1 /* 2 * QEMU float support 3 * 4 * The code in this source file is derived from release 2a of the SoftFloat 5 * IEC/IEEE Floating-point Arithmetic Package. Those parts of the code (and 6 * some later contributions) are provided under that license, as detailed below. 7 * It has subsequently been modified by contributors to the QEMU Project, 8 * so some portions are provided under: 9 * the SoftFloat-2a license 10 * the BSD license 11 * GPL-v2-or-later 12 * 13 * Any future contributions to this file after December 1st 2014 will be 14 * taken to be licensed under the Softfloat-2a license unless specifically 15 * indicated otherwise. 16 */ 17 18 /* 19 =============================================================================== 20 This C header file is part of the SoftFloat IEC/IEEE Floating-point 21 Arithmetic Package, Release 2a. 22 23 Written by John R. Hauser. This work was made possible in part by the 24 International Computer Science Institute, located at Suite 600, 1947 Center 25 Street, Berkeley, California 94704. Funding was partially provided by the 26 National Science Foundation under grant MIP-9311980. The original version 27 of this code was written as part of a project to build a fixed-point vector 28 processor in collaboration with the University of California at Berkeley, 29 overseen by Profs. Nelson Morgan and John Wawrzynek. More information 30 is available through the Web page `http://HTTP.CS.Berkeley.EDU/~jhauser/ 31 arithmetic/SoftFloat.html'. 32 33 THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE. Although reasonable effort 34 has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT 35 TIMES RESULT IN INCORRECT BEHAVIOR. USE OF THIS SOFTWARE IS RESTRICTED TO 36 PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY 37 AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE. 38 39 Derivative works are acceptable, even for commercial purposes, so long as 40 (1) they include prominent notice that the work is derivative, and (2) they 41 include prominent notice akin to these four paragraphs for those parts of 42 this code that are retained. 43 44 =============================================================================== 45 */ 46 47 /* BSD licensing: 48 * Copyright (c) 2006, Fabrice Bellard 49 * All rights reserved. 50 * 51 * Redistribution and use in source and binary forms, with or without 52 * modification, are permitted provided that the following conditions are met: 53 * 54 * 1. Redistributions of source code must retain the above copyright notice, 55 * this list of conditions and the following disclaimer. 56 * 57 * 2. Redistributions in binary form must reproduce the above copyright notice, 58 * this list of conditions and the following disclaimer in the documentation 59 * and/or other materials provided with the distribution. 60 * 61 * 3. Neither the name of the copyright holder nor the names of its contributors 62 * may be used to endorse or promote products derived from this software without 63 * specific prior written permission. 64 * 65 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" 66 * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 67 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 68 * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE 69 * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 70 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 71 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 72 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 73 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 74 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF 75 * THE POSSIBILITY OF SUCH DAMAGE. 76 */ 77 78 /* Portions of this work are licensed under the terms of the GNU GPL, 79 * version 2 or later. See the COPYING file in the top-level directory. 80 */ 81 82 #ifndef SOFTFLOAT_H 83 #define SOFTFLOAT_H 84 85 /*---------------------------------------------------------------------------- 86 | Software IEC/IEEE floating-point ordering relations 87 *----------------------------------------------------------------------------*/ 88 89 typedef enum { 90 float_relation_less = -1, 91 float_relation_equal = 0, 92 float_relation_greater = 1, 93 float_relation_unordered = 2 94 } FloatRelation; 95 96 #include "fpu/softfloat-types.h" 97 #include "fpu/softfloat-helpers.h" 98 99 /*---------------------------------------------------------------------------- 100 | Routine to raise any or all of the software IEC/IEEE floating-point 101 | exception flags. 102 *----------------------------------------------------------------------------*/ 103 void float_raise(uint8_t flags, float_status *status); 104 105 /*---------------------------------------------------------------------------- 106 | If `a' is denormal and we are in flush-to-zero mode then set the 107 | input-denormal exception and return zero. Otherwise just return the value. 108 *----------------------------------------------------------------------------*/ 109 float16 float16_squash_input_denormal(float16 a, float_status *status); 110 float32 float32_squash_input_denormal(float32 a, float_status *status); 111 float64 float64_squash_input_denormal(float64 a, float_status *status); 112 113 /*---------------------------------------------------------------------------- 114 | Options to indicate which negations to perform in float*_muladd() 115 | Using these differs from negating an input or output before calling 116 | the muladd function in that this means that a NaN doesn't have its 117 | sign bit inverted before it is propagated. 118 | We also support halving the result before rounding, as a special 119 | case to support the ARM fused-sqrt-step instruction FRSQRTS. 120 *----------------------------------------------------------------------------*/ 121 enum { 122 float_muladd_negate_c = 1, 123 float_muladd_negate_product = 2, 124 float_muladd_negate_result = 4, 125 float_muladd_halve_result = 8, 126 }; 127 128 /*---------------------------------------------------------------------------- 129 | Software IEC/IEEE integer-to-floating-point conversion routines. 130 *----------------------------------------------------------------------------*/ 131 132 float16 int16_to_float16_scalbn(int16_t a, int, float_status *status); 133 float16 int32_to_float16_scalbn(int32_t a, int, float_status *status); 134 float16 int64_to_float16_scalbn(int64_t a, int, float_status *status); 135 float16 uint16_to_float16_scalbn(uint16_t a, int, float_status *status); 136 float16 uint32_to_float16_scalbn(uint32_t a, int, float_status *status); 137 float16 uint64_to_float16_scalbn(uint64_t a, int, float_status *status); 138 139 float16 int16_to_float16(int16_t a, float_status *status); 140 float16 int32_to_float16(int32_t a, float_status *status); 141 float16 int64_to_float16(int64_t a, float_status *status); 142 float16 uint16_to_float16(uint16_t a, float_status *status); 143 float16 uint32_to_float16(uint32_t a, float_status *status); 144 float16 uint64_to_float16(uint64_t a, float_status *status); 145 146 float32 int16_to_float32_scalbn(int16_t, int, float_status *status); 147 float32 int32_to_float32_scalbn(int32_t, int, float_status *status); 148 float32 int64_to_float32_scalbn(int64_t, int, float_status *status); 149 float32 uint16_to_float32_scalbn(uint16_t, int, float_status *status); 150 float32 uint32_to_float32_scalbn(uint32_t, int, float_status *status); 151 float32 uint64_to_float32_scalbn(uint64_t, int, float_status *status); 152 153 float32 int16_to_float32(int16_t, float_status *status); 154 float32 int32_to_float32(int32_t, float_status *status); 155 float32 int64_to_float32(int64_t, float_status *status); 156 float32 uint16_to_float32(uint16_t, float_status *status); 157 float32 uint32_to_float32(uint32_t, float_status *status); 158 float32 uint64_to_float32(uint64_t, float_status *status); 159 160 float64 int16_to_float64_scalbn(int16_t, int, float_status *status); 161 float64 int32_to_float64_scalbn(int32_t, int, float_status *status); 162 float64 int64_to_float64_scalbn(int64_t, int, float_status *status); 163 float64 uint16_to_float64_scalbn(uint16_t, int, float_status *status); 164 float64 uint32_to_float64_scalbn(uint32_t, int, float_status *status); 165 float64 uint64_to_float64_scalbn(uint64_t, int, float_status *status); 166 167 float64 int16_to_float64(int16_t, float_status *status); 168 float64 int32_to_float64(int32_t, float_status *status); 169 float64 int64_to_float64(int64_t, float_status *status); 170 float64 uint16_to_float64(uint16_t, float_status *status); 171 float64 uint32_to_float64(uint32_t, float_status *status); 172 float64 uint64_to_float64(uint64_t, float_status *status); 173 174 floatx80 int32_to_floatx80(int32_t, float_status *status); 175 floatx80 int64_to_floatx80(int64_t, float_status *status); 176 177 float128 int32_to_float128(int32_t, float_status *status); 178 float128 int64_to_float128(int64_t, float_status *status); 179 float128 uint64_to_float128(uint64_t, float_status *status); 180 181 /*---------------------------------------------------------------------------- 182 | Software half-precision conversion routines. 183 *----------------------------------------------------------------------------*/ 184 185 float16 float32_to_float16(float32, bool ieee, float_status *status); 186 float32 float16_to_float32(float16, bool ieee, float_status *status); 187 float16 float64_to_float16(float64 a, bool ieee, float_status *status); 188 float64 float16_to_float64(float16 a, bool ieee, float_status *status); 189 190 int16_t float16_to_int16_scalbn(float16, FloatRoundMode, int, float_status *); 191 int32_t float16_to_int32_scalbn(float16, FloatRoundMode, int, float_status *); 192 int64_t float16_to_int64_scalbn(float16, FloatRoundMode, int, float_status *); 193 194 int16_t float16_to_int16(float16, float_status *status); 195 int32_t float16_to_int32(float16, float_status *status); 196 int64_t float16_to_int64(float16, float_status *status); 197 198 int16_t float16_to_int16_round_to_zero(float16, float_status *status); 199 int32_t float16_to_int32_round_to_zero(float16, float_status *status); 200 int64_t float16_to_int64_round_to_zero(float16, float_status *status); 201 202 uint16_t float16_to_uint16_scalbn(float16 a, FloatRoundMode, 203 int, float_status *status); 204 uint32_t float16_to_uint32_scalbn(float16 a, FloatRoundMode, 205 int, float_status *status); 206 uint64_t float16_to_uint64_scalbn(float16 a, FloatRoundMode, 207 int, float_status *status); 208 209 uint16_t float16_to_uint16(float16 a, float_status *status); 210 uint32_t float16_to_uint32(float16 a, float_status *status); 211 uint64_t float16_to_uint64(float16 a, float_status *status); 212 213 uint16_t float16_to_uint16_round_to_zero(float16 a, float_status *status); 214 uint32_t float16_to_uint32_round_to_zero(float16 a, float_status *status); 215 uint64_t float16_to_uint64_round_to_zero(float16 a, float_status *status); 216 217 /*---------------------------------------------------------------------------- 218 | Software half-precision operations. 219 *----------------------------------------------------------------------------*/ 220 221 float16 float16_round_to_int(float16, float_status *status); 222 float16 float16_add(float16, float16, float_status *status); 223 float16 float16_sub(float16, float16, float_status *status); 224 float16 float16_mul(float16, float16, float_status *status); 225 float16 float16_muladd(float16, float16, float16, int, float_status *status); 226 float16 float16_div(float16, float16, float_status *status); 227 float16 float16_scalbn(float16, int, float_status *status); 228 float16 float16_min(float16, float16, float_status *status); 229 float16 float16_max(float16, float16, float_status *status); 230 float16 float16_minnum(float16, float16, float_status *status); 231 float16 float16_maxnum(float16, float16, float_status *status); 232 float16 float16_minnummag(float16, float16, float_status *status); 233 float16 float16_maxnummag(float16, float16, float_status *status); 234 float16 float16_sqrt(float16, float_status *status); 235 FloatRelation float16_compare(float16, float16, float_status *status); 236 FloatRelation float16_compare_quiet(float16, float16, float_status *status); 237 238 bool float16_is_quiet_nan(float16, float_status *status); 239 bool float16_is_signaling_nan(float16, float_status *status); 240 float16 float16_silence_nan(float16, float_status *status); 241 242 static inline bool float16_is_any_nan(float16 a) 243 { 244 return ((float16_val(a) & ~0x8000) > 0x7c00); 245 } 246 247 static inline bool float16_is_neg(float16 a) 248 { 249 return float16_val(a) >> 15; 250 } 251 252 static inline bool float16_is_infinity(float16 a) 253 { 254 return (float16_val(a) & 0x7fff) == 0x7c00; 255 } 256 257 static inline bool float16_is_zero(float16 a) 258 { 259 return (float16_val(a) & 0x7fff) == 0; 260 } 261 262 static inline bool float16_is_zero_or_denormal(float16 a) 263 { 264 return (float16_val(a) & 0x7c00) == 0; 265 } 266 267 static inline float16 float16_abs(float16 a) 268 { 269 /* Note that abs does *not* handle NaN specially, nor does 270 * it flush denormal inputs to zero. 271 */ 272 return make_float16(float16_val(a) & 0x7fff); 273 } 274 275 static inline float16 float16_chs(float16 a) 276 { 277 /* Note that chs does *not* handle NaN specially, nor does 278 * it flush denormal inputs to zero. 279 */ 280 return make_float16(float16_val(a) ^ 0x8000); 281 } 282 283 static inline float16 float16_set_sign(float16 a, int sign) 284 { 285 return make_float16((float16_val(a) & 0x7fff) | (sign << 15)); 286 } 287 288 #define float16_zero make_float16(0) 289 #define float16_half make_float16(0x3800) 290 #define float16_one make_float16(0x3c00) 291 #define float16_one_point_five make_float16(0x3e00) 292 #define float16_two make_float16(0x4000) 293 #define float16_three make_float16(0x4200) 294 #define float16_infinity make_float16(0x7c00) 295 296 /*---------------------------------------------------------------------------- 297 | The pattern for a default generated half-precision NaN. 298 *----------------------------------------------------------------------------*/ 299 float16 float16_default_nan(float_status *status); 300 301 /*---------------------------------------------------------------------------- 302 | Software IEC/IEEE single-precision conversion routines. 303 *----------------------------------------------------------------------------*/ 304 305 int16_t float32_to_int16_scalbn(float32, FloatRoundMode, int, float_status *); 306 int32_t float32_to_int32_scalbn(float32, FloatRoundMode, int, float_status *); 307 int64_t float32_to_int64_scalbn(float32, FloatRoundMode, int, float_status *); 308 309 int16_t float32_to_int16(float32, float_status *status); 310 int32_t float32_to_int32(float32, float_status *status); 311 int64_t float32_to_int64(float32, float_status *status); 312 313 int16_t float32_to_int16_round_to_zero(float32, float_status *status); 314 int32_t float32_to_int32_round_to_zero(float32, float_status *status); 315 int64_t float32_to_int64_round_to_zero(float32, float_status *status); 316 317 uint16_t float32_to_uint16_scalbn(float32, FloatRoundMode, int, float_status *); 318 uint32_t float32_to_uint32_scalbn(float32, FloatRoundMode, int, float_status *); 319 uint64_t float32_to_uint64_scalbn(float32, FloatRoundMode, int, float_status *); 320 321 uint16_t float32_to_uint16(float32, float_status *status); 322 uint32_t float32_to_uint32(float32, float_status *status); 323 uint64_t float32_to_uint64(float32, float_status *status); 324 325 uint16_t float32_to_uint16_round_to_zero(float32, float_status *status); 326 uint32_t float32_to_uint32_round_to_zero(float32, float_status *status); 327 uint64_t float32_to_uint64_round_to_zero(float32, float_status *status); 328 329 float64 float32_to_float64(float32, float_status *status); 330 floatx80 float32_to_floatx80(float32, float_status *status); 331 float128 float32_to_float128(float32, float_status *status); 332 333 /*---------------------------------------------------------------------------- 334 | Software IEC/IEEE single-precision operations. 335 *----------------------------------------------------------------------------*/ 336 float32 float32_round_to_int(float32, float_status *status); 337 float32 float32_add(float32, float32, float_status *status); 338 float32 float32_sub(float32, float32, float_status *status); 339 float32 float32_mul(float32, float32, float_status *status); 340 float32 float32_div(float32, float32, float_status *status); 341 float32 float32_rem(float32, float32, float_status *status); 342 float32 float32_muladd(float32, float32, float32, int, float_status *status); 343 float32 float32_sqrt(float32, float_status *status); 344 float32 float32_exp2(float32, float_status *status); 345 float32 float32_log2(float32, float_status *status); 346 FloatRelation float32_compare(float32, float32, float_status *status); 347 FloatRelation float32_compare_quiet(float32, float32, float_status *status); 348 float32 float32_min(float32, float32, float_status *status); 349 float32 float32_max(float32, float32, float_status *status); 350 float32 float32_minnum(float32, float32, float_status *status); 351 float32 float32_maxnum(float32, float32, float_status *status); 352 float32 float32_minnummag(float32, float32, float_status *status); 353 float32 float32_maxnummag(float32, float32, float_status *status); 354 bool float32_is_quiet_nan(float32, float_status *status); 355 bool float32_is_signaling_nan(float32, float_status *status); 356 float32 float32_silence_nan(float32, float_status *status); 357 float32 float32_scalbn(float32, int, float_status *status); 358 359 static inline float32 float32_abs(float32 a) 360 { 361 /* Note that abs does *not* handle NaN specially, nor does 362 * it flush denormal inputs to zero. 363 */ 364 return make_float32(float32_val(a) & 0x7fffffff); 365 } 366 367 static inline float32 float32_chs(float32 a) 368 { 369 /* Note that chs does *not* handle NaN specially, nor does 370 * it flush denormal inputs to zero. 371 */ 372 return make_float32(float32_val(a) ^ 0x80000000); 373 } 374 375 static inline bool float32_is_infinity(float32 a) 376 { 377 return (float32_val(a) & 0x7fffffff) == 0x7f800000; 378 } 379 380 static inline bool float32_is_neg(float32 a) 381 { 382 return float32_val(a) >> 31; 383 } 384 385 static inline bool float32_is_zero(float32 a) 386 { 387 return (float32_val(a) & 0x7fffffff) == 0; 388 } 389 390 static inline bool float32_is_any_nan(float32 a) 391 { 392 return ((float32_val(a) & ~(1 << 31)) > 0x7f800000UL); 393 } 394 395 static inline bool float32_is_zero_or_denormal(float32 a) 396 { 397 return (float32_val(a) & 0x7f800000) == 0; 398 } 399 400 static inline bool float32_is_normal(float32 a) 401 { 402 return (((float32_val(a) >> 23) + 1) & 0xff) >= 2; 403 } 404 405 static inline bool float32_is_denormal(float32 a) 406 { 407 return float32_is_zero_or_denormal(a) && !float32_is_zero(a); 408 } 409 410 static inline bool float32_is_zero_or_normal(float32 a) 411 { 412 return float32_is_normal(a) || float32_is_zero(a); 413 } 414 415 static inline float32 float32_set_sign(float32 a, int sign) 416 { 417 return make_float32((float32_val(a) & 0x7fffffff) | (sign << 31)); 418 } 419 420 static inline bool float32_eq(float32 a, float32 b, float_status *s) 421 { 422 return float32_compare(a, b, s) == float_relation_equal; 423 } 424 425 static inline bool float32_le(float32 a, float32 b, float_status *s) 426 { 427 return float32_compare(a, b, s) <= float_relation_equal; 428 } 429 430 static inline bool float32_lt(float32 a, float32 b, float_status *s) 431 { 432 return float32_compare(a, b, s) < float_relation_equal; 433 } 434 435 static inline bool float32_unordered(float32 a, float32 b, float_status *s) 436 { 437 return float32_compare(a, b, s) == float_relation_unordered; 438 } 439 440 static inline bool float32_eq_quiet(float32 a, float32 b, float_status *s) 441 { 442 return float32_compare_quiet(a, b, s) == float_relation_equal; 443 } 444 445 static inline bool float32_le_quiet(float32 a, float32 b, float_status *s) 446 { 447 return float32_compare_quiet(a, b, s) <= float_relation_equal; 448 } 449 450 static inline bool float32_lt_quiet(float32 a, float32 b, float_status *s) 451 { 452 return float32_compare_quiet(a, b, s) < float_relation_equal; 453 } 454 455 static inline bool float32_unordered_quiet(float32 a, float32 b, 456 float_status *s) 457 { 458 return float32_compare_quiet(a, b, s) == float_relation_unordered; 459 } 460 461 #define float32_zero make_float32(0) 462 #define float32_half make_float32(0x3f000000) 463 #define float32_one make_float32(0x3f800000) 464 #define float32_one_point_five make_float32(0x3fc00000) 465 #define float32_two make_float32(0x40000000) 466 #define float32_three make_float32(0x40400000) 467 #define float32_infinity make_float32(0x7f800000) 468 469 /*---------------------------------------------------------------------------- 470 | Packs the sign `zSign', exponent `zExp', and significand `zSig' into a 471 | single-precision floating-point value, returning the result. After being 472 | shifted into the proper positions, the three fields are simply added 473 | together to form the result. This means that any integer portion of `zSig' 474 | will be added into the exponent. Since a properly normalized significand 475 | will have an integer portion equal to 1, the `zExp' input should be 1 less 476 | than the desired result exponent whenever `zSig' is a complete, normalized 477 | significand. 478 *----------------------------------------------------------------------------*/ 479 480 static inline float32 packFloat32(bool zSign, int zExp, uint32_t zSig) 481 { 482 return make_float32( 483 (((uint32_t)zSign) << 31) + (((uint32_t)zExp) << 23) + zSig); 484 } 485 486 /*---------------------------------------------------------------------------- 487 | The pattern for a default generated single-precision NaN. 488 *----------------------------------------------------------------------------*/ 489 float32 float32_default_nan(float_status *status); 490 491 /*---------------------------------------------------------------------------- 492 | Software IEC/IEEE double-precision conversion routines. 493 *----------------------------------------------------------------------------*/ 494 495 int16_t float64_to_int16_scalbn(float64, FloatRoundMode, int, float_status *); 496 int32_t float64_to_int32_scalbn(float64, FloatRoundMode, int, float_status *); 497 int64_t float64_to_int64_scalbn(float64, FloatRoundMode, int, float_status *); 498 499 int16_t float64_to_int16(float64, float_status *status); 500 int32_t float64_to_int32(float64, float_status *status); 501 int64_t float64_to_int64(float64, float_status *status); 502 503 int16_t float64_to_int16_round_to_zero(float64, float_status *status); 504 int32_t float64_to_int32_round_to_zero(float64, float_status *status); 505 int64_t float64_to_int64_round_to_zero(float64, float_status *status); 506 507 uint16_t float64_to_uint16_scalbn(float64, FloatRoundMode, int, float_status *); 508 uint32_t float64_to_uint32_scalbn(float64, FloatRoundMode, int, float_status *); 509 uint64_t float64_to_uint64_scalbn(float64, FloatRoundMode, int, float_status *); 510 511 uint16_t float64_to_uint16(float64, float_status *status); 512 uint32_t float64_to_uint32(float64, float_status *status); 513 uint64_t float64_to_uint64(float64, float_status *status); 514 515 uint16_t float64_to_uint16_round_to_zero(float64, float_status *status); 516 uint32_t float64_to_uint32_round_to_zero(float64, float_status *status); 517 uint64_t float64_to_uint64_round_to_zero(float64, float_status *status); 518 519 float32 float64_to_float32(float64, float_status *status); 520 floatx80 float64_to_floatx80(float64, float_status *status); 521 float128 float64_to_float128(float64, float_status *status); 522 523 /*---------------------------------------------------------------------------- 524 | Software IEC/IEEE double-precision operations. 525 *----------------------------------------------------------------------------*/ 526 float64 float64_round_to_int(float64, float_status *status); 527 float64 float64_add(float64, float64, float_status *status); 528 float64 float64_sub(float64, float64, float_status *status); 529 float64 float64_mul(float64, float64, float_status *status); 530 float64 float64_div(float64, float64, float_status *status); 531 float64 float64_rem(float64, float64, float_status *status); 532 float64 float64_muladd(float64, float64, float64, int, float_status *status); 533 float64 float64_sqrt(float64, float_status *status); 534 float64 float64_log2(float64, float_status *status); 535 FloatRelation float64_compare(float64, float64, float_status *status); 536 FloatRelation float64_compare_quiet(float64, float64, float_status *status); 537 float64 float64_min(float64, float64, float_status *status); 538 float64 float64_max(float64, float64, float_status *status); 539 float64 float64_minnum(float64, float64, float_status *status); 540 float64 float64_maxnum(float64, float64, float_status *status); 541 float64 float64_minnummag(float64, float64, float_status *status); 542 float64 float64_maxnummag(float64, float64, float_status *status); 543 bool float64_is_quiet_nan(float64 a, float_status *status); 544 bool float64_is_signaling_nan(float64, float_status *status); 545 float64 float64_silence_nan(float64, float_status *status); 546 float64 float64_scalbn(float64, int, float_status *status); 547 548 static inline float64 float64_abs(float64 a) 549 { 550 /* Note that abs does *not* handle NaN specially, nor does 551 * it flush denormal inputs to zero. 552 */ 553 return make_float64(float64_val(a) & 0x7fffffffffffffffLL); 554 } 555 556 static inline float64 float64_chs(float64 a) 557 { 558 /* Note that chs does *not* handle NaN specially, nor does 559 * it flush denormal inputs to zero. 560 */ 561 return make_float64(float64_val(a) ^ 0x8000000000000000LL); 562 } 563 564 static inline bool float64_is_infinity(float64 a) 565 { 566 return (float64_val(a) & 0x7fffffffffffffffLL ) == 0x7ff0000000000000LL; 567 } 568 569 static inline bool float64_is_neg(float64 a) 570 { 571 return float64_val(a) >> 63; 572 } 573 574 static inline bool float64_is_zero(float64 a) 575 { 576 return (float64_val(a) & 0x7fffffffffffffffLL) == 0; 577 } 578 579 static inline bool float64_is_any_nan(float64 a) 580 { 581 return ((float64_val(a) & ~(1ULL << 63)) > 0x7ff0000000000000ULL); 582 } 583 584 static inline bool float64_is_zero_or_denormal(float64 a) 585 { 586 return (float64_val(a) & 0x7ff0000000000000LL) == 0; 587 } 588 589 static inline bool float64_is_normal(float64 a) 590 { 591 return (((float64_val(a) >> 52) + 1) & 0x7ff) >= 2; 592 } 593 594 static inline bool float64_is_denormal(float64 a) 595 { 596 return float64_is_zero_or_denormal(a) && !float64_is_zero(a); 597 } 598 599 static inline bool float64_is_zero_or_normal(float64 a) 600 { 601 return float64_is_normal(a) || float64_is_zero(a); 602 } 603 604 static inline float64 float64_set_sign(float64 a, int sign) 605 { 606 return make_float64((float64_val(a) & 0x7fffffffffffffffULL) 607 | ((int64_t)sign << 63)); 608 } 609 610 static inline bool float64_eq(float64 a, float64 b, float_status *s) 611 { 612 return float64_compare(a, b, s) == float_relation_equal; 613 } 614 615 static inline bool float64_le(float64 a, float64 b, float_status *s) 616 { 617 return float64_compare(a, b, s) <= float_relation_equal; 618 } 619 620 static inline bool float64_lt(float64 a, float64 b, float_status *s) 621 { 622 return float64_compare(a, b, s) < float_relation_equal; 623 } 624 625 static inline bool float64_unordered(float64 a, float64 b, float_status *s) 626 { 627 return float64_compare(a, b, s) == float_relation_unordered; 628 } 629 630 static inline bool float64_eq_quiet(float64 a, float64 b, float_status *s) 631 { 632 return float64_compare_quiet(a, b, s) == float_relation_equal; 633 } 634 635 static inline bool float64_le_quiet(float64 a, float64 b, float_status *s) 636 { 637 return float64_compare_quiet(a, b, s) <= float_relation_equal; 638 } 639 640 static inline bool float64_lt_quiet(float64 a, float64 b, float_status *s) 641 { 642 return float64_compare_quiet(a, b, s) < float_relation_equal; 643 } 644 645 static inline bool float64_unordered_quiet(float64 a, float64 b, 646 float_status *s) 647 { 648 return float64_compare_quiet(a, b, s) == float_relation_unordered; 649 } 650 651 #define float64_zero make_float64(0) 652 #define float64_half make_float64(0x3fe0000000000000LL) 653 #define float64_one make_float64(0x3ff0000000000000LL) 654 #define float64_one_point_five make_float64(0x3FF8000000000000ULL) 655 #define float64_two make_float64(0x4000000000000000ULL) 656 #define float64_three make_float64(0x4008000000000000ULL) 657 #define float64_ln2 make_float64(0x3fe62e42fefa39efLL) 658 #define float64_infinity make_float64(0x7ff0000000000000LL) 659 660 /*---------------------------------------------------------------------------- 661 | The pattern for a default generated double-precision NaN. 662 *----------------------------------------------------------------------------*/ 663 float64 float64_default_nan(float_status *status); 664 665 /*---------------------------------------------------------------------------- 666 | Software IEC/IEEE extended double-precision conversion routines. 667 *----------------------------------------------------------------------------*/ 668 int32_t floatx80_to_int32(floatx80, float_status *status); 669 int32_t floatx80_to_int32_round_to_zero(floatx80, float_status *status); 670 int64_t floatx80_to_int64(floatx80, float_status *status); 671 int64_t floatx80_to_int64_round_to_zero(floatx80, float_status *status); 672 float32 floatx80_to_float32(floatx80, float_status *status); 673 float64 floatx80_to_float64(floatx80, float_status *status); 674 float128 floatx80_to_float128(floatx80, float_status *status); 675 676 /*---------------------------------------------------------------------------- 677 | The pattern for an extended double-precision inf. 678 *----------------------------------------------------------------------------*/ 679 extern const floatx80 floatx80_infinity; 680 681 /*---------------------------------------------------------------------------- 682 | Software IEC/IEEE extended double-precision operations. 683 *----------------------------------------------------------------------------*/ 684 floatx80 floatx80_round(floatx80 a, float_status *status); 685 floatx80 floatx80_round_to_int(floatx80, float_status *status); 686 floatx80 floatx80_add(floatx80, floatx80, float_status *status); 687 floatx80 floatx80_sub(floatx80, floatx80, float_status *status); 688 floatx80 floatx80_mul(floatx80, floatx80, float_status *status); 689 floatx80 floatx80_div(floatx80, floatx80, float_status *status); 690 floatx80 floatx80_rem(floatx80, floatx80, float_status *status); 691 floatx80 floatx80_sqrt(floatx80, float_status *status); 692 FloatRelation floatx80_compare(floatx80, floatx80, float_status *status); 693 FloatRelation floatx80_compare_quiet(floatx80, floatx80, float_status *status); 694 int floatx80_is_quiet_nan(floatx80, float_status *status); 695 int floatx80_is_signaling_nan(floatx80, float_status *status); 696 floatx80 floatx80_silence_nan(floatx80, float_status *status); 697 floatx80 floatx80_scalbn(floatx80, int, float_status *status); 698 699 static inline floatx80 floatx80_abs(floatx80 a) 700 { 701 a.high &= 0x7fff; 702 return a; 703 } 704 705 static inline floatx80 floatx80_chs(floatx80 a) 706 { 707 a.high ^= 0x8000; 708 return a; 709 } 710 711 static inline bool floatx80_is_infinity(floatx80 a) 712 { 713 #if defined(TARGET_M68K) 714 return (a.high & 0x7fff) == floatx80_infinity.high && !(a.low << 1); 715 #else 716 return (a.high & 0x7fff) == floatx80_infinity.high && 717 a.low == floatx80_infinity.low; 718 #endif 719 } 720 721 static inline bool floatx80_is_neg(floatx80 a) 722 { 723 return a.high >> 15; 724 } 725 726 static inline bool floatx80_is_zero(floatx80 a) 727 { 728 return (a.high & 0x7fff) == 0 && a.low == 0; 729 } 730 731 static inline bool floatx80_is_zero_or_denormal(floatx80 a) 732 { 733 return (a.high & 0x7fff) == 0; 734 } 735 736 static inline bool floatx80_is_any_nan(floatx80 a) 737 { 738 return ((a.high & 0x7fff) == 0x7fff) && (a.low<<1); 739 } 740 741 static inline bool floatx80_eq(floatx80 a, floatx80 b, float_status *s) 742 { 743 return floatx80_compare(a, b, s) == float_relation_equal; 744 } 745 746 static inline bool floatx80_le(floatx80 a, floatx80 b, float_status *s) 747 { 748 return floatx80_compare(a, b, s) <= float_relation_equal; 749 } 750 751 static inline bool floatx80_lt(floatx80 a, floatx80 b, float_status *s) 752 { 753 return floatx80_compare(a, b, s) < float_relation_equal; 754 } 755 756 static inline bool floatx80_unordered(floatx80 a, floatx80 b, float_status *s) 757 { 758 return floatx80_compare(a, b, s) == float_relation_unordered; 759 } 760 761 static inline bool floatx80_eq_quiet(floatx80 a, floatx80 b, float_status *s) 762 { 763 return floatx80_compare_quiet(a, b, s) == float_relation_equal; 764 } 765 766 static inline bool floatx80_le_quiet(floatx80 a, floatx80 b, float_status *s) 767 { 768 return floatx80_compare_quiet(a, b, s) <= float_relation_equal; 769 } 770 771 static inline bool floatx80_lt_quiet(floatx80 a, floatx80 b, float_status *s) 772 { 773 return floatx80_compare_quiet(a, b, s) < float_relation_equal; 774 } 775 776 static inline bool floatx80_unordered_quiet(floatx80 a, floatx80 b, 777 float_status *s) 778 { 779 return floatx80_compare_quiet(a, b, s) == float_relation_unordered; 780 } 781 782 /*---------------------------------------------------------------------------- 783 | Return whether the given value is an invalid floatx80 encoding. 784 | Invalid floatx80 encodings arise when the integer bit is not set, but 785 | the exponent is not zero. The only times the integer bit is permitted to 786 | be zero is in subnormal numbers and the value zero. 787 | This includes what the Intel software developer's manual calls pseudo-NaNs, 788 | pseudo-infinities and un-normal numbers. It does not include 789 | pseudo-denormals, which must still be correctly handled as inputs even 790 | if they are never generated as outputs. 791 *----------------------------------------------------------------------------*/ 792 static inline bool floatx80_invalid_encoding(floatx80 a) 793 { 794 return (a.low & (1ULL << 63)) == 0 && (a.high & 0x7FFF) != 0; 795 } 796 797 #define floatx80_zero make_floatx80(0x0000, 0x0000000000000000LL) 798 #define floatx80_one make_floatx80(0x3fff, 0x8000000000000000LL) 799 #define floatx80_ln2 make_floatx80(0x3ffe, 0xb17217f7d1cf79acLL) 800 #define floatx80_pi make_floatx80(0x4000, 0xc90fdaa22168c235LL) 801 #define floatx80_half make_floatx80(0x3ffe, 0x8000000000000000LL) 802 803 /*---------------------------------------------------------------------------- 804 | Returns the fraction bits of the extended double-precision floating-point 805 | value `a'. 806 *----------------------------------------------------------------------------*/ 807 808 static inline uint64_t extractFloatx80Frac(floatx80 a) 809 { 810 return a.low; 811 } 812 813 /*---------------------------------------------------------------------------- 814 | Returns the exponent bits of the extended double-precision floating-point 815 | value `a'. 816 *----------------------------------------------------------------------------*/ 817 818 static inline int32_t extractFloatx80Exp(floatx80 a) 819 { 820 return a.high & 0x7FFF; 821 } 822 823 /*---------------------------------------------------------------------------- 824 | Returns the sign bit of the extended double-precision floating-point value 825 | `a'. 826 *----------------------------------------------------------------------------*/ 827 828 static inline bool extractFloatx80Sign(floatx80 a) 829 { 830 return a.high >> 15; 831 } 832 833 /*---------------------------------------------------------------------------- 834 | Packs the sign `zSign', exponent `zExp', and significand `zSig' into an 835 | extended double-precision floating-point value, returning the result. 836 *----------------------------------------------------------------------------*/ 837 838 static inline floatx80 packFloatx80(bool zSign, int32_t zExp, uint64_t zSig) 839 { 840 floatx80 z; 841 842 z.low = zSig; 843 z.high = (((uint16_t)zSign) << 15) + zExp; 844 return z; 845 } 846 847 /*---------------------------------------------------------------------------- 848 | Normalizes the subnormal extended double-precision floating-point value 849 | represented by the denormalized significand `aSig'. The normalized exponent 850 | and significand are stored at the locations pointed to by `zExpPtr' and 851 | `zSigPtr', respectively. 852 *----------------------------------------------------------------------------*/ 853 854 void normalizeFloatx80Subnormal(uint64_t aSig, int32_t *zExpPtr, 855 uint64_t *zSigPtr); 856 857 /*---------------------------------------------------------------------------- 858 | Takes two extended double-precision floating-point values `a' and `b', one 859 | of which is a NaN, and returns the appropriate NaN result. If either `a' or 860 | `b' is a signaling NaN, the invalid exception is raised. 861 *----------------------------------------------------------------------------*/ 862 863 floatx80 propagateFloatx80NaN(floatx80 a, floatx80 b, float_status *status); 864 865 /*---------------------------------------------------------------------------- 866 | Takes an abstract floating-point value having sign `zSign', exponent `zExp', 867 | and extended significand formed by the concatenation of `zSig0' and `zSig1', 868 | and returns the proper extended double-precision floating-point value 869 | corresponding to the abstract input. Ordinarily, the abstract value is 870 | rounded and packed into the extended double-precision format, with the 871 | inexact exception raised if the abstract input cannot be represented 872 | exactly. However, if the abstract value is too large, the overflow and 873 | inexact exceptions are raised and an infinity or maximal finite value is 874 | returned. If the abstract value is too small, the input value is rounded to 875 | a subnormal number, and the underflow and inexact exceptions are raised if 876 | the abstract input cannot be represented exactly as a subnormal extended 877 | double-precision floating-point number. 878 | If `roundingPrecision' is 32 or 64, the result is rounded to the same 879 | number of bits as single or double precision, respectively. Otherwise, the 880 | result is rounded to the full precision of the extended double-precision 881 | format. 882 | The input significand must be normalized or smaller. If the input 883 | significand is not normalized, `zExp' must be 0; in that case, the result 884 | returned is a subnormal number, and it must not require rounding. The 885 | handling of underflow and overflow follows the IEC/IEEE Standard for Binary 886 | Floating-Point Arithmetic. 887 *----------------------------------------------------------------------------*/ 888 889 floatx80 roundAndPackFloatx80(int8_t roundingPrecision, bool zSign, 890 int32_t zExp, uint64_t zSig0, uint64_t zSig1, 891 float_status *status); 892 893 /*---------------------------------------------------------------------------- 894 | Takes an abstract floating-point value having sign `zSign', exponent 895 | `zExp', and significand formed by the concatenation of `zSig0' and `zSig1', 896 | and returns the proper extended double-precision floating-point value 897 | corresponding to the abstract input. This routine is just like 898 | `roundAndPackFloatx80' except that the input significand does not have to be 899 | normalized. 900 *----------------------------------------------------------------------------*/ 901 902 floatx80 normalizeRoundAndPackFloatx80(int8_t roundingPrecision, 903 bool zSign, int32_t zExp, 904 uint64_t zSig0, uint64_t zSig1, 905 float_status *status); 906 907 /*---------------------------------------------------------------------------- 908 | The pattern for a default generated extended double-precision NaN. 909 *----------------------------------------------------------------------------*/ 910 floatx80 floatx80_default_nan(float_status *status); 911 912 /*---------------------------------------------------------------------------- 913 | Software IEC/IEEE quadruple-precision conversion routines. 914 *----------------------------------------------------------------------------*/ 915 int32_t float128_to_int32(float128, float_status *status); 916 int32_t float128_to_int32_round_to_zero(float128, float_status *status); 917 int64_t float128_to_int64(float128, float_status *status); 918 int64_t float128_to_int64_round_to_zero(float128, float_status *status); 919 uint64_t float128_to_uint64(float128, float_status *status); 920 uint64_t float128_to_uint64_round_to_zero(float128, float_status *status); 921 uint32_t float128_to_uint32(float128, float_status *status); 922 uint32_t float128_to_uint32_round_to_zero(float128, float_status *status); 923 float32 float128_to_float32(float128, float_status *status); 924 float64 float128_to_float64(float128, float_status *status); 925 floatx80 float128_to_floatx80(float128, float_status *status); 926 927 /*---------------------------------------------------------------------------- 928 | Software IEC/IEEE quadruple-precision operations. 929 *----------------------------------------------------------------------------*/ 930 float128 float128_round_to_int(float128, float_status *status); 931 float128 float128_add(float128, float128, float_status *status); 932 float128 float128_sub(float128, float128, float_status *status); 933 float128 float128_mul(float128, float128, float_status *status); 934 float128 float128_div(float128, float128, float_status *status); 935 float128 float128_rem(float128, float128, float_status *status); 936 float128 float128_sqrt(float128, float_status *status); 937 FloatRelation float128_compare(float128, float128, float_status *status); 938 FloatRelation float128_compare_quiet(float128, float128, float_status *status); 939 bool float128_is_quiet_nan(float128, float_status *status); 940 bool float128_is_signaling_nan(float128, float_status *status); 941 float128 float128_silence_nan(float128, float_status *status); 942 float128 float128_scalbn(float128, int, float_status *status); 943 944 static inline float128 float128_abs(float128 a) 945 { 946 a.high &= 0x7fffffffffffffffLL; 947 return a; 948 } 949 950 static inline float128 float128_chs(float128 a) 951 { 952 a.high ^= 0x8000000000000000LL; 953 return a; 954 } 955 956 static inline bool float128_is_infinity(float128 a) 957 { 958 return (a.high & 0x7fffffffffffffffLL) == 0x7fff000000000000LL && a.low == 0; 959 } 960 961 static inline bool float128_is_neg(float128 a) 962 { 963 return a.high >> 63; 964 } 965 966 static inline bool float128_is_zero(float128 a) 967 { 968 return (a.high & 0x7fffffffffffffffLL) == 0 && a.low == 0; 969 } 970 971 static inline bool float128_is_zero_or_denormal(float128 a) 972 { 973 return (a.high & 0x7fff000000000000LL) == 0; 974 } 975 976 static inline bool float128_is_normal(float128 a) 977 { 978 return (((a.high >> 48) + 1) & 0x7fff) >= 2; 979 } 980 981 static inline bool float128_is_denormal(float128 a) 982 { 983 return float128_is_zero_or_denormal(a) && !float128_is_zero(a); 984 } 985 986 static inline bool float128_is_any_nan(float128 a) 987 { 988 return ((a.high >> 48) & 0x7fff) == 0x7fff && 989 ((a.low != 0) || ((a.high & 0xffffffffffffLL) != 0)); 990 } 991 992 static inline bool float128_eq(float128 a, float128 b, float_status *s) 993 { 994 return float128_compare(a, b, s) == float_relation_equal; 995 } 996 997 static inline bool float128_le(float128 a, float128 b, float_status *s) 998 { 999 return float128_compare(a, b, s) <= float_relation_equal; 1000 } 1001 1002 static inline bool float128_lt(float128 a, float128 b, float_status *s) 1003 { 1004 return float128_compare(a, b, s) < float_relation_equal; 1005 } 1006 1007 static inline bool float128_unordered(float128 a, float128 b, float_status *s) 1008 { 1009 return float128_compare(a, b, s) == float_relation_unordered; 1010 } 1011 1012 static inline bool float128_eq_quiet(float128 a, float128 b, float_status *s) 1013 { 1014 return float128_compare_quiet(a, b, s) == float_relation_equal; 1015 } 1016 1017 static inline bool float128_le_quiet(float128 a, float128 b, float_status *s) 1018 { 1019 return float128_compare_quiet(a, b, s) <= float_relation_equal; 1020 } 1021 1022 static inline bool float128_lt_quiet(float128 a, float128 b, float_status *s) 1023 { 1024 return float128_compare_quiet(a, b, s) < float_relation_equal; 1025 } 1026 1027 static inline bool float128_unordered_quiet(float128 a, float128 b, 1028 float_status *s) 1029 { 1030 return float128_compare_quiet(a, b, s) == float_relation_unordered; 1031 } 1032 1033 #define float128_zero make_float128(0, 0) 1034 1035 /*---------------------------------------------------------------------------- 1036 | The pattern for a default generated quadruple-precision NaN. 1037 *----------------------------------------------------------------------------*/ 1038 float128 float128_default_nan(float_status *status); 1039 1040 #endif /* SOFTFLOAT_H */ 1041