| 1 | /** @file
|
|---|
| 2 | * IPRT / No-CRT - fenv.h, AMD64.
|
|---|
| 3 | */
|
|---|
| 4 |
|
|---|
| 5 | /*
|
|---|
| 6 | * Copyright (C) 2006-2010 Oracle Corporation
|
|---|
| 7 | *
|
|---|
| 8 | * This file is part of VirtualBox Open Source Edition (OSE), as
|
|---|
| 9 | * available from http://www.virtualbox.org. This file is free software;
|
|---|
| 10 | * you can redistribute it and/or modify it under the terms of the GNU
|
|---|
| 11 | * General Public License (GPL) as published by the Free Software
|
|---|
| 12 | * Foundation, in version 2 as it comes in the "COPYING" file of the
|
|---|
| 13 | * VirtualBox OSE distribution. VirtualBox OSE is distributed in the
|
|---|
| 14 | * hope that it will be useful, but WITHOUT ANY WARRANTY of any kind.
|
|---|
| 15 | *
|
|---|
| 16 | * The contents of this file may alternatively be used under the terms
|
|---|
| 17 | * of the Common Development and Distribution License Version 1.0
|
|---|
| 18 | * (CDDL) only, as it comes in the "COPYING.CDDL" file of the
|
|---|
| 19 | * VirtualBox OSE distribution, in which case the provisions of the
|
|---|
| 20 | * CDDL are applicable instead of those of the GPL.
|
|---|
| 21 | *
|
|---|
| 22 | * You may elect to license modified versions of this file under the
|
|---|
| 23 | * terms and conditions of either the GPL or the CDDL or both.
|
|---|
| 24 | * --------------------------------------------------------------------
|
|---|
| 25 | *
|
|---|
| 26 | * This code is based on:
|
|---|
| 27 | *
|
|---|
| 28 | * Copyright (c) 2004-2005 David Schultz <[email protected]>
|
|---|
| 29 | * All rights reserved.
|
|---|
| 30 | *
|
|---|
| 31 | * Redistribution and use in source and binary forms, with or without
|
|---|
| 32 | * modification, are permitted provided that the following conditions
|
|---|
| 33 | * are met:
|
|---|
| 34 | * 1. Redistributions of source code must retain the above copyright
|
|---|
| 35 | * notice, this list of conditions and the following disclaimer.
|
|---|
| 36 | * 2. Redistributions in binary form must reproduce the above copyright
|
|---|
| 37 | * notice, this list of conditions and the following disclaimer in the
|
|---|
| 38 | * documentation and/or other materials provided with the distribution.
|
|---|
| 39 | *
|
|---|
| 40 | * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
|
|---|
| 41 | * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
|---|
| 42 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
|---|
| 43 | * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
|
|---|
| 44 | * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
|---|
| 45 | * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
|
|---|
| 46 | * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
|
|---|
| 47 | * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
|---|
| 48 | * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
|
|---|
| 49 | * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
|
|---|
| 50 | * SUCH DAMAGE.
|
|---|
| 51 | */
|
|---|
| 52 |
|
|---|
| 53 | #ifndef ___iprt_nocrt_amd64_fenv_h
|
|---|
| 54 | #define ___iprt_nocrt_amd64_fenv_h
|
|---|
| 55 |
|
|---|
| 56 | #include <iprt/types.h>
|
|---|
| 57 |
|
|---|
| 58 | typedef struct {
|
|---|
| 59 | struct {
|
|---|
| 60 | uint32_t __control;
|
|---|
| 61 | uint32_t __status;
|
|---|
| 62 | uint32_t __tag;
|
|---|
| 63 | char __other[16];
|
|---|
| 64 | } __x87;
|
|---|
| 65 | uint32_t __mxcsr;
|
|---|
| 66 | } fenv_t;
|
|---|
| 67 |
|
|---|
| 68 | typedef uint16_t fexcept_t;
|
|---|
| 69 |
|
|---|
| 70 | /* Exception flags */
|
|---|
| 71 | #define FE_INVALID 0x01
|
|---|
| 72 | #define FE_DENORMAL 0x02
|
|---|
| 73 | #define FE_DIVBYZERO 0x04
|
|---|
| 74 | #define FE_OVERFLOW 0x08
|
|---|
| 75 | #define FE_UNDERFLOW 0x10
|
|---|
| 76 | #define FE_INEXACT 0x20
|
|---|
| 77 | #define FE_ALL_EXCEPT (FE_DIVBYZERO | FE_DENORMAL | FE_INEXACT | \
|
|---|
| 78 | FE_INVALID | FE_OVERFLOW | FE_UNDERFLOW)
|
|---|
| 79 |
|
|---|
| 80 | /* Rounding modes */
|
|---|
| 81 | #define FE_TONEAREST 0x0000
|
|---|
| 82 | #define FE_DOWNWARD 0x0400
|
|---|
| 83 | #define FE_UPWARD 0x0800
|
|---|
| 84 | #define FE_TOWARDZERO 0x0c00
|
|---|
| 85 | #define _ROUND_MASK (FE_TONEAREST | FE_DOWNWARD | \
|
|---|
| 86 | FE_UPWARD | FE_TOWARDZERO)
|
|---|
| 87 |
|
|---|
| 88 | /*
|
|---|
| 89 | * As compared to the x87 control word, the SSE unit's control word
|
|---|
| 90 | * has the rounding control bits offset by 3 and the exception mask
|
|---|
| 91 | * bits offset by 7.
|
|---|
| 92 | */
|
|---|
| 93 | #define _SSE_ROUND_SHIFT 3
|
|---|
| 94 | #define _SSE_EMASK_SHIFT 7
|
|---|
| 95 |
|
|---|
| 96 | RT_C_DECLS_BEGIN
|
|---|
| 97 |
|
|---|
| 98 | /* Default floating-point environment */
|
|---|
| 99 | extern const fenv_t RT_NOCRT(__fe_dfl_env);
|
|---|
| 100 | #define FE_DFL_ENV (&__fe_dfl_env)
|
|---|
| 101 |
|
|---|
| 102 | #define __fldcw(__cw) __asm __volatile("fldcw %0" : : "m" (__cw))
|
|---|
| 103 | #define __fldenv(__env) __asm __volatile("fldenv %0" : : "m" (__env))
|
|---|
| 104 | #define __fnclex() __asm __volatile("fnclex")
|
|---|
| 105 | #define __fnstenv(__env) __asm __volatile("fnstenv %0" : "=m" (*(__env)))
|
|---|
| 106 | #define __fnstcw(__cw) __asm __volatile("fnstcw %0" : "=m" (*(__cw)))
|
|---|
| 107 | #define __fnstsw(__sw) __asm __volatile("fnstsw %0" : "=am" (*(__sw)))
|
|---|
| 108 | #define __fwait() __asm __volatile("fwait")
|
|---|
| 109 | #define __ldmxcsr(__csr) __asm __volatile("ldmxcsr %0" : : "m" (__csr))
|
|---|
| 110 | #define __stmxcsr(__csr) __asm __volatile("stmxcsr %0" : "=m" (*(__csr)))
|
|---|
| 111 |
|
|---|
| 112 | DECLINLINE(int)
|
|---|
| 113 | feclearexcept(int __excepts)
|
|---|
| 114 | {
|
|---|
| 115 | fenv_t __env;
|
|---|
| 116 |
|
|---|
| 117 | if (__excepts == FE_ALL_EXCEPT) {
|
|---|
| 118 | __fnclex();
|
|---|
| 119 | } else {
|
|---|
| 120 | __fnstenv(&__env.__x87);
|
|---|
| 121 | __env.__x87.__status &= ~__excepts;
|
|---|
| 122 | __fldenv(__env.__x87);
|
|---|
| 123 | }
|
|---|
| 124 | __stmxcsr(&__env.__mxcsr);
|
|---|
| 125 | __env.__mxcsr &= ~__excepts;
|
|---|
| 126 | __ldmxcsr(__env.__mxcsr);
|
|---|
| 127 | return (0);
|
|---|
| 128 | }
|
|---|
| 129 |
|
|---|
| 130 | DECLINLINE(int)
|
|---|
| 131 | fegetexceptflag(fexcept_t *__flagp, int __excepts)
|
|---|
| 132 | {
|
|---|
| 133 | int __mxcsr, __status;
|
|---|
| 134 |
|
|---|
| 135 | __stmxcsr(&__mxcsr);
|
|---|
| 136 | __fnstsw(&__status);
|
|---|
| 137 | *__flagp = (__mxcsr | __status) & __excepts;
|
|---|
| 138 | return (0);
|
|---|
| 139 | }
|
|---|
| 140 |
|
|---|
| 141 | int RT_NOCRT(fesetexceptflag)(const fexcept_t *__flagp, int __excepts);
|
|---|
| 142 | int RT_NOCRT(feraiseexcept)(int __excepts);
|
|---|
| 143 |
|
|---|
| 144 | DECLINLINE(int)
|
|---|
| 145 | fetestexcept(int __excepts)
|
|---|
| 146 | {
|
|---|
| 147 | int __mxcsr, __status;
|
|---|
| 148 |
|
|---|
| 149 | __stmxcsr(&__mxcsr);
|
|---|
| 150 | __fnstsw(&__status);
|
|---|
| 151 | return ((__status | __mxcsr) & __excepts);
|
|---|
| 152 | }
|
|---|
| 153 |
|
|---|
| 154 | DECLINLINE(int)
|
|---|
| 155 | fegetround(void)
|
|---|
| 156 | {
|
|---|
| 157 | int __control;
|
|---|
| 158 |
|
|---|
| 159 | /*
|
|---|
| 160 | * We assume that the x87 and the SSE unit agree on the
|
|---|
| 161 | * rounding mode. Reading the control word on the x87 turns
|
|---|
| 162 | * out to be about 5 times faster than reading it on the SSE
|
|---|
| 163 | * unit on an Opteron 244.
|
|---|
| 164 | */
|
|---|
| 165 | __fnstcw(&__control);
|
|---|
| 166 | return (__control & _ROUND_MASK);
|
|---|
| 167 | }
|
|---|
| 168 |
|
|---|
| 169 | DECLINLINE(int)
|
|---|
| 170 | fesetround(int __round)
|
|---|
| 171 | {
|
|---|
| 172 | int __mxcsr, __control;
|
|---|
| 173 |
|
|---|
| 174 | if (__round & ~_ROUND_MASK)
|
|---|
| 175 | return (-1);
|
|---|
| 176 |
|
|---|
| 177 | __fnstcw(&__control);
|
|---|
| 178 | __control &= ~_ROUND_MASK;
|
|---|
| 179 | __control |= __round;
|
|---|
| 180 | __fldcw(__control);
|
|---|
| 181 |
|
|---|
| 182 | __stmxcsr(&__mxcsr);
|
|---|
| 183 | __mxcsr &= ~(_ROUND_MASK << _SSE_ROUND_SHIFT);
|
|---|
| 184 | __mxcsr |= __round << _SSE_ROUND_SHIFT;
|
|---|
| 185 | __ldmxcsr(__mxcsr);
|
|---|
| 186 |
|
|---|
| 187 | return (0);
|
|---|
| 188 | }
|
|---|
| 189 |
|
|---|
| 190 | int RT_NOCRT(fegetenv)(fenv_t *__envp);
|
|---|
| 191 | int RT_NOCRT(feholdexcept)(fenv_t *__envp);
|
|---|
| 192 |
|
|---|
| 193 | DECLINLINE(int)
|
|---|
| 194 | fesetenv(const fenv_t *__envp)
|
|---|
| 195 | {
|
|---|
| 196 |
|
|---|
| 197 | __fldenv(__envp->__x87);
|
|---|
| 198 | __ldmxcsr(__envp->__mxcsr);
|
|---|
| 199 | return (0);
|
|---|
| 200 | }
|
|---|
| 201 |
|
|---|
| 202 | int RT_NOCRT(feupdateenv)(const fenv_t *__envp);
|
|---|
| 203 | int RT_NOCRT(feenableexcept)(int __mask);
|
|---|
| 204 | int RT_NOCRT(fedisableexcept)(int __mask);
|
|---|
| 205 |
|
|---|
| 206 | DECLINLINE(int)
|
|---|
| 207 | fegetexcept(void)
|
|---|
| 208 | {
|
|---|
| 209 | int __control;
|
|---|
| 210 |
|
|---|
| 211 | /*
|
|---|
| 212 | * We assume that the masks for the x87 and the SSE unit are
|
|---|
| 213 | * the same.
|
|---|
| 214 | */
|
|---|
| 215 | __fnstcw(&__control);
|
|---|
| 216 | return (~__control & FE_ALL_EXCEPT);
|
|---|
| 217 | }
|
|---|
| 218 |
|
|---|
| 219 | RT_C_DECLS_END
|
|---|
| 220 |
|
|---|
| 221 | #ifndef RT_WITHOUT_NOCRT_WRAPPERS
|
|---|
| 222 | # define fesetexceptflag RT_NOCRT(fesetexceptflag)
|
|---|
| 223 | # define feraiseexcept RT_NOCRT(feraiseexcept)
|
|---|
| 224 | # define fegetenv RT_NOCRT(fegetenv)
|
|---|
| 225 | # define feholdexcept RT_NOCRT(feholdexcept)
|
|---|
| 226 | # define feupdateenv RT_NOCRT(feupdateenv)
|
|---|
| 227 | # define feenableexcept RT_NOCRT(feenableexcept)
|
|---|
| 228 | # define fedisableexcept RT_NOCRT(fedisableexcept)
|
|---|
| 229 | # define __fe_dfl_env RT_NOCRT(__fe_dfl_env)
|
|---|
| 230 | #endif
|
|---|
| 231 |
|
|---|
| 232 | #endif /* !__iprt_nocrt_amd64_fenv_h__ */
|
|---|
