| 1 | ; $Id: expf.asm 98103 2023-01-17 14:15:46Z vboxsync $
|
|---|
| 2 | ;; @file
|
|---|
| 3 | ; IPRT - No-CRT expf - AMD64 & X86.
|
|---|
| 4 | ;
|
|---|
| 5 |
|
|---|
| 6 | ;
|
|---|
| 7 | ; Copyright (C) 2006-2023 Oracle and/or its affiliates.
|
|---|
| 8 | ;
|
|---|
| 9 | ; This file is part of VirtualBox base platform packages, as
|
|---|
| 10 | ; available from https://www.virtualbox.org.
|
|---|
| 11 | ;
|
|---|
| 12 | ; This program is free software; you can redistribute it and/or
|
|---|
| 13 | ; modify it under the terms of the GNU General Public License
|
|---|
| 14 | ; as published by the Free Software Foundation, in version 3 of the
|
|---|
| 15 | ; License.
|
|---|
| 16 | ;
|
|---|
| 17 | ; This program is distributed in the hope that it will be useful, but
|
|---|
| 18 | ; WITHOUT ANY WARRANTY; without even the implied warranty of
|
|---|
| 19 | ; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|---|
| 20 | ; General Public License for more details.
|
|---|
| 21 | ;
|
|---|
| 22 | ; You should have received a copy of the GNU General Public License
|
|---|
| 23 | ; along with this program; if not, see <https://www.gnu.org/licenses>.
|
|---|
| 24 | ;
|
|---|
| 25 | ; The contents of this file may alternatively be used under the terms
|
|---|
| 26 | ; of the Common Development and Distribution License Version 1.0
|
|---|
| 27 | ; (CDDL), a copy of it is provided in the "COPYING.CDDL" file included
|
|---|
| 28 | ; in the VirtualBox distribution, in which case the provisions of the
|
|---|
| 29 | ; CDDL are applicable instead of those of the GPL.
|
|---|
| 30 | ;
|
|---|
| 31 | ; You may elect to license modified versions of this file under the
|
|---|
| 32 | ; terms and conditions of either the GPL or the CDDL or both.
|
|---|
| 33 | ;
|
|---|
| 34 | ; SPDX-License-Identifier: GPL-3.0-only OR CDDL-1.0
|
|---|
| 35 | ;
|
|---|
| 36 |
|
|---|
| 37 |
|
|---|
| 38 | %define RT_ASM_WITH_SEH64
|
|---|
| 39 | %include "iprt/asmdefs.mac"
|
|---|
| 40 | %include "iprt/x86.mac"
|
|---|
| 41 |
|
|---|
| 42 |
|
|---|
| 43 | BEGINCODE
|
|---|
| 44 |
|
|---|
| 45 | extern NAME(RT_NOCRT(feraiseexcept))
|
|---|
| 46 |
|
|---|
| 47 | ;;
|
|---|
| 48 | ; Compute the e (2.7182818...) to the power of rd.
|
|---|
| 49 | ; @returns st(0) / xmm0
|
|---|
| 50 | ; @param rd [xSP + xCB*2] / xmm0
|
|---|
| 51 | RT_NOCRT_BEGINPROC expf
|
|---|
| 52 | push xBP
|
|---|
| 53 | SEH64_PUSH_xBP
|
|---|
| 54 | mov xBP, xSP
|
|---|
| 55 | SEH64_SET_FRAME_xBP 0
|
|---|
| 56 | sub xSP, 20h
|
|---|
| 57 | SEH64_ALLOCATE_STACK 20h
|
|---|
| 58 | SEH64_END_PROLOGUE
|
|---|
| 59 |
|
|---|
| 60 | ;
|
|---|
| 61 | ; Load the input into st0.
|
|---|
| 62 | ;
|
|---|
| 63 | %ifdef RT_ARCH_AMD64
|
|---|
| 64 | movss [xBP - 10h], xmm0
|
|---|
| 65 | fld dword [xBP - 10h]
|
|---|
| 66 | %else
|
|---|
| 67 | fld dword [xBP + xCB*2]
|
|---|
| 68 | %endif
|
|---|
| 69 |
|
|---|
| 70 | ;
|
|---|
| 71 | ; Weed out non-normal values.
|
|---|
| 72 | ;
|
|---|
| 73 | fxam
|
|---|
| 74 | fnstsw ax
|
|---|
| 75 | mov cx, ax
|
|---|
| 76 | and ax, X86_FSW_C3 | X86_FSW_C2 | X86_FSW_C0
|
|---|
| 77 | cmp ax, X86_FSW_C2 ; Normal finite number (excluding zero)
|
|---|
| 78 | je .finite
|
|---|
| 79 | cmp ax, X86_FSW_C3 ; Zero
|
|---|
| 80 | je .zero
|
|---|
| 81 | cmp ax, X86_FSW_C3 | X86_FSW_C2 ; Denormals
|
|---|
| 82 | je .finite
|
|---|
| 83 | cmp ax, X86_FSW_C0 | X86_FSW_C2 ; Infinity.
|
|---|
| 84 | je .inf
|
|---|
| 85 | jmp .nan
|
|---|
| 86 |
|
|---|
| 87 | .finite:
|
|---|
| 88 | ;
|
|---|
| 89 | ; Convert to power of 2 and it'll be the same as exp2.
|
|---|
| 90 | ;
|
|---|
| 91 | fldl2e ; -> st0=log2(e); st1=input
|
|---|
| 92 | fmulp ; -> st0=input*log2(e)
|
|---|
| 93 |
|
|---|
| 94 | ;
|
|---|
| 95 | ; Split the job in two on the fraction and integer input parts.
|
|---|
| 96 | ;
|
|---|
| 97 | fld st0 ; Push a copy of the input on the stack.
|
|---|
| 98 | frndint ; st0 = (int)(input*log2(e))
|
|---|
| 99 | fsub st1, st0 ; st1 = input*log2(e) - (int)input*log2(e); i.e. st1 = fraction, st0 = integer.
|
|---|
| 100 | fxch ; st0 = fraction, st1 = integer.
|
|---|
| 101 |
|
|---|
| 102 | ; 1. Calculate on the fraction.
|
|---|
| 103 | f2xm1 ; st0 = 2**fraction - 1.0
|
|---|
| 104 | fld1
|
|---|
| 105 | faddp ; st0 = 2**fraction
|
|---|
| 106 |
|
|---|
| 107 | ; 2. Apply the integer power of two.
|
|---|
| 108 | fscale ; st0 = result; st1 = integer part of input.
|
|---|
| 109 | fstp st1 ; st0 = result; no st1.
|
|---|
| 110 |
|
|---|
| 111 | ;
|
|---|
| 112 | ; Return st0.
|
|---|
| 113 | ;
|
|---|
| 114 | .return_val:
|
|---|
| 115 | %ifdef RT_ARCH_AMD64
|
|---|
| 116 | fstp dword [xBP - 10h]
|
|---|
| 117 | movss xmm0, [xBP - 10h]
|
|---|
| 118 | %endif
|
|---|
| 119 | .return:
|
|---|
| 120 | leave
|
|---|
| 121 | ret
|
|---|
| 122 |
|
|---|
| 123 | ;
|
|---|
| 124 | ; +/-0.0: Return +1.0
|
|---|
| 125 | ;
|
|---|
| 126 | .zero:
|
|---|
| 127 | ffreep st0
|
|---|
| 128 | fld1
|
|---|
| 129 | jmp .return_val
|
|---|
| 130 |
|
|---|
| 131 | ;
|
|---|
| 132 | ; -Inf: Return +0.0.
|
|---|
| 133 | ; +Inf: Return +Inf. Join path with NaN.
|
|---|
| 134 | ;
|
|---|
| 135 | .inf:
|
|---|
| 136 | test cx, X86_FSW_C1 ; sign bit
|
|---|
| 137 | jz .nan
|
|---|
| 138 | ffreep st0
|
|---|
| 139 | fldz
|
|---|
| 140 | jmp .return_val
|
|---|
| 141 |
|
|---|
| 142 | ;
|
|---|
| 143 | ; NaN: Return the input NaN value as is, if we can.
|
|---|
| 144 | ;
|
|---|
| 145 | .nan:
|
|---|
| 146 | %ifdef RT_ARCH_AMD64
|
|---|
| 147 | ffreep st0
|
|---|
| 148 | %endif
|
|---|
| 149 | jmp .return
|
|---|
| 150 | ENDPROC RT_NOCRT(expf)
|
|---|
| 151 |
|
|---|
