int64x2_t)__builtin_neon_vreinterpretv2div2di (__a)
6770
6771 #define vreinterpretq_s64_p8(__a) \
6772   (int64x2_t)__builtin_neon_vreinterpretv2div16qi (__a)
6773
6774 #define vreinterpretq_s64_p16(__a) \
6775   (int64x2_t)__builtin_neon_vreinterpretv2div8hi (__a)
6776
6777 #define vreinterpret_u64_s8(__a) \
6778   (uint64x1_t)__builtin_neon_vreinterpretv1div8qi (__a)
6779
6780 #define vreinterpret_u64_s16(__a) \
6781   (uint64x1_t)__builtin_neon_vreinterpretv1div4hi (__a)
6782
6783 #define vreinterpret_u64_s32(__a) \
6784   (uint64x1_t)__builtin_neon_vreinterpretv1div2si (__a)
6785
6786 #define vreinterpret_u64_s64(__a) \
6787   (uint64x1_t)__builtin_neon_vreinterpretv1div1di (__a)
6788
6789 #define vreinterpret_u64_f32(__a) \
6790   (uint64x1_t)__builtin_neon_vreinterpretv1div2sf (__a)
6791
6792 #define vreinterpret_u64_u8(__a) \
6793   (uint64x1_t)__builtin_neon_vreinterpretv1div8qi (__a)
6794
6795 #define vreinterpret_u64_u16(__a) \
6796   (uint64x1_t)__builtin_neon_vreinterpretv1div4hi (__a)
6797
6798 #define vreinterpret_u64_u32(__a) \
6799   (uint64x1_t)__builtin_neon_vreinterpretv1div2si (__a)
6800
6801 #define vreinterpret_u64_p8(__a) \
6802   (uint64x1_t)__builtin_neon_vreinterpretv1div8qi (__a)
6803
6804 #define vreinterpret_u64_p16(__a) \
6805   (uint64x1_t)__builtin_neon_vreinterpretv1div4hi (__a)
6806
6807 #define vreinterpretq_u64_s8(__a) \
6808   (uint64x2_t)__builtin_neon_vreinterpretv2div16qi (__a)
6809
6810 #define vreinterpretq_u64_s16(__a) \
6811   (uint64x2_t)__builtin_neon_vreinterpretv2div8hi (__a)
6812
6813 #define vreinterpretq_u64_s32(__a) \
6814   (uint64x2_t)__builtin_neon_vreinterpretv2div4si (__a)
6815
6816 #define vreinterpretq_u64_s64(__a) \
6817   (uint64x2_t)__builtin_neon_vreinterpretv2div2di (__a)
6818
6819 #define vreinterpretq_u64_f32(__a) \
6820   (uint64x2_t)__builtin_neon_vreinterpretv2div4sf (__a)
6821
6822 #define vreinterpretq_u64_u8(__a) \
6823   (uint64x2_t)__builtin_neon_vreinterpretv2div16qi (__a)
6824
6825 #define vreinterpretq_u64_u16(__a) \
6826   (uint64x2_t)__builtin_neon_vreinterpretv2div8hi (__a)
6827
6828 #define vreinterpretq_u64_u32(__a) \
6829   (uint64x2_t)__builtin_neon_vreinterpretv2div4si (__a)
6830
6831 #define vreinterpretq_u64_p8(__a) \
6832   (uint64x2_t)__builtin_neon_vreinterpretv2div16qi (__a)
6833
6834 #define vreinterpretq_u64_p16(__a) \
6835   (uint64x2_t)__builtin_neon_vreinterpretv2div8hi (__a)
6836
6837 #define vreinterpret_s8_s16(__a) \
6838   (int8x8_t)__builtin_neon_vreinterpretv8qiv4hi (__a)
6839
6840 #define vreinterpret_s8_s32(__a) \
6841   (int8x8_t)__builtin_neon_vreinterpretv8qiv2si (__a)
6842
6843 #define vreinterpret_s8_s64(__a) \
6844   (int8x8_t)__builtin_neon_vreinterpretv8qiv1di (__a)
6845
6846 #define vreinterpret_s8_f32(__a) \
6847   (int8x8_t)__builtin_neon_vreinterpretv8qiv2sf (__a)
6848
6849 #define vreinterpret_s8_u8(__a) \
6850   (int8x8_t)__builtin_neon_vreinterpretv8qiv8qi (__a)
6851
6852 #define vreinterpret_s8_u16(__a) \
6853   (int8x8_t)__builtin_neon_vreinterpretv8qiv4hi (__a)
6854
6855 #define vreinterpret_s8_u32(__a) \
6856   (int8x8_t)__builtin_neon_vreinterpretv8qiv2si (__a)
6857
6858 #define vreinterpret_s8_u64(__a) \
6859   (int8x8_t)__builtin_neon_vreinterpretv8qiv1di (__a)
6860
6861 #define vreinterpret_s8_p8(__a) \
6862   (int8x8_t)__builtin_neon_vreinterpretv8qiv8qi (__a)
6863
6864 #define vreinterpret_s8_p16(__a) \
6865   (int8x8_t)__builtin_neon_vreinterpretv8qiv4hi (__a)
6866
6867 #define vreinterpretq_s8_s16(__a) \
6868   (int8x16_t)__builtin_neon_vreinterpretv16qiv8hi (__a)
6869
6870 #define vreinterpretq_s8_s32(__a) \
6871   (int8x16_t)__builtin_neon_vreinterpretv16qiv4si (__a)
6872
6873 #define vreinterpretq_s8_s64(__a) \
6874   (int8x16_t)__builtin_neon_vreinterpretv16qiv2di (__a)
6875
6876 #define vreinterpretq_s8_f32(__a) \
6877   (int8x16_t)__builtin_neon_vreinterpretv16qiv4sf (__a)
6878
6879 #define vreinterpretq_s8_u8(__a) \
6880   (int8x16_t)__builtin_neon_vreinterpretv16qiv16qi (__a)
6881
6882 #define vreinterpretq_s8_u16(__a) \
6883   (int8x16_t)__builtin_neon_vreinterpretv16qiv8hi (__a)
6884
6885 #define vreinterpretq_s8_u32(__a) \
6886   (int8x16_t)__builtin_neon_vreinterpretv16qiv4si (__a)
6887
6888 #define vreinterpretq_s8_u64(__a) \
6889   (int8x16_t)__builtin_neon_vreinterpretv16qiv2di (__a)
6890
6891 #define vreinterpretq_s8_p8(__a) \
6892   (int8x16_t)__builtin_neon_vreinterpretv16qiv16qi (__a)
6893
6894 #define vreinterpretq_s8_p16(__a) \
6895   (int8x16_t)__builtin_neon_vreinterpretv16qiv8hi (__a)
6896
6897 #define vreinterpret_s16_s8(__a) \
6898   (int16x4_t)__builtin_neon_vreinterpretv4hiv8qi (__a)
6899
6900 #define vreinterpret_s16_s32(__a) \
6901   (int16x4_t)__builtin_neon_vreinterpretv4hiv2si (__a)
6902
6903 #define vreinterpret_s16_s64(__a) \
6904   (int16x4_t)__builtin_neon_vreinterpretv4hiv1di (__a)
6905
6906 #define vreinterpret_s16_f32(__a) \
6907   (int16x4_t)__builtin_neon_vreinterpretv4hiv2sf (__a)
6908
6909 #define vreinterpret_s16_u8(__a) \
6910   (int16x4_t)__builtin_neon_vreinterpretv4hiv8qi (__a)
6911
6912 #define vreinterpret_s16_u16(__a) \
6913   (int16x4_t)__builtin_neon_vreinterpretv4hiv4hi (__a)
6914
6915 #define vreinterpret_s16_u32(__a) \
6916   (int16x4_t)__builtin_neon_vreinterpretv4hiv2si (__a)
6917
6918 #define vreinterpret_s16_u64(__a) \
6919   (int16x4_t)__builtin_neon_vreinterpretv4hiv1di (__a)
6920
6921 #define vreinterpret_s16_p8(__a) \
6922   (int16x4_t)__builtin_neon_vreinterpretv4hiv8qi (__a)
6923
6924 #define vreinterpret_s16_p16(__a) \
6925   (int16x4_t)__builtin_neon_vreinterpretv4hiv4hi (__a)
6926
6927 #define vreinterpretq_s16_s8(__a) \
6928   (int16x8_t)__builtin_neon_vreinterpretv8hiv16qi (__a)
6929
6930 #define vreinterpretq_s16_s32(__a) \
6931   (int16x8_t)__builtin_neon_vreinterpretv8hiv4si (__a)
6932
6933 #define vreinterpretq_s16_s64(__a) \
6934   (int16x8_t)__builtin_neon_vreinterpretv8hiv2di (__a)
6935
6936 #define vreinterpretq_s16_f32(__a) \
6937   (int16x8_t)__builtin_neon_vreinterpretv8hiv4sf (__a)
6938
6939 #define vreinterpretq_s16_u8(__a) \
6940   (int16x8_t)__builtin_neon_vreinterpretv8hiv16qi (__a)
6941
6942 #define vreinterpretq_s16_u16(__a) \
6943   (int16x8_t)__builtin_neon_vreinterpretv8hiv8hi (__a)
6944
6945 #define vreinterpretq_s16_u32(__a) \
6946   (int16x8_t)__builtin_neon_vreinterpretv8hiv4si (__a)
6947
6948 #define vreinterpretq_s16_u64(__a) \
6949   (int16x8_t)__builtin_neon_vreinterpretv8hiv2di (__a)
6950
6951 #define vreinterpretq_s16_p8(__a) \
6952   (int16x8_t)__builtin_neon_vreinterpretv8hiv16qi (__a)
6953
6954 #define vreinterpretq_s16_p16(__a) \
6955   (int16x8_t)__builtin_neon_vreinterpretv8hiv8hi (__a)
6956
6957 #define vreinterpret_s32_s8(__a) \
6958   (int32x2_t)__builtin_neon_vreinterpretv2siv8qi (__a)
6959
6960 #define vreinterpret_s32_s16(__a) \
6961   (int32x2_t)__builtin_neon_vreinterpretv2siv4hi (__a)
6962
6963 #define vreinterpret_s32_s64(__a) \
6964   (int32x2_t)__builtin_neon_vreinterpretv2siv1di (__a)
6965
6966 #define vreinterpret_s32_f32(__a) \
6967   (int32x2_t)__builtin_neon_vreinterpretv2siv2sf (__a)
6968
6969 #define vreinterpret_s32_u8(__a) \
6970   (int32x2_t)__builtin_neon_vreinterpretv2siv8qi (__a)
6971
6972 #define vreinterpret_s32_u16(__a) \
6973   (int32x2_t)__builtin_neon_vreinterpretv2siv4hi (__a)
6974
6975 #define vreinterpret_s32_u32(__a) \
6976   (int32x2_t)__builtin_neon_vreinterpretv2siv2si (__a)
6977
6978 #define vreinterpret_s32_u64(__a) \
6979   (int32x2_t)__builtin_neon_vreinterpretv2siv1di (__a)
6980
6981 #define vreinterpret_s32_p8(__a) \
6982   (int32x2_t)__builtin_neon_vreinterpretv2siv8qi (__a)
6983
6984 #define vreinterpret_s32_p16(__a) \
6985   (int32x2_t)__builtin_neon_vreinterpretv2siv4hi (__a)
6986
6987 #define vreinterpretq_s32_s8(__a) \
6988   (int32x4_t)__builtin_neon_vreinterpretv4siv16qi (__a)
6989
6990 #define vreinterpretq_s32_s16(__a) \
6991   (int32x4_t)__builtin_neon_vreinterpretv4siv8hi (__a)
6992
6993 #define vreinterpretq_s32_s64(__a) \
6994   (int32x4_t)__builtin_neon_vreinterpretv4siv2di (__a)
6995
6996 #define vreinterpretq_s32_f32(__a) \
6997   (int32x4_t)__builtin_neon_vreinterpretv4siv4sf (__a)
6998
6999 #define vreinterpretq_s32_u8(__a) \
7000   (int32x4_t)__builtin_neon_vreinterpretv4siv16qi (__a)
7001
7002 #define vreinterpretq_s32_u16(__a) \
7003   (int32x4_t)__builtin_neon_vreinterpretv4siv8hi (__a)
7004
7005 #define vreinterpretq_s32_u32(__a) \
7006   (int32x4_t)__builtin_neon_vreinterpretv4siv4si (__a)
7007
7008 #define vreinterpretq_s32_u64(__a) \
7009   (int32x4_t)__builtin_neon_vreinterpretv4siv2di (__a)
7010
7011 #define vreinterpretq_s32_p8(__a) \
7012   (int32x4_t)__builtin_neon_vreinterpretv4siv16qi (__a)
7013
7014 #define vreinterpretq_s32_p16(__a) \
7015   (int32x4_t)__builtin_neon_vreinterpretv4siv8hi (__a)
7016
7017 #define vreinterpret_u8_s8(__a) \
7018   (uint8x8_t)__builtin_neon_vreinterpretv8qiv8qi (__a)
7019
7020 #define vreinterpret_u8_s16(__a) \
7021   (uint8x8_t)__builtin_neon_vreinterpretv8qiv4hi (__a)
7022
7023 #define vreinterpret_u8_s32(__a) \
7024   (uint8x8_t)__builtin_neon_vreinterpretv8qiv2si (__a)
7025
7026 #define vreinterpret_u8_s64(__a) \
7027   (uint8x8_t)__builtin_neon_vreinterpretv8qiv1di (__a)
7028
7029 #define vreinterpret_u8_f32(__a) \
7030   (uint8x8_t)__builtin_neon_vreinterpretv8qiv2sf (__a)
7031
7032 #define vreinterpret_u8_u16(__a) \
7033   (uint8x8_t)__builtin_neon_vreinterpretv8qiv4hi (__a)
7034
7035 #define vreinterpret_u8_u32(__a) \
7036   (uint8x8_t)__builtin_neon_vreinterpretv8qiv2si (__a)
7037
7038 #define vreinterpret_u8_u64(__a) \
7039   (uint8x8_t)__builtin_neon_vreinterpretv8qiv1di (__a)
7040
7041 #define vreinterpret_u8_p8(__a) \
7042   (uint8x8_t)__builtin_neon_vreinterpretv8qiv8qi (__a)
7043
7044 #define vreinterpret_u8_p16(__a) \
7045   (uint8x8_t)__builtin_neon_vreinterpretv8qiv4hi (__a)
7046
7047 #define vreinterpretq_u8_s8(__a) \
7048   (uint8x16_t)__builtin_neon_vreinterpretv16qiv16qi (__a)
7049
7050 #define vreinterpretq_u8_s16(__a) \
7051   (uint8x16_t)__builtin_neon_vreinterpretv16qiv8hi (__a)
7052
7053 #define vreinterpretq_u8_s32(__a) \
7054   (uint8x16_t)__builtin_neon_vreinterpretv16qiv4si (__a)
7055
7056 #define vreinterpretq_u8_s64(__a) \
7057   (uint8x16_t)__builtin_neon_vreinterpretv16qiv2di (__a)
7058
7059 #define vreinterpretq_u8_f32(__a) \
7060   (uint8x16_t)__builtin_neon_vreinterpretv16qiv4sf (__a)
7061
7062 #define vreinterpretq_u8_u16(__a) \
7063   (uint8x16_t)__builtin_neon_vreinterpretv16qiv8hi (__a)
7064
7065 #define vreinterpretq_u8_u32(__a) \
7066   (uint8x16_t)__builtin_neon_vreinterpretv16qiv4si (__a)
7067
7068 #define vreinterpretq_u8_u64(__a) \
7069   (uint8x16_t)__builtin_neon_vreinterpretv16qiv2di (__a)
7070
7071 #define vreinterpretq_u8_p8(__a) \
7072   (uint8x16_t)__builtin_neon_vreinterpretv16qiv16qi (__a)
7073
7074 #define vreinterpretq_u8_p16(__a) \
7075   (uint8x16_t)__builtin_neon_vreinterpretv16qiv8hi (__a)
7076
7077 #define vreinterpret_u16_s8(__a) \
7078   (uint16x4_t)__builtin_neon_vreinterpretv4hiv8qi (__a)
7079
7080 #define vreinterpret_u16_s16(__a) \
7081   (uint16x4_t)__builtin_neon_vreinterpretv4hiv4hi (__a)
7082
7083 #define vreinterpret_u16_s32(__a) \
7084   (uint16x4_t)__builtin_neon_vreinterpretv4hiv2si (__a)
7085
7086 #define vreinterpret_u16_s64(__a) \
7087   (uint16x4_t)__builtin_neon_vreinterpretv4hiv1di (__a)
7088
7089 #define vreinterpret_u16_f32(__a) \
7090   (uint16x4_t)__builtin_neon_vreinterpretv4hiv2sf (__a)
7091
7092 #define vreinterpret_u16_u8(__a) \
7093   (uint16x4_t)__builtin_neon_vreinterpretv4hiv8qi (__a)
7094
7095 #define vreinterpret_u16_u32(__a) \
7096   (uint16x4_t)__builtin_neon_vreinterpretv4hiv2si (__a)
7097
7098 #define vreinterpret_u16_u64(__a) \
7099   (uint16x4_t)__builtin_neon_vreinterpretv4hiv1di (__a)
7100
7101 #define vreinterpret_u16_p8(__a) \
7102   (uint16x4_t)__builtin_neon_vreinterpretv4hiv8qi (__a)
7103
7104 #define vreinterpret_u16_p16(__a) \
7105   (uint16x4_t)__builtin_neon_vreinterpretv4hiv4hi (__a)
7106
7107 #define vreinterpretq_u16_s8(__a) \
7108   (uint16x8_t)__builtin_neon_vreinterpretv8hiv16qi (__a)
7109
7110 #define vreinterpretq_u16_s16(__a) \
7111   (uint16x8_t)__builtin_neon_vreinterpretv8hiv8hi (__a)
7112
7113 #define vreinterpretq_u16_s32(__a) \
7114   (uint16x8_t)__builtin_neon_vreinterpretv8hiv4si (__a)
7115
7116 #define vreinterpretq_u16_s64(__a) \
7117   (uint16x8_t)__builtin_neon_vreinterpretv8hiv2di (__a)
7118
7119 #define vreinterpretq_u16_f32(__a) \
7120   (uint16x8_t)__builtin_neon_vreinterpretv8hiv4sf (__a)
7121
7122 #define vreinterpretq_u16_u8(__a) \
7123   (uint16x8_t)__builtin_neon_vreinterpretv8hiv16qi (__a)
7124
7125 #define vreinterpretq_u16_u32(__a) \
7126   (uint16x8_t)__builtin_neon_vreinterpretv8hiv4si (__a)
7127
7128 #define vreinterpretq_u16_u64(__a) \
7129   (uint16x8_t)__builtin_neon_vreinterpretv8hiv2di (__a)
7130
7131 #define vreinterpretq_u16_p8(__a) \
7132   (uint16x8_t)__builtin_neon_vreinterpretv8hiv16qi (__a)
7133
7134 #define vreinterpretq_u16_p16(__a) \
7135   (uint16x8_t)__builtin_neon_vreinterpretv8hiv8hi (__a)
7136
7137 #define vreinterpret_u32_s8(__a) \
7138   (uint32x2_t)__builtin_neon_vreinterpretv2siv8qi (__a)
7139
7140 #define vreinterpret_u32_s16(__a) \
7141   (uint32x2_t)__builtin_neon_vreinterpretv2siv4hi (__a)
7142
7143 #define vreinterpret_u32_s32(__a) \
7144   (uint32x2_t)__builtin_neon_vreinterpretv2siv2si (__a)
7145
7146 #define vreinterpret_u32_s64(__a) \
7147   (uint32x2_t)__builtin_neon_vreinterpretv2siv1di (__a)
7148
7149 #define vreinterpret_u32_f32(__a) \
7150   (uint32x2_t)__builtin_neon_vreinterpretv2siv2sf (__a)
7151
7152 #define vreinterpret_u32_u8(__a) \
7153   (uint32x2_t)__builtin_neon_vreinterpretv2siv8qi (__a)
7154
7155 #define vreinterpret_u32_u16(__a) \
7156   (uint32x2_t)__builtin_neon_vreinterpretv2siv4hi (__a)
7157
7158 #define vreinterpret_u32_u64(__a) \
7159   (uint32x2_t)__builtin_neon_vreinterpretv2siv1di (__a)
7160
7161 #define vreinterpret_u32_p8(__a) \
7162   (uint32x2_t)__builtin_neon_vreinterpretv2siv8qi (__a)
7163
7164 #define vreinterpret_u32_p16(__a) \
7165   (uint32x2_t)__builtin_neon_vreinterpretv2siv4hi (__a)
7166
7167 #define vreinterpretq_u32_s8(__a) \
7168   (uint32x4_t)__builtin_neon_vreinterpretv4siv16qi (__a)
7169
7170 #define vreinterpretq_u32_s16(__a) \
7171   (uint32x4_t)__builtin_neon_vreinterpretv4siv8hi (__a)
7172
7173 #define vreinterpretq_u32_s32(__a) \
7174   (uint32x4_t)__builtin_neon_vreinterpretv4siv4si (__a)
7175
7176 #define vreinterpretq_u32_s64(__a) \
7177   (uint32x4_t)__builtin_neon_vreinterpretv4siv2di (__a)
7178
7179 #define vreinterpretq_u32_f32(__a) \
7180   (uint32x4_t)__builtin_neon_vreinterpretv4siv4sf (__a)
7181
7182 #define vreinterpretq_u32_u8(__a) \
7183   (uint32x4_t)__builtin_neon_vreinterpretv4siv16qi (__a)
7184
7185 #define vreinterpretq_u32_u16(__a) \
7186   (uint32x4_t)__builtin_neon_vreinterpretv4siv8hi (__a)
7187
7188 #define vreinterpretq_u32_u64(__a) \
7189   (uint32x4_t)__builtin_neon_vreinterpretv4siv2di (__a)
7190
7191 #define vreinterpretq_u32_p8(__a) \
7192   (uint32x4_t)__builtin_neon_vreinterpretv4siv16qi (__a)
7193
7194 #define vreinterpretq_u32_p16(__a) \
7195   (uint32x4_t)__builtin_neon_vreinterpretv4siv8hi (__a)
7196
7197 #ifdef __cplusplus
7198 }
7199 #endif
7200 #endif
7201 #endif