Skip to content

R
riscv-gcc-1
Commit a71c0334 by Michael Meissner Committed by Michael Meissner
Options

float128-sed: New files to convert TF names to KF names for PowerPC IEEE 128-bit floating... 

2016-01-21  Michael Meissner  <meissner@linux.vnet.ibm.com>
	    Steven Munroe <munroesj@linux.vnet.ibm.com>
	    Tulio Magno Quites Machado Filho <tulioqm@br.ibm.com>

	* config/rs6000/float128-sed: New files to convert TF names to KF
	names for PowerPC IEEE 128-bit floating point support.
	* config/rs6000/float128-sed-hw: Likewise.

	* config/rs6000/float128-hw.c: New file for ISA 3.0 IEEE 128-bit
	floating point hardware support.

	* config/rs6000/float128-ifunc.c: New file to pick either IEEE
	128-bit floating point software emulation or use ISA 3.0 hardware
	support if it is available.

	* config/rs6000/quad-float128.h: New file to support IEEE 128-bit
	floating point.

	* config/rs6000/extendkftf2-sw.c: New file, convert IEEE 128-bit
	floating point to IBM extended double.

	* config/rs6000/trunctfkf2-sw.c: New file, convert IBM extended
	double to IEEE 128-bit floating point.

	* config/rs6000/t-float128: New Makefile fragments to enable
	building __float128 emulation support.
	* config/rs6000/t-float128-hw: Likewise.

	* config/rs6000/sfp-exceptions.c: New file to provide exception
	support for IEEE 128-bit floating point.

	* config/rs6000/floattikf.c: New files for converting between IEEE
	128-bit floating point and signed/unsigned 128-bit integers.
	* config/rs6000/fixunskfti.c: Likewise.
	* config/rs6000/fixkfti.c: Likewise.
	* config/rs6000/floatuntikf.c: Likewise.

	* config/rs6000/sfp-machine.h (_FP_W_TYPE_SIZE): Use 64-bit types
	when building on 64-bit systems, or when VSX is enabled.
	(_FP_W_TYPE): Likewise.
	(_FP_WS_TYPE): Likewise.
	(_FP_I_TYPE): Likewise.
	(TItype): Define on 64-bit systems.
	(UTItype): Likewise.
	(TI_BITS): Likewise.
	(_FP_MUL_MEAT_D): Add support for using 64-bit types.
	(_FP_MUL_MEAT_Q): Likewise.
	(_FP_DIV_MEAT_D): Likewise.
	(_FP_DIV_MEAT_Q): Likewise.
	(_FP_NANFRAC_D): Likewise.
	(_FP_NANFRAC_Q): Likewise.
	(ISA_BIT): Add exception support if we are being compiled on a
	machine with hardware floating point support to build the IEEE
	128-bit emulation functions.
	(FP_EX_INVALID): Likewise.
	(FP_EX_OVERFLOW): Likewise.
	(FP_EX_UNDERFLOW): Likewise.
	(FP_EX_DIVZERO): Likewise.
	(FP_EX_INEXACT): Likewise.
	(FP_EX_ALL): Likewise.
	(__sfp_handle_exceptions): Likewise.
	(FP_HANDLE_EXCEPTIONS): Likewise.
	(FP_RND_NEAREST): Likewise.
	(FP_RND_ZERO): Likewise.
	(FP_RND_PINF): Likewise.
	(FP_RND_MINF): Likewise.
	(FP_RND_MASK): Likewise.
	(_FP_DECL_EX): Likewise.
	(FP_INIT_ROUNDMODE): Likewise.
	(FP_ROUNDMODE): Likewise.

	* libgcc/config.host (powerpc*-*-linux*): If compiler can compile
	VSX code, enable IEEE 128-bit floating point.  If the compiler can
	compile IEEE 128-bit floating point code with ISA 3.0 IEEE 128-bit
	floating point hardware instructions and it supports declaring
	functions with the ifunc attribute, enable ifunc functions to
	switch between software and hardware support.
	* configure.ac (powerpc*-*-linux*): Likewise.
	* configure: Regenerate.



Co-Authored-By: Steven Munroe <munroesj@linux.vnet.ibm.com>
Co-Authored-By: Tulio Magno Quites Machado Filho <tulioqm@br.ibm.com>

From-SVN: r232685
parent c8d3e088
Showing with 1581 additions and 3 deletions
libgcc/ChangeLog
2016-01-21 Michael Meissner <meissner@linux.vnet.ibm.com>
Steven Munroe <munroesj@linux.vnet.ibm.com>
Tulio Magno Quites Machado Filho <tulioqm@br.ibm.com>
* config/rs6000/float128-sed: New files to convert TF names to KF
names for PowerPC IEEE 128-bit floating point support.
* config/rs6000/float128-sed-hw: Likewise.
* config/rs6000/float128-hw.c: New file for ISA 3.0 IEEE 128-bit
floating point hardware support.
* config/rs6000/float128-ifunc.c: New file to pick either IEEE
128-bit floating point software emulation or use ISA 3.0 hardware
support if it is available.
* config/rs6000/quad-float128.h: New file to support IEEE 128-bit
floating point.
* config/rs6000/extendkftf2-sw.c: New file, convert IEEE 128-bit
floating point to IBM extended double.
* config/rs6000/trunctfkf2-sw.c: New file, convert IBM extended
double to IEEE 128-bit floating point.
* config/rs6000/t-float128: New Makefile fragments to enable
building __float128 emulation support.
* config/rs6000/t-float128-hw: Likewise.
* config/rs6000/sfp-exceptions.c: New file to provide exception
support for IEEE 128-bit floating point.
* config/rs6000/floattikf.c: New files for converting between IEEE
128-bit floating point and signed/unsigned 128-bit integers.
* config/rs6000/fixunskfti.c: Likewise.
* config/rs6000/fixkfti.c: Likewise.
* config/rs6000/floatuntikf.c: Likewise.
* config/rs6000/sfp-machine.h (_FP_W_TYPE_SIZE): Use 64-bit types
when building on 64-bit systems, or when VSX is enabled.
(_FP_W_TYPE): Likewise.
(_FP_WS_TYPE): Likewise.
(_FP_I_TYPE): Likewise.
(TItype): Define on 64-bit systems.
(UTItype): Likewise.
(TI_BITS): Likewise.
(_FP_MUL_MEAT_D): Add support for using 64-bit types.
(_FP_MUL_MEAT_Q): Likewise.
(_FP_DIV_MEAT_D): Likewise.
(_FP_DIV_MEAT_Q): Likewise.
(_FP_NANFRAC_D): Likewise.
(_FP_NANFRAC_Q): Likewise.
(ISA_BIT): Add exception support if we are being compiled on a
machine with hardware floating point support to build the IEEE
128-bit emulation functions.
(FP_EX_INVALID): Likewise.
(FP_EX_OVERFLOW): Likewise.
(FP_EX_UNDERFLOW): Likewise.
(FP_EX_DIVZERO): Likewise.
(FP_EX_INEXACT): Likewise.
(FP_EX_ALL): Likewise.
(__sfp_handle_exceptions): Likewise.
(FP_HANDLE_EXCEPTIONS): Likewise.
(FP_RND_NEAREST): Likewise.
(FP_RND_ZERO): Likewise.
(FP_RND_PINF): Likewise.
(FP_RND_MINF): Likewise.
(FP_RND_MASK): Likewise.
(_FP_DECL_EX): Likewise.
(FP_INIT_ROUNDMODE): Likewise.
(FP_ROUNDMODE): Likewise.
* libgcc/config.host (powerpc*-*-linux*): If compiler can compile
VSX code, enable IEEE 128-bit floating point. If the compiler can
compile IEEE 128-bit floating point code with ISA 3.0 IEEE 128-bit
floating point hardware instructions and it supports declaring
functions with the ifunc attribute, enable ifunc functions to
switch between software and hardware support.
* configure.ac (powerpc*-*-linux*): Likewise.
* configure: Regenerate.
2016-01-15 Nick Clifton <nickc@redhat.com>
* config/msp430/t-msp430 (lib2_mul_none.o): Only use the first
......
libgcc/config.host
......@@ -1063,6 +1063,15 @@ powerpc*-*-linux*)
exit 1
;;
esac
if test $libgcc_cv_powerpc_float128 = yes; then
tmake_file="${tmake_file} rs6000/t-float128"
fi
if test $libgcc_cv_powerpc_float128_hw = yes; then
tmake_file="${tmake_file} rs6000/t-float128-hw"
fi
extra_parts="$extra_parts ecrti.o ecrtn.o ncrti.o ncrtn.o"
md_unwind_header=rs6000/linux-unwind.h
;;
......
libgcc/config/rs6000/extendkftf2-sw.c 0 → 100644
/* Software IEEE 128-bit floating-point emulation for PowerPC.
Copyright (C) 2016 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Michael Meissner (meissner@linux.vnet.ibm.com)
Code is based on the main soft-fp library written by:
Richard Henderson (rth@cygnus.com) and
Jakub Jelinek (jj@ultra.linux.cz).
The GNU C Library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
In addition to the permissions in the GNU Lesser General Public
License, the Free Software Foundation gives you unlimited
permission to link the compiled version of this file into
combinations with other programs, and to distribute those
combinations without any restriction coming from the use of this
file. (The Lesser General Public License restrictions do apply in
other respects; for example, they cover modification of the file,
and distribution when not linked into a combine executable.)
The GNU C Library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with the GNU C Library; if not, see
<http://www.gnu.org/licenses/>. */
/* Convert IEEE 128-bit floating point to IBM long double. */
#ifdef __FLOAT128_HARDWARE__
#error "This module must not be compiled with IEEE 128-bit hardware support"
#endif
#include "soft-fp.h"
#include "quad-float128.h"
IBM128_TYPE
__extendkftf2_sw (__float128 value)
{
IBM128_TYPE ret;
CVT_FLOAT128_TO_IBM128 (ret, value);
return ret;
}
libgcc/config/rs6000/fixkfti.c 0 → 100644
/* Software floating-point emulation, convert IEEE quad to 128bit signed
integer.
Copyright (C) 2016 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Steven Munroe (munroesj@linux.vnet.ibm.com)
Code is based on the main soft-fp library written by:
Uros Bizjak (ubizjak@gmail.com).
The GNU C Library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
In addition to the permissions in the GNU Lesser General Public
License, the Free Software Foundation gives you unlimited
permission to link the compiled version of this file into
combinations with other programs, and to distribute those
combinations without any restriction coming from the use of this
file. (The Lesser General Public License restrictions do apply in
other respects; for example, they cover modification of the file,
and distribution when not linked into a combine executable.)
The GNU C Library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with the GNU C Library; if not, see
<http://www.gnu.org/licenses/>. */
#ifdef _ARCH_PPC64
#include "soft-fp.h"
#include "quad-float128.h"
TItype
__fixkfti (TFtype a)
{
FP_DECL_EX;
FP_DECL_Q (A);
UTItype r;
FP_INIT_EXCEPTIONS;
FP_UNPACK_RAW_Q (A, a);
FP_TO_INT_Q (r, A, TI_BITS, 1);
FP_HANDLE_EXCEPTIONS;
return r;
}
#endif
libgcc/config/rs6000/fixunskfti.c 0 → 100644
/* Software floating-point emulation, convert IEEE quad to 128bit unsigned
integer.
Copyright (C) 2016 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Steven Munroe (munroesj@linux.vnet.ibm.com)
Code is based on the main soft-fp library written by:
Uros Bizjak (ubizjak@gmail.com).
The GNU C Library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
In addition to the permissions in the GNU Lesser General Public
License, the Free Software Foundation gives you unlimited
permission to link the compiled version of this file into
combinations with other programs, and to distribute those
combinations without any restriction coming from the use of this
file. (The Lesser General Public License restrictions do apply in
other respects; for example, they cover modification of the file,
and distribution when not linked into a combine executable.)
The GNU C Library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with the GNU C Library; if not, see
<http://www.gnu.org/licenses/>. */
#ifdef _ARCH_PPC64
#include "soft-fp.h"
#include "quad-float128.h"
UTItype
__fixunskfti (TFtype a)
{
FP_DECL_EX;
FP_DECL_Q (A);
UTItype r;
FP_INIT_EXCEPTIONS;
FP_UNPACK_RAW_Q (A, a);
FP_TO_INT_Q (r, A, TI_BITS, 0);
FP_HANDLE_EXCEPTIONS;
return r;
}
#endif
libgcc/config/rs6000/float128-hw.c 0 → 100644
/* Automatic switching between software and hardware IEEE 128-bit
floating-point emulation for PowerPC.
Copyright (C) 2016 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Michael Meissner (meissner@linux.vnet.ibm.com)
Code is based on the main soft-fp library written by:
Richard Henderson (rth@cygnus.com) and
Jakub Jelinek (jj@ultra.linux.cz).
The GNU C Library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
In addition to the permissions in the GNU Lesser General Public
License, the Free Software Foundation gives you unlimited
permission to link the compiled version of this file into
combinations with other programs, and to distribute those
combinations without any restriction coming from the use of this
file. (The Lesser General Public License restrictions do apply in
other respects; for example, they cover modification of the file,
and distribution when not linked into a combine executable.)
The GNU C Library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with the GNU C Library; if not, see
<http://www.gnu.org/licenses/>. */
#include <soft-fp.h>
#include <quad-float128.h>
#ifndef __FLOAT128_HARDWARE__
#error "This module must be compiled with IEEE 128-bit hardware support"
#endif
TFtype
__addkf3_hw (TFtype a, TFtype b)
{
return a + b;
}
TFtype
__subkf3_hw (TFtype a, TFtype b)
{
return a - b;
}
TFtype
__mulkf3_hw (TFtype a, TFtype b)
{
return a * b;
}
TFtype
__divkf3_hw (TFtype a, TFtype b)
{
return a / b;
}
TFtype
__negkf2_hw (TFtype a)
{
return -a;
}
TFtype
__floatsikf_hw (SItype_ppc a)
{
return (TFtype) a;
}
TFtype
__floatunsikf_hw (USItype_ppc a)
{
return (TFtype) a;
}
TFtype
__floatdikf_hw (DItype_ppc a)
{
return (TFtype) a;
}
TFtype
__floatundikf_hw (UDItype_ppc a)
{
return (TFtype) a;
}
SItype_ppc
__fixkfsi_hw (TFtype a)
{
return (SItype_ppc) a;
}
USItype_ppc
__fixunskfsi_hw (TFtype a)
{
return (USItype_ppc) a;
}
DItype_ppc
__fixkfdi_hw (TFtype a)
{
return (DItype_ppc) a;
}
UDItype_ppc
__fixunskfdi_hw (TFtype a)
{
return (UDItype_ppc) a;
}
TFtype
__extendsfkf2_hw (float a)
{
return (TFtype) a;
}
TFtype
__extenddfkf2_hw (double a)
{
return (TFtype) a;
}
float
__trunckfsf2_hw (TFtype a)
{
return (float) a;
}
double
__trunckfdf2_hw (TFtype a)
{
return (double) a;
}
/* __eqkf2 returns 0 if equal, or 1 if not equal or NaN. */
CMPtype
__eqkf2_hw (TFtype a, TFtype b)
{
return (a != b);
}
/* __gekf2 returns -1 if a < b, 0 if a == b, +1 if a > b, or -2 if NaN. */
CMPtype
__gekf2_hw (TFtype a, TFtype b)
{
if (a < b)
return -1;
else if (__builtin_isunordered (a, b))
return -2;
else if (a == b)
return 0;
return 1;
}
/* __lekf2 returns -1 if a < b, 0 if a == b, +1 if a > b, or +2 if NaN. */
CMPtype
__lekf2_hw (TFtype a, TFtype b)
{
if (a < b)
return -1;
else if (__builtin_isunordered (a, b))
return 2;
else if (a == b)
return 0;
return 1;
}
/* __unordkf2 returns 1 if NaN or 0 otherwise. */
CMPtype
__unordkf2_hw (TFtype a, TFtype b)
{
return (__builtin_isunordered (a, b)) ? 1 : 0;
}
/* Convert __float128 to __ibm128. */
IBM128_TYPE
__extendkftf2_hw (TFtype value)
{
IBM128_TYPE ret;
CVT_FLOAT128_TO_IBM128 (ret, value);
return ret;
}
/* Convert __ibm128 to __float128. */
TFtype
__trunctfkf2_hw (IBM128_TYPE value)
{
TFtype ret;
CVT_IBM128_TO_FLOAT128 (ret, value);
return ret;
}
libgcc/config/rs6000/float128-ifunc.c 0 → 100644
/* Automatic switching between software and hardware IEEE 128-bit
floating-point emulation for PowerPC.
Copyright (C) 2016 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Michael Meissner (meissner@linux.vnet.ibm.com)
Code is based on the main soft-fp library written by:
Richard Henderson (rth@cygnus.com) and
Jakub Jelinek (jj@ultra.linux.cz).
The GNU C Library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
In addition to the permissions in the GNU Lesser General Public
License, the Free Software Foundation gives you unlimited
permission to link the compiled version of this file into
combinations with other programs, and to distribute those
combinations without any restriction coming from the use of this
file. (The Lesser General Public License restrictions do apply in
other respects; for example, they cover modification of the file,
and distribution when not linked into a combine executable.)
The GNU C Library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with the GNU C Library; if not, see
<http://www.gnu.org/licenses/>. */
#include <soft-fp.h>
#include <quad-float128.h>
#include <string.h>
#include <stdlib.h>
#include <ctype.h>
#ifndef FLOAT128_HW_INSNS
#error "float128-ifunc.c needs access to ISA 3.0 instructions and ifunc"
#endif
#ifdef __FLOAT128_HARDWARE__
#error "This module must not be compiled with IEEE 128-bit hardware support"
#endif
#include <sys/auxv.h>
/* Use the namespace clean version of getauxval. However, not all versions of
sys/auxv.h declare it, so declare it here. This code is intended to be
temporary until a suitable version of __builtin_cpu_supports is added that
allows us to tell quickly if the machine supports IEEE 128-bit hardware. */
extern unsigned long __getauxval (unsigned long);
static int
have_ieee_hw_p (void)
{
static int ieee_hw_p = -1;
if (ieee_hw_p < 0)
{
char *p = (char *) __getauxval (AT_PLATFORM);
ieee_hw_p = 0;
/* Don't use atoi/strtol/strncmp/etc. These may require the normal
environment to be setup to set errno to 0, and the ifunc resolvers run
before the whole glibc environment is initialized. */
if (p && p[0] == 'p' && p[1] == 'o' && p[2] == 'w' && p[3] == 'e'
&& p[4] == 'r')
{
long n = 0;
char ch;
p += 5;
while ((ch = *p++) >= '0' && (ch <= '9'))
n = (n * 10) + (ch - '0');
if (n >= 9)
ieee_hw_p = 1;
}
}
return ieee_hw_p;
}
#define SW_OR_HW(SW, HW) (have_ieee_hw_p () ? HW : SW)
/* Resolvers. */
/* We do not provide ifunc resolvers for __fixkfti, __fixunskfti, __floattikf,
and __floatuntikf. There is no ISA 3.0 instruction that converts between
128-bit integer types and 128-bit IEEE floating point, or vice versa. So
use the emulator functions for these conversions. */
static void *__addkf3_resolve (void);
static void *__subkf3_resolve (void);
static void *__mulkf3_resolve (void);
static void *__divkf3_resolve (void);
static void *__negkf2_resolve (void);
static void *__eqkf2_resolve (void);
static void *__nekf2_resolve (void);
static void *__gekf2_resolve (void);
static void *__gtkf2_resolve (void);
static void *__lekf2_resolve (void);
static void *__ltkf2_resolve (void);
static void *__unordkf2_resolve (void);
static void *__extendsfkf2_resolve (void);
static void *__extenddfkf2_resolve (void);
static void *__trunckfsf2_resolve (void);
static void *__trunckfdf2_resolve (void);
static void *__fixkfsi_resolve (void);
static void *__fixkfdi_resolve (void);
static void *__fixunskfsi_resolve (void);
static void *__fixunskfdi_resolve (void);
static void *__floatsikf_resolve (void);
static void *__floatdikf_resolve (void);
static void *__floatunsikf_resolve (void);
static void *__floatundikf_resolve (void);
static void *__extendkftf2_resolve (void);
static void *__trunctfkf2_resolve (void);
static void *
__addkf3_resolve (void)
{
return (void *) SW_OR_HW (__addkf3_sw, __addkf3_hw);
}
static void *
__subkf3_resolve (void)
{
return (void *) SW_OR_HW (__subkf3_sw, __subkf3_hw);
}
static void *
__mulkf3_resolve (void)
{
return (void *) SW_OR_HW (__mulkf3_sw, __mulkf3_hw);
}
static void *
__divkf3_resolve (void)
{
return (void *) SW_OR_HW (__divkf3_sw, __divkf3_hw);
}
static void *
__negkf2_resolve (void)
{
return (void *) SW_OR_HW (__negkf2_sw, __negkf2_hw);
}
static void *
__floatsikf_resolve (void)
{
return (void *) SW_OR_HW (__floatsikf_sw, __floatsikf_hw);
}
static void *
__floatdikf_resolve (void)
{
return (void *) SW_OR_HW (__floatdikf_sw, __floatdikf_hw);
}
static void *
__floatunsikf_resolve (void)
{
return (void *) SW_OR_HW (__floatunsikf_sw, __floatunsikf_hw);
}
static void *
__floatundikf_resolve (void)
{
return (void *) SW_OR_HW (__floatundikf_sw, __floatundikf_hw);
}
static void *
__fixkfsi_resolve (void)
{
return (void *) SW_OR_HW (__fixkfsi_sw, __fixkfsi_hw);
}
static void *
__fixkfdi_resolve (void)
{
return (void *) SW_OR_HW (__fixkfdi_sw, __fixkfdi_hw);
}
static void *
__fixunskfsi_resolve (void)
{
return (void *) SW_OR_HW (__fixunskfsi_sw, __fixunskfsi_hw);
}
static void *
__fixunskfdi_resolve (void)
{
return (void *) SW_OR_HW (__fixunskfdi_sw, __fixunskfdi_hw);
}
static void *
__extendsfkf2_resolve (void)
{
return (void *) SW_OR_HW (__extendsfkf2_sw, __extendsfkf2_hw);
}
static void *
__extenddfkf2_resolve (void)
{
return (void *) SW_OR_HW (__extenddfkf2_sw, __extenddfkf2_hw);
}
static void *
__trunckfsf2_resolve (void)
{
return (void *) SW_OR_HW (__trunckfsf2_sw, __trunckfsf2_hw);
}
static void *
__trunckfdf2_resolve (void)
{
return (void *) SW_OR_HW (__trunckfdf2_sw, __trunckfdf2_hw);
}
static void *
__extendkftf2_resolve (void)
{
return (void *) SW_OR_HW (__extendkftf2_sw, __extendkftf2_hw);
}
static void *
__trunctfkf2_resolve (void)
{
return (void *) SW_OR_HW (__trunctfkf2_sw, __trunctfkf2_hw);
}
static void *
__eqkf2_resolve (void)
{
return (void *) SW_OR_HW (__eqkf2_sw, __eqkf2_hw);
}
static void *
__gekf2_resolve (void)
{
return (void *) SW_OR_HW (__gekf2_sw, __gekf2_hw);
}
static void *
__lekf2_resolve (void)
{
return (void *) SW_OR_HW (__lekf2_sw, __lekf2_hw);
}
static void *
__unordkf2_resolve (void)
{
return (void *) SW_OR_HW (__unordkf2_sw, __unordkf2_hw);
}
/* Resolve __nekf2, __gtkf2, __ltkf2 like __eqkf2, __gekf2, and __lekf2, since
the functions return the same values. */
static void *
__nekf2_resolve (void)
{
return (void *) SW_OR_HW (__eqkf2_sw, __eqkf2_hw);
}
static void *
__gtkf2_resolve (void)
{
return (void *) SW_OR_HW (__gekf2_sw, __gekf2_hw);
}
static void *
__ltkf2_resolve (void)
{
return (void *) SW_OR_HW (__lekf2_sw, __lekf2_hw);
}
/* Ifunc definitions. */
TFtype __addkf3 (TFtype, TFtype)
__attribute__ ((__ifunc__ ("__addkf3_resolve")));
TFtype __subkf3 (TFtype, TFtype)
__attribute__ ((__ifunc__ ("__subkf3_resolve")));
TFtype __mulkf3 (TFtype, TFtype)
__attribute__ ((__ifunc__ ("__mulkf3_resolve")));
TFtype __divkf3 (TFtype, TFtype)
__attribute__ ((__ifunc__ ("__divkf3_resolve")));
TFtype __negkf2 (TFtype)
__attribute__ ((__ifunc__ ("__negkf2_resolve")));
CMPtype __eqkf2 (TFtype, TFtype)
__attribute__ ((__ifunc__ ("__eqkf2_resolve")));
CMPtype __nekf2 (TFtype, TFtype)
__attribute__ ((__ifunc__ ("__nekf2_resolve")));
CMPtype __gekf2 (TFtype, TFtype)
__attribute__ ((__ifunc__ ("__gekf2_resolve")));
CMPtype __gtkf2 (TFtype, TFtype)
__attribute__ ((__ifunc__ ("__gtkf2_resolve")));
CMPtype __lekf2 (TFtype, TFtype)
__attribute__ ((__ifunc__ ("__lekf2_resolve")));
CMPtype __ltkf2 (TFtype, TFtype)
__attribute__ ((__ifunc__ ("__ltkf2_resolve")));
CMPtype __unordkf2 (TFtype, TFtype)
__attribute__ ((__ifunc__ ("__unordkf2_resolve")));
TFtype __extendsfkf2 (float)
__attribute__ ((__ifunc__ ("__extendsfkf2_resolve")));
TFtype __extenddfkf2 (double)
__attribute__ ((__ifunc__ ("__extenddfkf2_resolve")));
float __trunckfsf2 (TFtype)
__attribute__ ((__ifunc__ ("__trunckfsf2_resolve")));
double __trunckfdf2 (TFtype)
__attribute__ ((__ifunc__ ("__trunckfdf2_resolve")));
SItype_ppc __fixkfsi (TFtype)
__attribute__ ((__ifunc__ ("__fixkfsi_resolve")));
DItype_ppc __fixkfdi (TFtype)
__attribute__ ((__ifunc__ ("__fixkfdi_resolve")));
USItype_ppc __fixunskfsi (TFtype)
__attribute__ ((__ifunc__ ("__fixunskfsi_resolve")));
UDItype_ppc __fixunskfdi (TFtype)
__attribute__ ((__ifunc__ ("__fixunskfdi_resolve")));
TFtype __floatsikf (SItype_ppc)
__attribute__ ((__ifunc__ ("__floatsikf_resolve")));
TFtype __floatdikf (DItype_ppc)
__attribute__ ((__ifunc__ ("__floatdikf_resolve")));
TFtype __floatunsikf (USItype_ppc)
__attribute__ ((__ifunc__ ("__floatunsikf_resolve")));
TFtype __floatundikf (UDItype_ppc)
__attribute__ ((__ifunc__ ("__floatundikf_resolve")));
IBM128_TYPE __extendkftf2 (TFtype)
__attribute__ ((__ifunc__ ("__extendkftf2_resolve")));
TFtype __trunctfkf2 (IBM128_TYPE)
__attribute__ ((__ifunc__ ("__trunctfkf2_resolve")));
libgcc/config/rs6000/float128-sed 0 → 100644
s/__addtf3/__addkf3/g
s/__divtf3/__divkf3/g
s/__eqtf2/__eqkf2/g
s/__extenddftf2/__extenddfkf2/g
s/__extendsftf2/__extendsfkf2/g
s/__fixtfdi/__fixkfdi/g
s/__fixtfsi/__fixkfsi/g
s/__fixunstfdi/__fixunskfdi/g
s/__fixunstfsi/__fixunskfsi/g
s/__floatditf/__floatdikf/g
s/__floatsitf/__floatsikf/g
s/__floatunditf/__floatundikf/g
s/__floatunsitf/__floatunsikf/g
s/__getf2/__gekf2/g
s/__gttf2/__gtkf2/g
s/__letf2/__lekf2/g
s/__lttf2/__ltkf2/g
s/__multf3/__mulkf3/g
s/__negtf2/__negkf2/g
s/__netf2/__nekf2/g
s/quad[.]h/quad-float128.h/g
s/__subtf3/__subkf3/g
s/__trunctfdf2/__trunckfdf2/g
s/__trunctfsf2/__trunckfsf2/g
s/__unordtf2/__unordkf2/g
libgcc/config/rs6000/floattikf.c 0 → 100644
/* Software floating-point emulation, convert a 128bit signed integer to IEEE
quad.
Copyright (C) 2016 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Steven Munroe (munroesj@linux.vnet.ibm.com)
Code is based on the main soft-fp library written by:
Uros Bizjak (ubizjak@gmail.com).
The GNU C Library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
In addition to the permissions in the GNU Lesser General Public
License, the Free Software Foundation gives you unlimited
permission to link the compiled version of this file into
combinations with other programs, and to distribute those
combinations without any restriction coming from the use of this
file. (The Lesser General Public License restrictions do apply in
other respects; for example, they cover modification of the file,
and distribution when not linked into a combine executable.)
The GNU C Library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with the GNU C Library; if not, see
<http://www.gnu.org/licenses/>. */
#ifdef _ARCH_PPC64
#include "soft-fp.h"
#include "quad-float128.h"
TFtype
__floattikf (TItype i)
{
FP_DECL_EX;
FP_DECL_Q (A);
TFtype a;
FP_INIT_ROUNDMODE;
FP_FROM_INT_Q (A, i, TI_BITS, UTItype);
FP_PACK_RAW_Q (a, A);
FP_HANDLE_EXCEPTIONS;
return a;
}
#endif
libgcc/config/rs6000/floatuntikf.c 0 → 100644
/* Software floating-point emulation, convert a 128bit unsigned integer to IEEE
quad.
Copyright (C) 2016 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Steven Munroe (munroesj@linux.vnet.ibm.com)
Code is based on the main soft-fp library written by:
Uros Bizjak (ubizjak@gmail.com).
The GNU C Library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
In addition to the permissions in the GNU Lesser General Public
License, the Free Software Foundation gives you unlimited
permission to link the compiled version of this file into
combinations with other programs, and to distribute those
combinations without any restriction coming from the use of this
file. (The Lesser General Public License restrictions do apply in
other respects; for example, they cover modification of the file,
and distribution when not linked into a combine executable.)
The GNU C Library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with the GNU C Library; if not, see
<http://www.gnu.org/licenses/>. */
#ifdef _ARCH_PPC64
#include "soft-fp.h"
#include "quad-float128.h"
TFtype
__floatuntikf (UTItype i)
{
FP_DECL_EX;
FP_DECL_Q (A);
TFtype a;
FP_INIT_ROUNDMODE;
FP_FROM_INT_Q (A, i, TI_BITS, UTItype);
FP_PACK_RAW_Q (a, A);
FP_HANDLE_EXCEPTIONS;
return a;
}
#endif
libgcc/config/rs6000/quad-float128.h 0 → 100644
/* Software floating-point emulation.
Definitions for IEEE Quad Precision on the PowerPC.
Copyright (C) 2016 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Michael Meissner (meissner@linux.vnet.ibm.com).
The GNU C Library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
In addition to the permissions in the GNU Lesser General Public
License, the Free Software Foundation gives you unlimited
permission to link the compiled version of this file into
combinations with other programs, and to distribute those
combinations without any restriction coming from the use of this
file. (The Lesser General Public License restrictions do apply in
other respects; for example, they cover modification of the file,
and distribution when not linked into a combine executable.)
The GNU C Library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with the GNU C Library; if not, see
<http://www.gnu.org/licenses/>. */
/* quad.h defines the TFtype type by:
typedef float TFtype __attribute__ ((mode (TF)));
This define forces it to use KFmode (aka, ieee 128-bit floating point). */
#define TF KF
/* Force the use of the VSX instruction set. */
#if defined(_ARCH_PPC) && (!defined(__VSX__) || !defined(__FLOAT128__))
#pragma GCC target ("vsx,float128")
#endif
#include <quad.h>
#ifdef __LONG_DOUBLE_IEEE128__
#define IBM128_TYPE __ibm128
#else
#define IBM128_TYPE long double
#endif
/* Add prototypes of the library functions created. In case the appropriate
int/long types are not declared in scope by the time quad.h is included,
provide our own version. */
typedef int SItype_ppc __attribute__ ((__mode__ (__SI__)));
typedef int DItype_ppc __attribute__ ((__mode__ (__DI__)));
typedef unsigned USItype_ppc __attribute__ ((__mode__ (__SI__)));
typedef unsigned UDItype_ppc __attribute__ ((__mode__ (__DI__)));
#ifdef _ARCH_PPC64
typedef int TItype_ppc __attribute__ ((__mode__ (__TI__)));
typedef unsigned UTItype_ppc __attribute__ ((__mode__ (__TI__)));
#endif
/* Software emulation functions. */
extern TFtype __addkf3_sw (TFtype, TFtype);
extern TFtype __subkf3_sw (TFtype, TFtype);
extern TFtype __mulkf3_sw (TFtype, TFtype);
extern TFtype __divkf3_sw (TFtype, TFtype);
extern TFtype __negkf2_sw (TFtype);
extern CMPtype __eqkf2_sw (TFtype, TFtype);
extern CMPtype __gekf2_sw (TFtype, TFtype);
extern CMPtype __lekf2_sw (TFtype, TFtype);
extern CMPtype __unordkf2_sw (TFtype, TFtype);
extern TFtype __extendsfkf2_sw (float);
extern TFtype __extenddfkf2_sw (double);
extern float __trunckfsf2_sw (TFtype);
extern double __trunckfdf2_sw (TFtype);
extern SItype_ppc __fixkfsi_sw (TFtype);
extern DItype_ppc __fixkfdi_sw (TFtype);
extern USItype_ppc __fixunskfsi_sw (TFtype);
extern UDItype_ppc __fixunskfdi_sw (TFtype);
extern TFtype __floatsikf_sw (SItype_ppc);
extern TFtype __floatdikf_sw (DItype_ppc);
extern TFtype __floatunsikf_sw (USItype_ppc);
extern TFtype __floatundikf_sw (UDItype_ppc);
extern IBM128_TYPE __extendkftf2_sw (TFtype);
extern TFtype __trunctfkf2_sw (IBM128_TYPE);
#ifdef _ARCH_PPC64
/* We do not provide ifunc resolvers for __fixkfti, __fixunskfti, __floattikf,
and __floatuntikf. There is no ISA 3.0 instruction that converts between
128-bit integer types and 128-bit IEEE floating point, or vice versa. So
use the emulator functions for these conversions. */
extern TItype_ppc __fixkfti (TFtype);
extern UTItype_ppc __fixunskfti (TFtype);
extern TFtype __floattikf (TItype_ppc);
extern TFtype __floatuntikf (UTItype_ppc);
#endif
/* Functions using the ISA 3.0 hardware support. If the code is compiled with
-mcpu=power9, it will not use these functions, but if it was compiled with
-mcpu=power7 or -mcpu=power8 and run on a ISA 3.0 system, it will use the
hardware instruction. */
extern TFtype __addkf3_hw (TFtype, TFtype);
extern TFtype __subkf3_hw (TFtype, TFtype);
extern TFtype __mulkf3_hw (TFtype, TFtype);
extern TFtype __divkf3_hw (TFtype, TFtype);
extern TFtype __negkf2_hw (TFtype);
extern CMPtype __eqkf2_hw (TFtype, TFtype);
extern CMPtype __gekf2_hw (TFtype, TFtype);
extern CMPtype __lekf2_hw (TFtype, TFtype);
extern CMPtype __unordkf2_hw (TFtype, TFtype);
extern TFtype __extendsfkf2_hw (float);
extern TFtype __extenddfkf2_hw (double);
extern float __trunckfsf2_hw (TFtype);
extern double __trunckfdf2_hw (TFtype);
extern SItype_ppc __fixkfsi_hw (TFtype);
extern DItype_ppc __fixkfdi_hw (TFtype);
extern USItype_ppc __fixunskfsi_hw (TFtype);
extern UDItype_ppc __fixunskfdi_hw (TFtype);
extern TFtype __floatsikf_hw (SItype_ppc);
extern TFtype __floatdikf_hw (DItype_ppc);
extern TFtype __floatunsikf_hw (USItype_ppc);
extern TFtype __floatundikf_hw (UDItype_ppc);
extern IBM128_TYPE __extendkftf2_hw (TFtype);
extern TFtype __trunctfkf2_hw (IBM128_TYPE);
/* Ifunc function declarations, to automatically switch between software
emulation and hardware support. */
extern TFtype __addkf3 (TFtype, TFtype);
extern TFtype __subkf3 (TFtype, TFtype);
extern TFtype __mulkf3 (TFtype, TFtype);
extern TFtype __divkf3 (TFtype, TFtype);
extern TFtype __negkf2 (TFtype);
extern CMPtype __eqkf2 (TFtype, TFtype);
extern CMPtype __nekf2 (TFtype, TFtype);
extern CMPtype __gekf2 (TFtype, TFtype);
extern CMPtype __gtkf2 (TFtype, TFtype);
extern CMPtype __lekf2 (TFtype, TFtype);
extern CMPtype __ltkf2 (TFtype, TFtype);
extern CMPtype __unordkf2 (TFtype, TFtype);
extern TFtype __extendsfkf2 (float);
extern TFtype __extenddfkf2 (double);
extern float __trunckfsf2 (TFtype);
extern double __trunckfdf2 (TFtype);
extern SItype_ppc __fixkfsi (TFtype);
extern DItype_ppc __fixkfdi (TFtype);
extern USItype_ppc __fixunskfsi (TFtype);
extern UDItype_ppc __fixunskfdi (TFtype);
extern TFtype __floatsikf (SItype_ppc);
extern TFtype __floatdikf (DItype_ppc);
extern TFtype __floatunsikf (USItype_ppc);
extern TFtype __floatundikf (UDItype_ppc);
extern IBM128_TYPE __extendkftf2 (TFtype);
extern TFtype __trunctfkf2 (IBM128_TYPE);
/* Implementation of conversions between __ibm128 and __float128, to allow the
same code to be used on systems with IEEE 128-bit emulation and with IEEE
128-bit hardware support. */
union ibm128_union {
IBM128_TYPE ibm128;
double dbl[2];
};
#define CVT_FLOAT128_TO_IBM128(RESULT, VALUE) \
{ \
double __high, __low; \
__float128 __value = (VALUE); \
union ibm128_union u; \
\
__high = (double) __value; \
if (__builtin_isnan (__high) || __builtin_isinf (__high)) \
__low = 0.0; \
\
else \
{ \
double __high_temp; \
\
__low = (double) (__value - (__float128) __high); \
/* Renormalize low/high and move them into canonical IBM long \
double form. */ \
__high_temp = __high + __low; \
__low = (__high - __high_temp) + __low; \
__high = __high_temp; \
} \
\
u.dbl[0] = __high; \
u.dbl[1] = __low; \
RESULT = u.ibm128; \
}
#define CVT_IBM128_TO_FLOAT128(RESULT, VALUE) \
{ \
union ibm128_union u; \
double __high, __low; \
\
u.ibm128 = (VALUE); \
__high = u.dbl[0]; \
__low = u.dbl[1]; \
\
/* Handle the special cases of NAN and infinity. */ \
if (__builtin_isnan (__high) || __builtin_isinf (__high)) \
RESULT = (__float128) __high; \
\
/* If low is 0.0, there no need to do the add. In addition, \
avoiding the add produces the correct sign if high is -0.0. */ \
else if (__low == 0.0) \
RESULT = (__float128) __high; \
\
else \
RESULT = ((__float128) __high) + ((__float128) __low); \
}
libgcc/config/rs6000/sfp-exceptions.c 0 → 100644
/* Copyright (C) 2016 Free Software Foundation, Inc.
This file is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by the
Free Software Foundation; either version 3, or (at your option) any
later version.
This file is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
General Public License for more details.
Under Section 7 of GPL version 3, you are granted additional
permissions described in the GCC Runtime Library Exception, version
3.1, as published by the Free Software Foundation.
You should have received a copy of the GNU General Public License and
a copy of the GCC Runtime Library Exception along with this program;
see the files COPYING3 and COPYING.RUNTIME respectively. If not, see
<http://www.gnu.org/licenses/>. */
#include "sfp-machine.h"
/* Only provide exception support if we have hardware floating point and we can
execute the mtfsf instruction. This would only be true if we are using the
emulation routines for IEEE 128-bit floating point on pre-ISA 3.0 machines
without the IEEE 128-bit floating point support. */
#ifndef __NO_FPRS__
void
__sfp_handle_exceptions (int _fex)
{
const double fp_max = __DBL_MAX__;
const double fp_min = __DBL_MIN__;
const double fp_zero = (double) 0.0;
const double fp_one = 1.0;
double tmp;
if (_fex & FP_EX_INVALID)
{
__asm__ __volatile__ ("fdiv %0, %1, %1"
: "=f" (tmp)
: "f" (fp_zero));
}
if (_fex & FP_EX_DIVZERO)
{
__asm__ __volatile__ ("fdiv %0, %1, %2"
: "=f" (tmp)
: "f" (fp_one), "f" (fp_zero));
}
if (_fex & FP_EX_OVERFLOW)
{
__asm__ __volatile__ ("fadd %0, %1, %1"
: "=f" (tmp)
: "f" (fp_max));
}
if (_fex & FP_EX_UNDERFLOW)
{
__asm__ __volatile__ ("fmul %0, %1, %1"
: "=f" (tmp)
: "f" (fp_min));
}
if (_fex & FP_EX_INEXACT)
{
__asm__ __volatile__ ("fsub %0, %1, %2"
: "=f" (tmp)
: "f" (fp_max), "f" (fp_one));
}
}
#endif /* !__NO_FPRS__ */
libgcc/config/rs6000/sfp-machine.h
/* Decide whether to use 64 or 32-bit types to do the emulation. If we are
doing IEEE-128 with VSX, use 64-bit emulation even if we are compiling for a
32-bit target. */
#if defined(_ARCH_PPC64) || defined(__VSX__) || defined(__FLOAT128__)
#define _FP_W_TYPE_SIZE 64
#define _FP_W_TYPE unsigned long long
#define _FP_WS_TYPE signed long long
#define _FP_I_TYPE long long
#ifdef _ARCH_PPC64
typedef int TItype __attribute__ ((mode (TI)));
typedef unsigned int UTItype __attribute__ ((mode (TI)));
#define TI_BITS (__CHAR_BIT__ * (int)sizeof(TItype))
#endif
#else /* 32-bits */
#define _FP_W_TYPE_SIZE 32
#define _FP_W_TYPE unsigned long
#define _FP_WS_TYPE signed long
#define _FP_I_TYPE long
#define _FP_W_TYPE unsigned int
#define _FP_WS_TYPE signed int
#define _FP_I_TYPE int
#endif /* 32-bits */
/* The type of the result of a floating point comparison. This must
match `__libgcc_cmp_return__' in GCC for the target. */
......@@ -10,18 +29,39 @@ typedef int __gcc_CMPtype __attribute__ ((mode (__libgcc_cmp_return__)));
#define _FP_MUL_MEAT_S(R,X,Y) \
_FP_MUL_MEAT_1_wide(_FP_WFRACBITS_S,R,X,Y,umul_ppmm)
#if (_FP_W_TYPE_SIZE==64)
#define _FP_MUL_MEAT_D(R,X,Y) \
_FP_MUL_MEAT_1_wide(_FP_WFRACBITS_D,R,X,Y,umul_ppmm)
#define _FP_MUL_MEAT_Q(R,X,Y) \
_FP_MUL_MEAT_2_wide(_FP_WFRACBITS_Q,R,X,Y,umul_ppmm)
#else
#define _FP_MUL_MEAT_D(R,X,Y) \
_FP_MUL_MEAT_2_wide(_FP_WFRACBITS_D,R,X,Y,umul_ppmm)
#define _FP_MUL_MEAT_Q(R,X,Y) \
_FP_MUL_MEAT_4_wide(_FP_WFRACBITS_Q,R,X,Y,umul_ppmm)
#endif
#define _FP_DIV_MEAT_S(R,X,Y) _FP_DIV_MEAT_1_loop(S,R,X,Y)
#if (_FP_W_TYPE_SIZE==64)
#define _FP_DIV_MEAT_D(R,X,Y) _FP_DIV_MEAT_1_udiv(D,R,X,Y)
#define _FP_DIV_MEAT_Q(R,X,Y) _FP_DIV_MEAT_2_udiv(Q,R,X,Y)
#else
#define _FP_DIV_MEAT_D(R,X,Y) _FP_DIV_MEAT_2_udiv(D,R,X,Y)
#define _FP_DIV_MEAT_Q(R,X,Y) _FP_DIV_MEAT_4_udiv(Q,R,X,Y)
#endif
#define _FP_NANFRAC_S ((_FP_QNANBIT_S << 1) - 1)
#if (_FP_W_TYPE_SIZE==64)
#define _FP_NANFRAC_D ((_FP_QNANBIT_D << 1) - 1)
#define _FP_NANFRAC_Q ((_FP_QNANBIT_Q << 1) - 1), -1
#else
#define _FP_NANFRAC_D ((_FP_QNANBIT_D << 1) - 1), -1
#define _FP_NANFRAC_Q ((_FP_QNANBIT_Q << 1) - 1), -1, -1, -1
#endif
#define _FP_NANSIGN_S 0
#define _FP_NANSIGN_D 0
#define _FP_NANSIGN_Q 0
......@@ -64,6 +104,54 @@ typedef int __gcc_CMPtype __attribute__ ((mode (__libgcc_cmp_return__)));
# endif
#endif
/* Only provide exception support if we have hardware floating point using
floating point registers and we can execute the mtfsf instruction. This
would only be true if we are using the emulation routines for IEEE 128-bit
floating point on pre-ISA 3.0 machines without the IEEE 128-bit floating
point support. */
#ifndef ___NO_FPRS__
#define ISA_BIT(x) (1LL << (63 - x))
/* Use the same bits of the FPSCR. */
# define FP_EX_INVALID ISA_BIT(34)
# define FP_EX_OVERFLOW ISA_BIT(35)
# define FP_EX_UNDERFLOW ISA_BIT(36)
# define FP_EX_DIVZERO ISA_BIT(37)
# define FP_EX_INEXACT ISA_BIT(38)
# define FP_EX_ALL (FP_EX_INVALID | FP_EX_OVERFLOW \
| FP_EX_UNDERFLOW | FP_EX_DIVZERO \
| FP_EX_INEXACT)
void __sfp_handle_exceptions (int);
# define FP_HANDLE_EXCEPTIONS \
do { \
if (__builtin_expect (_fex, 0)) \
__sfp_handle_exceptions (_fex); \
} while (0);
/* A set bit indicates an exception is masked and a clear bit indicates it is
trapping. */
# define FP_TRAPPING_EXCEPTIONS (~_fpscr & (FP_EX_ALL >> 22))
# define FP_RND_NEAREST 0x0
# define FP_RND_ZERO 0x1
# define FP_RND_PINF 0x2
# define FP_RND_MINF 0x3
# define FP_RND_MASK 0x3
# define _FP_DECL_EX \
unsigned long long _fpscr __attribute__ ((unused)) = FP_RND_NEAREST
#define FP_INIT_ROUNDMODE \
do { \
__asm__ __volatile__ ("mtfsf 255, %0" \
: \
: "f" (_fpscr)); \
} while (0)
# define FP_ROUNDMODE (_fpscr & FP_RND_MASK)
#endif /* !__NO_FPRS__ */
/* Define ALIASNAME as a strong alias for NAME. */
# define strong_alias(name, aliasname) _strong_alias(name, aliasname)
......
libgcc/config/rs6000/t-float128 0 → 100644
# Support for adding __float128 to the powerpc.
# The standard 128-bit floating point support functions are TFmode. Most
# PowerPC targets use a long double format that has a pair of doubles to give
# you more precision, but no extra expoenent range. This long double format is
# mostly compatible with the format used by the IBM XL compilers. Some of the
# names used by the IBM double-double format use TF in them, so we rename
# all of the functions provided for the new IEEE 128-bit support.
#
# We use the TF functions in soft-fp for 128-bit floating point support, using
# sed to transform the names in the files from TF names to KF names.
# Emulator functions from the soft-fp directory
fp128_softfp_funcs = addkf3 subkf3 mulkf3 divkf3 negkf2 \
unordkf2 eqkf2 gekf2 lekf2 \
extendsfkf2 extenddfkf2 trunckfsf2 trunckfdf2 \
fixkfsi fixkfdi fixunskfsi fixunskfdi \
floatsikf floatdikf floatunsikf floatundikf
fp128_softfp_src = $(addsuffix -sw.c,$(fp128_softfp_funcs))
fp128_softfp_static_obj = $(addsuffix -sw$(objext),$(fp128_softfp_funcs))
fp128_softfp_shared_obj = $(addsuffix -sw_s$(objext),$(fp128_softfp_funcs))
fp128_softfp_obj = $(fp128_softfp_static_obj) $(fp128_softfp_shared_obj)
# New functions for software emulation
fp128_ppc_funcs = floattikf floatuntikf fixkfti fixunskfti \
extendkftf2-sw trunctfkf2-sw \
sfp-exceptions
fp128_ppc_src = $(addprefix $(srcdir)/config/rs6000/,$(addsuffix \
.c,$(fp128_ppc_funcs)))
fp128_ppc_static_obj = $(addsuffix $(objext),$(fp128_ppc_funcs))
fp128_ppc_shared_obj = $(addsuffix _s$(objext),$(fp128_ppc_funcs))
fp128_ppc_obj = $(fp128_ppc_static_obj) $(fp128_ppc_shared_obj)
# All functions
fp128_funcs = $(fp128_softfp_funcs) $(fp128_ppc_funcs) \
$(fp128_hw_funcs) $(fp128_ifunc_funcs)
fp128_src = $(fp128_softfp_src) $(fp128_ppc_src) \
$(fp128_hw_src) $(fp128_ifunc_src)
fp128_obj = $(fp128_softfp_obj) $(fp128_ppc_obj) \
$(fp128_hw_obj) $(fp128_ifunc_obj)
fp128_sed = $(srcdir)/config/rs6000/float128-sed$(fp128_sed_hw)
fp128_dep = $(fp128_sed) $(srcdir)/config/rs6000/t-float128
fp128_includes = $(srcdir)/soft-fp/double.h \
$(srcdir)/soft-fp/op-1.h \
$(srcdir)/soft-fp/op-4.h \
$(srcdir)/soft-fp/op-common.h \
$(srcdir)/soft-fp/single.h \
$(srcdir)/soft-fp/extended.h \
$(srcdir)/soft-fp/op-2.h \
$(srcdir)/soft-fp/op-8.h \
$(srcdir)/soft-fp/quad.h \
$(srcdir)/soft-fp/soft-fp.h
# Build the emulator without ISA 3.0 hardware support.
FP128_CFLAGS_SW = -Wno-type-limits -mvsx -mfloat128 \
-mno-float128-hardware \
-I$(srcdir)/soft-fp \
-I$(srcdir)/config/rs6000 \
$(FLOAT128_HW_INSNS)
$(fp128_softfp_obj) : INTERNAL_CFLAGS += $(FP128_CFLAGS_SW)
$(fp128_ppc_obj) : INTERNAL_CFLAGS += $(FP128_CFLAGS_SW)
$(fp128_obj) : $(fp128_includes)
$(fp128_obj) : $(srcdir)/config/rs6000/quad-float128.h
$(fp128_softfp_src) : $(srcdir)/soft-fp/$(subst -sw,,$(subst kf,tf,$@)) $(fp128_dep)
@src="$(srcdir)/soft-fp/$(subst -sw,,$(subst kf,tf,$@))"; \
echo "Create $@"; \
(echo "/* file created from $$src */"; \
echo; \
sed -f $(fp128_sed) < $$src) > $@
.PHONY: test clean-float128
test:
@echo "fp128_src:"; \
for x in $(fp128_src); do echo " $$x"; done; \
echo; \
echo "fp128_obj:"; \
for x in $(fp128_obj); do echo " $$x"; done;
clean-float128:
rm -rf $(fp128_softfp_src)
@$(MULTICLEAN) multi-clean DO=clean-float128
# For now, only put it in the static library
# LIB2ADD += $(fp128_src)
LIB2ADD_ST += $(fp128_src)
libgcc/config/rs6000/t-float128-hw 0 → 100644
# Support for adding __float128 hardware support to the powerpc.
# Tell the float128 functions that the ISA 3.0 hardware support can
# be compiled it to be selected via IFUNC functions.
FLOAT128_HW_INSNS = -DFLOAT128_HW_INSNS
# New functions for hardware support
fp128_hw_funcs = float128-hw
fp128_hw_src = $(srcdir)/config/rs6000/float128-hw.c
fp128_hw_static_obj = float128-hw$(objext)
fp128_hw_shared_obj = float128-hw_s$(objext)
fp128_hw_obj = $(fp128_hw_static_obj) $(fp128_hw_shared_obj)
fp128_ifunc_funcs = float128-ifunc
fp128_ifunc_src = $(srcdir)/config/rs6000/float128-ifunc.c
fp128_ifunc_static_obj = float128-ifunc$(objext)
fp128_ifunc_shared_obj = float128-ifunc_s$(objext)
fp128_ifunc_obj = $(fp128_ifunc_static_obj) $(fp128_ifunc_shared_obj)
# Add _sw suffix to names to allow ifunc support.
fp128_sed_hw = -hw
# Build the hardware support functions with appropriate hardware support
FP128_CFLAGS_HW = -Wno-type-limits -mvsx -mfloat128 \
-mpower8-vector -mpower9-vector \
-mfloat128-hardware \
-I$(srcdir)/soft-fp \
-I$(srcdir)/config/rs6000 \
$(FLOAT128_HW_INSNS)
$(fp128_hw_obj) : INTERNAL_CFLAGS += $(FP128_CFLAGS_HW)
$(fp128_hw_obj) : $(srcdir)/config/rs6000/t-float128-hw
$(fp128_ifunc_obj) : INTERNAL_CFLAGS += $(FP128_CFLAGS_SW)
$(fp128_ifunc_obj) : $(srcdir)/config/rs6000/t-float128-hw
libgcc/config/rs6000/trunctfkf2-sw.c 0 → 100644
/* Software IEEE 128-bit floating-point emulation for PowerPC.
Copyright (C) 2016 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Michael Meissner (meissner@linux.vnet.ibm.com)
Code is based on the main soft-fp library written by:
Richard Henderson (rth@cygnus.com) and
Jakub Jelinek (jj@ultra.linux.cz).
The GNU C Library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
In addition to the permissions in the GNU Lesser General Public
License, the Free Software Foundation gives you unlimited
permission to link the compiled version of this file into
combinations with other programs, and to distribute those
combinations without any restriction coming from the use of this
file. (The Lesser General Public License restrictions do apply in
other respects; for example, they cover modification of the file,
and distribution when not linked into a combine executable.)
The GNU C Library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with the GNU C Library; if not, see
<http://www.gnu.org/licenses/>. */
/* Convert IBM long double to IEEE 128-bit floating point. */
#ifdef __FLOAT128_HARDWARE__
#error "This module must not be compiled with IEEE 128-bit hardware support"
#endif
#include "soft-fp.h"
#include "quad-float128.h"
#ifndef FLOAT128_HW_INSNS
#define __trunctfkf2_sw __trunctfkf2
#endif
__float128
__trunctfkf2_sw (IBM128_TYPE value)
{
__float128 ret;
CVT_IBM128_TO_FLOAT128 (ret, value);
return ret;
}
libgcc/configure
......@@ -4766,6 +4766,61 @@ esac
;;
esac
case ${host} in
powerpc*-*-linux*)
{ $as_echo "$as_me:${as_lineno-$LINENO}: checking whether the PowerPC compiler can do __float128" >&5
$as_echo_n "checking whether the PowerPC compiler can do __float128... " >&6; }
if test "${libgcc_cv_powerpc_float128+set}" = set; then :
$as_echo_n "(cached) " >&6
else
cat confdefs.h - <<_ACEOF >conftest.$ac_ext
/* end confdefs.h. */
#pragma GCC target ("vsx,float128")
__float128 add (__float128 *a) { return *a + *(a+1); }
_ACEOF
if ac_fn_c_try_compile "$LINENO"; then :
libgcc_cv_powerpc_float128=yes
else
libgcc_cv_powerpc_float128=no
fi
rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
fi
{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $libgcc_cv_powerpc_float128" >&5
$as_echo "$libgcc_cv_powerpc_float128" >&6; }
{ $as_echo "$as_me:${as_lineno-$LINENO}: checking whether the PowerPC compiler can do hardware __float128" >&5
$as_echo_n "checking whether the PowerPC compiler can do hardware __float128... " >&6; }
if test "${libgcc_cv_powerpc_float128_hw+set}" = set; then :
$as_echo_n "(cached) " >&6
else
cat confdefs.h - <<_ACEOF >conftest.$ac_ext
/* end confdefs.h. */
#pragma GCC target ("cpu=power9,float128,float128-hardware")
#include <sys/auxv.h>
#ifndef AT_PLATFORM
#error "AT_PLATFORM is not defined"
#endif
__float128 add (__float128 a, __float128 b)
{
__float128 ret;
__asm__ ("xsaddqp %0,%1,%2" : "=v" (ret) : "v" (a), "v" (b));
return ret;
}
void *add_resolver (void) { return (void *) add; }
__float128 add_ifunc (__float128, __float128)
__attribute__ ((__ifunc__ ("add_resolver")));
_ACEOF
if ac_fn_c_try_compile "$LINENO"; then :
libgcc_cv_powerpc_float128_hw=yes
else
libgcc_cv_powerpc_float128_hw=no
fi
rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
fi
{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $libgcc_cv_powerpc_float128_hw" >&5
$as_echo "$libgcc_cv_powerpc_float128_hw" >&6; }
esac
# Collect host-machine-specific information.
. ${srcdir}/config.host
......
libgcc/configure.ac
......@@ -373,6 +373,37 @@ esac
;;
esac
case ${host} in
powerpc*-*-linux*)
AC_CACHE_CHECK([whether the PowerPC compiler can do __float128],
[libgcc_cv_powerpc_float128],
[AC_COMPILE_IFELSE(
[#pragma GCC target ("vsx,float128")
__float128 add (__float128 *a) { return *a + *(a+1); }],
[libgcc_cv_powerpc_float128=yes],
[libgcc_cv_powerpc_float128=no])])
AC_CACHE_CHECK([whether the PowerPC compiler can do hardware __float128],
[libgcc_cv_powerpc_float128_hw],
[AC_COMPILE_IFELSE(
[#pragma GCC target ("cpu=power9,float128,float128-hardware")
#include <sys/auxv.h>
#ifndef AT_PLATFORM
#error "AT_PLATFORM is not defined"
#endif
__float128 add (__float128 a, __float128 b)
{
__float128 ret;
__asm__ ("xsaddqp %0,%1,%2" : "=v" (ret) : "v" (a), "v" (b));
return ret;
}
void *add_resolver (void) { return (void *) add; }
__float128 add_ifunc (__float128, __float128)
__attribute__ ((__ifunc__ ("add_resolver")));],
[libgcc_cv_powerpc_float128_hw=yes],
[libgcc_cv_powerpc_float128_hw=no])])
esac
# Collect host-machine-specific information.
. ${srcdir}/config.host
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment