1 /* ------------------------------------------------------------------ 2 * Copyright (C) 1998-2009 PacketVideo 3 * 4 * Licensed under the Apache License, Version 2.0 (the "License"); 5 * you may not use this file except in compliance with the License. 6 * You may obtain a copy of the License at 7 * 8 * http://www.apache.org/licenses/LICENSE-2.0 9 * 10 * Unless required by applicable law or agreed to in writing, software 11 * distributed under the License is distributed on an "AS IS" BASIS, 12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either 13 * express or implied. 14 * See the License for the specific language governing permissions 15 * and limitations under the License. 16 * ------------------------------------------------------------------- 17 */ 18 /* 19 20 Pathname: ./c/include/fxp_mul32_arm_v5.h 21 22 ------------------------------------------------------------------------------ 23 REVISION HISTORY 24 25 Who: Date: 26 Description: 27 ------------------------------------------------------------------------------ 28 INCLUDE DESCRIPTION 29 30 ------------------------------------------------------------------------------ 31 */ 32 33 #ifndef FXP_MUL32_ARM_V5 34 #define FXP_MUL32_ARM_V5 35 36 37 #ifdef __cplusplus 38 extern "C" 39 { 40 #endif 41 42 #include "pv_audio_type_defs.h" 43 44 45 #if defined(PV_ARM_V5) 46 47 //#undef EXTENDED_ASM 48 #define EXTENDED_ASM 49 #define _ARM_V5_ 50 51 shft_lft_1(Int32 L_var1)52 __inline Int32 shft_lft_1(Int32 L_var1) 53 { 54 __asm 55 { 56 qadd L_var1, L_var1, L_var1 57 } 58 59 return L_var1; 60 } 61 62 fxp_mul_16_by_16(Int32 L_var1,Int32 L_var2)63 __inline Int32 fxp_mul_16_by_16(Int32 L_var1, Int32 L_var2) 64 { 65 __asm 66 { 67 smulbb L_var1, L_var1, L_var2 68 } 69 return L_var1; 70 } 71 72 fxp_mul_16_by_16bb(Int32 L_var1,Int32 L_var2)73 __inline Int32 fxp_mul_16_by_16bb(Int32 L_var1, Int32 L_var2) 74 { 75 __asm 76 { 77 smulbb L_var1, L_var1, L_var2 78 } 79 return L_var1; 80 } 81 82 fxp_mul_16_by_16tb(Int32 L_var1,Int32 L_var2)83 __inline Int32 fxp_mul_16_by_16tb(Int32 L_var1, Int32 L_var2) 84 { 85 __asm 86 { 87 smultb L_var1, L_var1, L_var2 88 } 89 return L_var1; 90 } 91 fxp_mul_16_by_16tt(Int32 L_var1,Int32 L_var2)92 __inline Int32 fxp_mul_16_by_16tt(Int32 L_var1, Int32 L_var2) 93 { 94 __asm 95 { 96 smultt L_var1, L_var1, L_var2 97 } 98 return L_var1; 99 } 100 fxp_mul_16_by_16bt(Int32 L_var1,Int32 L_var2)101 __inline Int32 fxp_mul_16_by_16bt(Int32 L_var1, Int32 L_var2) 102 { 103 __asm 104 { 105 smulbt L_var1, L_var1, L_var2 106 } 107 return L_var1; 108 } 109 110 111 fxp_mac_16_by_16(const Int32 L_var1,const Int32 L_var2,Int32 L_add)112 __inline Int32 fxp_mac_16_by_16(const Int32 L_var1, const Int32 L_var2, Int32 L_add) 113 { 114 __asm 115 { 116 smlabb L_add, L_var1, L_var2, L_add 117 } 118 return (L_add); 119 } 120 fxp_mac_16_by_16_bb(const Int32 L_var1,Int32 L_var2,Int32 L_add)121 __inline Int32 fxp_mac_16_by_16_bb(const Int32 L_var1, Int32 L_var2, Int32 L_add) 122 { 123 __asm 124 { 125 smlabb L_add, L_var1, L_var2, L_add 126 } 127 return L_add; 128 } 129 fxp_mac_16_by_16_bt(const Int32 L_var1,Int32 L_var2,Int32 L_add)130 __inline Int32 fxp_mac_16_by_16_bt(const Int32 L_var1, Int32 L_var2, Int32 L_add) 131 { 132 __asm 133 { 134 smlabt L_add, L_var1, L_var2, L_add 135 } 136 return L_add; 137 } 138 139 fxp_mac_16_by_16_tb(const Int32 L_var1,Int32 L_var2,Int32 L_add)140 __inline Int32 fxp_mac_16_by_16_tb(const Int32 L_var1, Int32 L_var2, Int32 L_add) 141 { 142 __asm 143 { 144 smlatb L_add, L_var1, L_var2, L_add 145 } 146 return L_add; 147 } 148 fxp_mac_16_by_16_tt(const Int32 L_var1,Int32 L_var2,Int32 L_add)149 __inline Int32 fxp_mac_16_by_16_tt(const Int32 L_var1, Int32 L_var2, Int32 L_add) 150 { 151 __asm 152 { 153 smlatt L_add, L_var1, L_var2, L_add 154 } 155 return L_add; 156 } 157 fxp_mac32_by_16(Int32 L_var1,const Int32 L_var2,Int32 L_add)158 __inline Int32 fxp_mac32_by_16(Int32 L_var1, const Int32 L_var2, Int32 L_add) 159 { 160 __asm 161 { 162 smlawb L_add, L_var1, L_var2, L_add 163 } 164 return (L_add); 165 } 166 167 fxp_mac64_Q31(int64 sum,const Int32 L_var1,const Int32 L_var2)168 __inline int64 fxp_mac64_Q31(int64 sum, const Int32 L_var1, const Int32 L_var2) 169 { 170 uint32 b = (UInt32)(sum); 171 int32 c = Int32(sum >> 32); 172 __asm 173 { 174 smlal b, c, L_var1, L_var2 175 } 176 return (((int64(c)) << 32) | b); 177 } 178 179 fxp_mac32_Q31(Int32 L_add,Int32 L_var1,const Int32 L_var2)180 __inline Int32 fxp_mac32_Q31(Int32 L_add, Int32 L_var1, const Int32 L_var2) 181 { 182 __asm 183 { 184 smlal L_var1, L_add, L_var2, L_var1 185 } 186 return L_add; 187 } 188 fxp_msu32_Q31(Int32 L_sub,Int32 L_var1,const Int32 L_var2)189 __inline Int32 fxp_msu32_Q31(Int32 L_sub, Int32 L_var1, const Int32 L_var2) 190 { 191 __asm 192 { 193 rsb L_var1, L_var1, #0 194 smlal L_var1, L_sub, L_var2, L_var1 195 } 196 return L_sub; 197 } 198 fxp_mul32_Q31(Int32 L_var1,const Int32 L_var2)199 __inline Int32 fxp_mul32_Q31(Int32 L_var1, const Int32 L_var2) 200 { 201 Int32 result64_hi; 202 __asm 203 { 204 smull L_var1, result64_hi, L_var2, L_var1 205 } 206 return (result64_hi); 207 } 208 fxp_mul32_Q30(const Int32 L_var1,const Int32 L_var2)209 __inline Int32 fxp_mul32_Q30(const Int32 L_var1, const Int32 L_var2) 210 { 211 Int32 result64_hi; 212 Int32 result64_lo; 213 __asm 214 { 215 smull result64_lo, result64_hi, L_var2, L_var1 216 mov result64_hi, result64_hi, asl #2 217 #ifdef EXTENDED_ASM 218 mov result64_lo, result64_lo, lsr #30 219 orr result64_hi, result64_lo, result64_hi 220 #else 221 orr result64_hi, result64_hi, result64_lo, lsr #30 222 #endif 223 } 224 return (result64_hi); 225 } 226 227 fxp_mac32_Q30(const Int32 L_var1,const Int32 L_var2,Int32 L_add)228 __inline Int32 fxp_mac32_Q30(const Int32 L_var1, const Int32 L_var2, Int32 L_add) 229 { 230 Int32 result64_hi; 231 Int32 result64_lo; 232 __asm 233 { 234 smull result64_lo, result64_hi, L_var2, L_var1 235 add L_add, L_add, result64_hi, asl #2 236 add L_add, L_add, result64_lo, lsr #30 237 } 238 return (L_add); 239 } 240 241 fxp_mul32_Q29(const Int32 L_var1,const Int32 L_var2)242 __inline Int32 fxp_mul32_Q29(const Int32 L_var1, const Int32 L_var2) 243 { 244 Int32 result64_hi; 245 Int32 result64_lo; 246 __asm 247 { 248 smull result64_lo, result64_hi, L_var2, L_var1 249 mov result64_hi, result64_hi, asl #3 250 #ifdef EXTENDED_ASM 251 mov result64_lo, result64_lo, lsr #29 252 orr result64_hi, result64_lo, result64_hi 253 #else 254 orr result64_hi, result64_hi, result64_lo, lsr #29 255 #endif 256 } 257 return (result64_hi); 258 } 259 260 261 fxp_mac32_Q29(const Int32 L_var1,const Int32 L_var2,Int32 L_add)262 __inline Int32 fxp_mac32_Q29(const Int32 L_var1, const Int32 L_var2, Int32 L_add) 263 { 264 Int32 result64_hi; 265 Int32 result64_lo; 266 __asm 267 { 268 smull result64_lo, result64_hi, L_var2, L_var1 269 add L_add, L_add, result64_hi, asl #3 270 add L_add, L_add, result64_lo, lsr #29 271 } 272 return (L_add); 273 } 274 275 fxp_msu32_Q29(const Int32 L_var1,const Int32 L_var2,Int32 L_sub)276 __inline Int32 fxp_msu32_Q29(const Int32 L_var1, const Int32 L_var2, Int32 L_sub) 277 { 278 Int32 result64_hi; 279 Int32 result64_lo; 280 __asm 281 { 282 smull result64_lo, result64_hi, L_var2, L_var1 283 sub L_sub, L_sub, result64_hi, asl #3 284 sub L_sub, L_sub, result64_lo, lsr #29 285 } 286 return (L_sub); 287 } 288 289 fxp_mul32_Q28(const Int32 L_var1,const Int32 L_var2)290 __inline Int32 fxp_mul32_Q28(const Int32 L_var1, const Int32 L_var2) 291 { 292 Int32 result64_hi; 293 Int32 result64_lo; 294 __asm 295 { 296 smull result64_lo, result64_hi, L_var2, L_var1 297 mov result64_hi, result64_hi, asl #4 298 #ifdef EXTENDED_ASM 299 mov result64_lo, result64_lo, lsr #28 300 orr result64_hi, result64_lo, result64_hi 301 #else 302 orr result64_hi, result64_hi, result64_lo, lsr #28 303 #endif 304 305 } 306 return (result64_hi); 307 } 308 fxp_mul32_Q27(const Int32 L_var1,const Int32 L_var2)309 __inline Int32 fxp_mul32_Q27(const Int32 L_var1, const Int32 L_var2) 310 { 311 Int32 result64_hi; 312 Int32 result64_lo; 313 __asm 314 { 315 smull result64_lo, result64_hi, L_var2, L_var1 316 mov result64_hi, result64_hi, asl #5 317 #ifdef EXTENDED_ASM 318 mov result64_lo, result64_lo, lsr #27 319 orr result64_hi, result64_lo, result64_hi 320 #else 321 orr result64_hi, result64_hi, result64_lo, lsr #27 322 #endif 323 } 324 return (result64_hi); 325 } 326 fxp_mul32_Q26(const Int32 L_var1,const Int32 L_var2)327 __inline Int32 fxp_mul32_Q26(const Int32 L_var1, const Int32 L_var2) 328 { 329 Int32 result64_hi; 330 Int32 result64_lo; 331 __asm 332 { 333 smull result64_lo, result64_hi, L_var2, L_var1 334 mov result64_hi, result64_hi, asl #6 335 #ifdef EXTENDED_ASM 336 mov result64_lo, result64_lo, lsr #26 337 orr result64_hi, result64_lo, result64_hi 338 #else 339 orr result64_hi, result64_hi, result64_lo, lsr #26 340 #endif 341 342 } 343 return (result64_hi); 344 } 345 fxp_mul32_Q20(const Int32 L_var1,const Int32 L_var2)346 __inline Int32 fxp_mul32_Q20(const Int32 L_var1, const Int32 L_var2) 347 { 348 Int32 result64_hi; 349 Int32 result64_lo; 350 __asm 351 { 352 smull result64_lo, result64_hi, L_var2, L_var1 353 mov result64_hi, result64_hi, asl #12 354 #ifdef EXTENDED_ASM 355 mov result64_lo, result64_lo, lsr #20 356 orr result64_hi, result64_lo, result64_hi 357 #else 358 orr result64_hi, result64_hi, result64_lo, lsr #20 359 #endif 360 } 361 return (result64_hi); 362 } 363 fxp_mul32_by_16(Int32 L_var1,const Int32 L_var2)364 __inline Int32 fxp_mul32_by_16(Int32 L_var1, const Int32 L_var2) 365 { 366 Int32 result64_hi; 367 __asm 368 { 369 smulwb result64_hi, L_var1, L_var2 370 } 371 return (result64_hi); 372 } 373 374 #define fxp_mul32_by_16b( a, b) fxp_mul32_by_16(a, b) 375 fxp_mul32_by_16t(Int32 L_var1,const Int32 L_var2)376 __inline Int32 fxp_mul32_by_16t(Int32 L_var1, const Int32 L_var2) 377 { 378 Int32 result64_hi; 379 __asm 380 { 381 smulwt result64_hi, L_var1, L_var2 382 } 383 return (result64_hi); 384 } 385 fxp_mul32_Q15(const Int32 L_var1,const Int32 L_var2)386 __inline Int32 fxp_mul32_Q15(const Int32 L_var1, const Int32 L_var2) 387 { 388 Int32 result64_hi; 389 Int32 result64_lo; 390 __asm 391 { 392 smull result64_lo, result64_hi, L_var2, L_var1 393 mov result64_hi, result64_hi, asl #17 394 #ifdef EXTENDED_ASM 395 mov result64_lo, result64_lo, lsr #15 396 orr result64_hi, result64_lo, result64_hi 397 #else 398 orr result64_hi, result64_hi, result64_lo, lsr #15 399 #endif 400 } 401 return (result64_hi); 402 } 403 404 cmplx_mul32_by_16(Int32 L_var1,const Int32 L_var2,const Int32 cmplx)405 __inline Int32 cmplx_mul32_by_16(Int32 L_var1, const Int32 L_var2, const Int32 cmplx) 406 { 407 Int32 result64_hi; 408 409 __asm 410 { 411 smulwt result64_hi, L_var1, cmplx 412 smlawb result64_hi, L_var2, cmplx, result64_hi 413 } 414 return (result64_hi); 415 416 } 417 fxp_mul32_Q14(const Int32 L_var1,const Int32 L_var2)418 __inline Int32 fxp_mul32_Q14(const Int32 L_var1, const Int32 L_var2) 419 { 420 Int32 result64_hi; 421 Int32 result64_lo; 422 __asm 423 { 424 smull result64_lo, result64_hi, L_var2, L_var1 425 mov result64_hi, result64_hi, asl #18 426 #ifdef EXTENDED_ASM 427 mov result64_lo, result64_lo, lsr #14 428 orr result64_hi, result64_lo, result64_hi 429 #else 430 orr result64_hi, result64_hi, result64_lo, lsr #14 431 #endif 432 } 433 return (result64_hi); 434 } 435 436 437 #define preload_cache( a) 438 439 440 441 442 #endif 443 444 #ifdef __cplusplus 445 } 446 #endif 447 448 449 #endif /* FXP_MUL32 */ 450 451