glibc/math/math_private.h
Ulrich Drepper 650425ceb4 Update.
1997-09-25 00:23  Ulrich Drepper  <drepper@cygnus.com>

	* elf/dl-profile.c: Correct implementation.

	* io/ftwtest-sh: Don't depend on TMPDIR == /tmp.

	* locale/setlocale.c: Rewrite a bit for more clarity.

	* math/Makefile (libm-calls): Add w_exp2.
	* math/math_private.h: Add prototypes for __ieee754_exp2{,f,l}.
	* sysdeps/libm-i387/s_exp2.S: Change name to __ieee754_exp2.
	* sysdeps/libm-i387/s_exp2f.S: Likewise.
	* sysdeps/libm-i387/s_exp2l.S: Likewise.
	* sysdeps/libm-ieee754/k_standard.c: Add error cases for exp2.

	* string/bits/string2.h (__strcpy_small): Optimize.
	(__stpcpy_small): Likewise.
	(strncpy): Use variable for dest argument since it's used more than
	once.
	(strncat): Likewise.
	(strcmp): Add optimization for this function.
	* sysdeps/i386/i486/string.h (strlen): Correctly use __builtin_strlen.
	(__strcpy_small): Optimize.
	(__stpcpy_small): Likewise.
	(__stpcpy_c): Correctly use __mempcpy_* macros.
	(__mempcpy_by2, __mempcpy_by4, __mempcpy_byn): Return pointer to
	byte following last copied.
	(strncat): Use variable for dest argument since it's used more than
	once.
	(strcmp): Add optimization for this function.

	* sysdeps/i386/dl-machine.h (ELF_MACHINE_RUNTIME_TRAMPOLINE): Use
	.text and .previous to select correct section.

1997-09-23 19:56  Philip Blundell  <Philip.Blundell@pobox.com>

	* configure.in: Add `--disable-versioning' option to suppress the
	use of symbol versions even if binutils claims to support it.

1997-09-24 20:10  Philip Blundell  <Philip.Blundell@pobox.com>

	* csu/Makefile (before-compile): Don't try to build abi-tags.h if
	not using ELF.

1997-09-21  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* elf/sofini.c [HAVE_DWARF2_UNWIND_INFO]: Supply termination for
	the frame unwind info section.

	* elf/soinit.c [HAVE_DWARF2_UNWIND_INFO]: Register and unregister
	the frame unwind info.

	* config.h.in (HAVE_DWARF2_UNWIND_INFO): Add #undef.

	* configure.in: Check whether gcc supports DWARF2 unwind info.

	* libc.map: Export frame handling tables.

1997-09-10 06:56  Geoff Keating  <geoffk@ozemail.com.au>

	* sysdeps/unix/sysv/linux/powerpc/bits/kernel_termios.h:
	Use the size of the kernel's termios structure for ioctls.

	* sysdeps/powerpc/dl-machine.h: Prepare for library profiling.

	* sysdeps/powerpc/bits/mathinline.h: Add slightly slower versions
	of the C9X FP comparison macros. Delete 'fabs' and 'sqrt' inline
	routines, because gcc has them as internals.

	* sysdeps/powerpc/Makefile (pic-ccflags): Define this instead of
	CFLAGS-.os.

	* sysdeps/powerpc/bzero.S: New file.
	* sysdeps/powerpc/strcat.c: New file.
	* sysdeps/powerpc/strcpy.S: New file.
	* sysdeps/powerpc/stpcpy.S: New file.

	* math/Makefile: Add atest_exp2, test-reduce.
	* math/atest_exp2.c: New file.
	* math/test-reduce.c: New file.
	* sysdeps/libm-ieee754/Dist: New file.
	* sysdeps/libm-ieee754/s_exp2.c: New file.
	* sysdeps/libm-ieee754/s_exp2f.c: New file.
	* sysdeps/libm-ieee754/t_exp2.h: New file.
	* sysdeps/libm-ieee754/t_exp2f.h: New file.
	* math/libm-test.c (exp2_test): Add some more tests.

	* Rules: Use empty.os instead of empty.o, since it gets linked into
	libc.so...
	* configure.in: Add --disable-static to disable building .a files.
	* config.make.in: Substitute the new variable.
	* Makeconfig: Don't build .o files if not building .a files.

	* elf/dl-runtime.c (fixup): Factor out call to elf_machine_relplt.
	(profile_fixup): Likewise.

1997-09-23  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* posix/globtest.c (main): Change to directory passed as first
	argument.

	* posix/globtest.sh: Don't cd before running the program, instead
	pass testdir as argument, so that $common_objpfx remains valid.

1997-09-23 18:01  Thorsten Kukuk  <kukuk@vt.uni-paderborn.de>

	* nis/nis_call.c (__do_niscall2): Fix multiple Server support.

	* nis/nis_findserver.c: Save latest working sockaddr_in.

1997-09-17 22:07  Zack Weinberg  <zack@rabi.phys.columbia.edu>

	* configure.in: Automatically determine whether as and ld are the
	GNU versions.
	(options):  --with-gnu-{as,ld,binutils} replaced by
	single option --with-binutils=PATH specifying a -B option to gcc.
	* aclocal.m4: Two new macros defined, LIBC_PROG_FOO_GNU and
	LIBC_PROG_BINUTILS.

	* configure.in: Allow the user to force configuration for
	unsupported platforms with an undocumented option.

1997-09-22 16:55  Thorsten Kukuk  <kukuk@vt.uni-paderborn.de>

	* sunrpc/get_myaddr.c (get_myaddress): Avoid loopback interfaces,
	return loopback address only if there is no other interface.

	* nis/Makefile: Add nis_callback libnsl-routines.

	* nis/nis_add.c: Use new __do_niscall* interface.
	* nis/nis_cache.c: Likewise.
	* nis/nis_checkpoint.c: Likewise.
	* nis/nis_lookup.c: Likewise.
	* nis/nis_mkdir.c: Likewise.
	* nis/nis_modify.c: Likewise.
	* nis/nis_ping.c: Likewise.
	* nis/nis_remove.c: Likewise.
	* nis/nis_rmdir.c: Likewise.
	* nis/nis_server.c: Likewise.
	* nis/nis_util.c: Likewise.

	* nis/rpcsvc/nis.h: Make C++ safe.

	* nis/nss_nisplus/nisplus-publickey.c (getsecretkey): Fix use
	of variables.

	* nis/nis_findserv.c: Make thread safe.

	* nis/nis_call.c: Add support for callback, Fix use of variables.
	* nis/nis_table.c: Add support for callback, FOLLOW_PATH and
	ALL_RESULTS.
	* nis/nis_callback.c: New, callback functions.
	* nis/nis_intern.h: Add callback declarations.

1997-09-20  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* string/bits/string2.h: Fix logic in preprocessor directive.
	(__strsep_1c, __strsep_g): Don't declare __retval as pointing to
	const, to save a cast and a possible warning.

1997-09-22 04:12  Ulrich Drepper  <drepper@cygnus.com>

	* sysdeps/i386/i486/atomicity.h (exchange_and_add): Really address
	memory in xadd not %1.

1997-09-21 13:56  Ulrich Drepper  <drepper@cygnus.com>

	* manual/maint.texi (Supported Configurations): Add SPARC64 to
	list of supported platforms.
1997-09-24 23:03:42 +00:00

361 lines
11 KiB
C

/*
* ====================================================
* Copyright (C) 1993 by Sun Microsystems, Inc. All rights reserved.
*
* Developed at SunPro, a Sun Microsystems, Inc. business.
* Permission to use, copy, modify, and distribute this
* software is freely granted, provided that this notice
* is preserved.
* ====================================================
*/
/*
* from: @(#)fdlibm.h 5.1 93/09/24
* $Id$
*/
#ifndef _MATH_PRIVATE_H_
#define _MATH_PRIVATE_H_
#include <machine/endian.h>
#include <sys/types.h>
/* The original fdlibm code used statements like:
n0 = ((*(int*)&one)>>29)^1; * index of high word *
ix0 = *(n0+(int*)&x); * high word of x *
ix1 = *((1-n0)+(int*)&x); * low word of x *
to dig two 32 bit words out of the 64 bit IEEE floating point
value. That is non-ANSI, and, moreover, the gcc instruction
scheduler gets it wrong. We instead use the following macros.
Unlike the original code, we determine the endianness at compile
time, not at run time; I don't see much benefit to selecting
endianness at run time. */
/* A union which permits us to convert between a double and two 32 bit
ints. */
#if BYTE_ORDER == BIG_ENDIAN
typedef union
{
double value;
struct
{
u_int32_t msw;
u_int32_t lsw;
} parts;
} ieee_double_shape_type;
#endif
#if BYTE_ORDER == LITTLE_ENDIAN
typedef union
{
double value;
struct
{
u_int32_t lsw;
u_int32_t msw;
} parts;
} ieee_double_shape_type;
#endif
/* Get two 32 bit ints from a double. */
#define EXTRACT_WORDS(ix0,ix1,d) \
do { \
ieee_double_shape_type ew_u; \
ew_u.value = (d); \
(ix0) = ew_u.parts.msw; \
(ix1) = ew_u.parts.lsw; \
} while (0)
/* Get the more significant 32 bit int from a double. */
#define GET_HIGH_WORD(i,d) \
do { \
ieee_double_shape_type gh_u; \
gh_u.value = (d); \
(i) = gh_u.parts.msw; \
} while (0)
/* Get the less significant 32 bit int from a double. */
#define GET_LOW_WORD(i,d) \
do { \
ieee_double_shape_type gl_u; \
gl_u.value = (d); \
(i) = gl_u.parts.lsw; \
} while (0)
/* Set a double from two 32 bit ints. */
#define INSERT_WORDS(d,ix0,ix1) \
do { \
ieee_double_shape_type iw_u; \
iw_u.parts.msw = (ix0); \
iw_u.parts.lsw = (ix1); \
(d) = iw_u.value; \
} while (0)
/* Set the more significant 32 bits of a double from an int. */
#define SET_HIGH_WORD(d,v) \
do { \
ieee_double_shape_type sh_u; \
sh_u.value = (d); \
sh_u.parts.msw = (v); \
(d) = sh_u.value; \
} while (0)
/* Set the less significant 32 bits of a double from an int. */
#define SET_LOW_WORD(d,v) \
do { \
ieee_double_shape_type sl_u; \
sl_u.value = (d); \
sl_u.parts.lsw = (v); \
(d) = sl_u.value; \
} while (0)
/* A union which permits us to convert between a float and a 32 bit
int. */
typedef union
{
float value;
u_int32_t word;
} ieee_float_shape_type;
/* Get a 32 bit int from a float. */
#define GET_FLOAT_WORD(i,d) \
do { \
ieee_float_shape_type gf_u; \
gf_u.value = (d); \
(i) = gf_u.word; \
} while (0)
/* Set a float from a 32 bit int. */
#define SET_FLOAT_WORD(d,i) \
do { \
ieee_float_shape_type sf_u; \
sf_u.word = (i); \
(d) = sf_u.value; \
} while (0)
/* A union which permits us to convert between a long double and
three 32 bit ints. */
#if BYTE_ORDER == BIG_ENDIAN
typedef union
{
long double value;
struct
{
unsigned int sign_exponent:16;
unsigned int empty:16;
u_int32_t msw;
u_int32_t lsw;
} parts;
} ieee_long_double_shape_type;
#endif
#if BYTE_ORDER == LITTLE_ENDIAN
typedef union
{
long double value;
struct
{
u_int32_t lsw;
u_int32_t msw;
unsigned int sign_exponent:16;
unsigned int empty:16;
} parts;
} ieee_long_double_shape_type;
#endif
/* Get three 32 bit ints from a double. */
#define GET_LDOUBLE_WORDS(exp,ix0,ix1,d) \
do { \
ieee_long_double_shape_type ew_u; \
ew_u.value = (d); \
(exp) = ew_u.parts.sign_exponent; \
(ix0) = ew_u.parts.msw; \
(ix1) = ew_u.parts.lsw; \
} while (0)
/* Set a double from two 32 bit ints. */
#define SET_LDOUBLE_WORDS(d,exp,ix0,ix1) \
do { \
ieee_long_double_shape_type iw_u; \
iw_u.parts.sign_exponent = (exp); \
iw_u.parts.msw = (ix0); \
iw_u.parts.lsw = (ix1); \
(d) = iw_u.value; \
} while (0)
/* Get the more significant 32 bits of a long double mantissa. */
#define GET_LDOUBLE_MSW(v,d) \
do { \
ieee_long_double_shape_type sh_u; \
sh_u.value = (d); \
(v) = sh_u.parts.msw; \
} while (0)
/* Set the more significant 32 bits of a long double mantissa from an int. */
#define SET_LDOUBLE_MSW(d,v) \
do { \
ieee_long_double_shape_type sh_u; \
sh_u.value = (d); \
sh_u.parts.msw = (v); \
(d) = sh_u.value; \
} while (0)
/* Get int from the exponent of a long double. */
#define GET_LDOUBLE_EXP(exp,d) \
do { \
ieee_long_double_shape_type ge_u; \
ge_u.value = (d); \
(exp) = ge_u.parts.sign_exponent; \
} while (0)
/* Set exponent of a long double from an int. */
#define SET_LDOUBLE_EXP(d,exp) \
do { \
ieee_long_double_shape_type se_u; \
se_u.value = (d); \
se_u.parts.sign_exponent = (exp); \
(d) = se_u.value; \
} while (0)
/* ieee style elementary functions */
extern double __ieee754_sqrt __P((double));
extern double __ieee754_acos __P((double));
extern double __ieee754_acosh __P((double));
extern double __ieee754_log __P((double));
extern double __ieee754_atanh __P((double));
extern double __ieee754_asin __P((double));
extern double __ieee754_atan2 __P((double,double));
extern double __ieee754_exp __P((double));
extern double __ieee754_exp2 __P((double));
extern double __ieee754_cosh __P((double));
extern double __ieee754_fmod __P((double,double));
extern double __ieee754_pow __P((double,double));
extern double __ieee754_lgamma_r __P((double,int *));
extern double __ieee754_gamma_r __P((double,int *));
extern double __ieee754_lgamma __P((double));
extern double __ieee754_gamma __P((double));
extern double __ieee754_log10 __P((double));
extern double __ieee754_sinh __P((double));
extern double __ieee754_hypot __P((double,double));
extern double __ieee754_j0 __P((double));
extern double __ieee754_j1 __P((double));
extern double __ieee754_y0 __P((double));
extern double __ieee754_y1 __P((double));
extern double __ieee754_jn __P((int,double));
extern double __ieee754_yn __P((int,double));
extern double __ieee754_remainder __P((double,double));
extern int32_t __ieee754_rem_pio2 __P((double,double*));
extern double __ieee754_scalb __P((double,double));
/* fdlibm kernel function */
extern double __kernel_standard __P((double,double,int));
extern double __kernel_sin __P((double,double,int));
extern double __kernel_cos __P((double,double));
extern double __kernel_tan __P((double,double,int));
extern int __kernel_rem_pio2 __P((double*,double*,int,int,int,
const int32_t*));
/* ieee style elementary float functions */
extern float __ieee754_sqrtf __P((float));
extern float __ieee754_acosf __P((float));
extern float __ieee754_acoshf __P((float));
extern float __ieee754_logf __P((float));
extern float __ieee754_atanhf __P((float));
extern float __ieee754_asinf __P((float));
extern float __ieee754_atan2f __P((float,float));
extern float __ieee754_expf __P((float));
extern float __ieee754_exp2f __P((float));
extern float __ieee754_coshf __P((float));
extern float __ieee754_fmodf __P((float,float));
extern float __ieee754_powf __P((float,float));
extern float __ieee754_lgammaf_r __P((float,int *));
extern float __ieee754_gammaf_r __P((float,int *));
extern float __ieee754_lgammaf __P((float));
extern float __ieee754_gammaf __P((float));
extern float __ieee754_log10f __P((float));
extern float __ieee754_sinhf __P((float));
extern float __ieee754_hypotf __P((float,float));
extern float __ieee754_j0f __P((float));
extern float __ieee754_j1f __P((float));
extern float __ieee754_y0f __P((float));
extern float __ieee754_y1f __P((float));
extern float __ieee754_jnf __P((int,float));
extern float __ieee754_ynf __P((int,float));
extern float __ieee754_remainderf __P((float,float));
extern int32_t __ieee754_rem_pio2f __P((float,float*));
extern float __ieee754_scalbf __P((float,float));
/* float versions of fdlibm kernel functions */
extern float __kernel_sinf __P((float,float,int));
extern float __kernel_cosf __P((float,float));
extern float __kernel_tanf __P((float,float,int));
extern int __kernel_rem_pio2f __P((float*,float*,int,int,int,
const int32_t*));
/* ieee style elementary long double functions */
extern long double __ieee754_sqrtl __P((long double));
extern long double __ieee754_acosl __P((long double));
extern long double __ieee754_acoshl __P((long double));
extern long double __ieee754_logl __P((long double));
extern long double __ieee754_atanhl __P((long double));
extern long double __ieee754_asinl __P((long double));
extern long double __ieee754_atan2l __P((long double,long double));
extern long double __ieee754_expl __P((long double));
extern long double __ieee754_exp2l __P((long double));
extern long double __ieee754_coshl __P((long double));
extern long double __ieee754_fmodl __P((long double,long double));
extern long double __ieee754_powl __P((long double,long double));
extern long double __ieee754_lgammal_r __P((long double,int *));
extern long double __ieee754_gammal_r __P((long double,int *));
extern long double __ieee754_lgammal __P((long double));
extern long double __ieee754_gammal __P((long double));
extern long double __ieee754_log10l __P((long double));
extern long double __ieee754_sinhl __P((long double));
extern long double __ieee754_hypotl __P((long double,long double));
extern long double __ieee754_j0l __P((long double));
extern long double __ieee754_j1l __P((long double));
extern long double __ieee754_y0l __P((long double));
extern long double __ieee754_y1l __P((long double));
extern long double __ieee754_jnl __P((int,long double));
extern long double __ieee754_ynl __P((int,long double));
extern long double __ieee754_remainderl __P((long double,long double));
extern int __ieee754_rem_pio2l __P((long double,long double*));
extern long double __ieee754_scalbl __P((long double,long double));
/* long double versions of fdlibm kernel functions */
extern long double __kernel_sinl __P((long double,long double,int));
extern long double __kernel_cosl __P((long double,long double));
extern long double __kernel_tanl __P((long double,long double,int));
extern int __kernel_rem_pio2l __P((long double*,long double*,int,int,
int,const int*));
#endif /* _MATH_PRIVATE_H_ */