glibc/sysdeps/libm-i387/s_asinhl.S
Ulrich Drepper d705269e75 Update.
1997-06-19 19:38  Ulrich Drepper  <drepper@cygnus.com>

	* features.h: Define __STDC_IEC_559__ and _STDC_IEC_559_COMPLEX__.

	* elf/dl-minimal.c (__dcgettext): Remove assertion.

	* inet/rcmd.c: Correct a few typos.  Reported by Eric Troan.

	* manual/Makefile (distribute): Add dir.
	* manual/dir: New file.

	* math/Makefile (libm-support): Rename s_rinttol, s_rinttoll,
	s_roundtol, and s_roundtoll to s_lrint, s_llrint, s_lround,
	and s_llround respectively.
	(libm-calls): Add e_gamma_r.
	* math/libm-test.c (check_int_exc): New function.
	(signbit_test): Remove test for sign of NaN.
	(gamma_test): Clear exception after test of existence.
	Correct tests and and epsilons.
	(lgamma_test): Likewise.
	(ilogb_test): Correct all tests.
	(scalb_test): Rewrite.
	(rinttol_test): Rename to lrint_test and correct tests.
	(rinttoll_test): Likewise.
	(roundtol_test): Likewise.
	(roundtoll_test): Likewise.
	(main): Call lrint/lround functions instead of rinttol/roundtol.
	* math/math.h: Change prototypes for rinttol/roundtol.
	* math/mathcalls.h: Rearrange prototypes according to ISO C9X draft.
	* sysdeps/generic/mathbits.h: Define FP_ILOGB0 and FP_ILOGBNAN.
	* sysdeps/i386/mathbits.h: Likewise.
	* sysdeps/libm-i387/e_scalb.S: Handle special cases correctly.
	* sysdeps/libm-i387/e_scalbf.S: Likewise.
	* sysdeps/libm-i387/e_scalbl.S: Likewise.
	* sysdeps/libm-i387/s_asinh.S: Handle -inf correctly.
	* sysdeps/libm-i387/s_asinhf.S: Likewise.
	* sysdeps/libm-i387/s_asinhl.S: Likewise.
	* sysdeps/libm-i387/s_ilogb.S: Optimize.
	* sysdeps/libm-i387/s_ilogbf.S: Likewise.
	* sysdeps/libm-i387/s_ilogbl.S: Likewise.
	* sysdeps/libm-i387/s_rinttol.S: Rename to...
	* sysdeps/libm-i387/s_lrint.S: ...this.
	* sysdeps/libm-i387/s_rinttoll.S: Rename to...
	* sysdeps/libm-i387/s_llrint.S: ...this.
	* sysdeps/libm-i387/s_remquo.S: Correctly set sign of remainder.
	* sysdeps/libm-i387/s_remquof.S: Likewise.
	* sysdeps/libm-i387/s_remquol.S: Likewise.
	* sysdeps/libm-i387/e_gamma_r.c: New file.  Implementation of gamma
	function according to ISO C.
	* sysdeps/libm-i387/e_gammaf_r.c: New file.
	* sysdeps/libm-i387/e_gammal_r.c: New file.
	* sysdeps/libm-i387/e_lgamma_r.c: Don't let optimize compile the
	generation of exceptions away.
	* sysdeps/libm-i387/e_lgammaf_r.c: Likewise.
	* sysdeps/libm-i387/k_standard.c: Correct return value for infinity
	points of gamma function when not SVID mode.
	* sysdeps/libm-i387/s_rinttoll.c: Renamed to...
	* sysdeps/libm-i387/s_llrint.c: ...this.
	* sysdeps/libm-i387/s_rinttol.c: Renamed to...
	* sysdeps/libm-i387/s_lrint.c: ...this.
	* sysdeps/libm-i387/s_roundtoll.c: Renamed to...
	* sysdeps/libm-i387/s_llround.c: ...this.
	* sysdeps/libm-i387/s_roundtol.c: Renamed to...
	* sysdeps/libm-i387/s_lround.c: ..this.
	* sysdeps/libm-i387/s_scalbn.c: Change second parameter according to
	ISO C.
	* sysdeps/libm-i387/s_scalbnf.c: Likewise.
	* sysdeps/libm-i387/s_scalbnl.c: Likewise.
	* sysdeps/libm-i387/w_gamma.c: Call __ieee754_gamma_r if library
	mode is not _SVID_.
	* sysdeps/libm-i387/w_gammaf.c: Likewise.
	* sysdeps/libm-i387/w_gammal.c: Likewise.
	* sysdeps/m68k/fpu/__math.h: Rename __rinttol to __lrint and
	rinttol to lrint.
	* sysdeps/m68k/fpu/s_rinttol.c: Renamed to...
	* sysdeps/m68k/fpu/s_lrint.c: ...this.
	* sysdeps/m68k/fpu/s_rinttoll.c: Renamed to...
	* sysdeps/m68k/fpu/s_llrint.c: ...this.

	* md5-crypt/Makefile: Link md5test program with md5.o.

	* stdio-common/temptest.c: Don't use __stdio_gen_tempname which
	is not exported by the libc.so.

	* stdio-common/vfscanf.c: Correct scanning of strings after last
	change.

	* sysdeps/unix/sysv/linux/i386/sysdep.S: Use .comm to define errno.

1997-06-19 07:37  H.J. Lu  <hjl@gnu.ai.mit.edu>

	* time/tzfile.c (__tzfile_read): Store getc () return in int.

1997-06-13  Miles Bader  <miles@gnu.ai.mit.edu>

	* argp/argp-parse.c (argp_version_parser): Include `(PROGRAM ERROR)' in
	the no-version error text to indicate that something's fucked.
	[!_] (N_): New macro.
	(argp_default_options, argp_version_options): Wrap doc strings in N_().

1997-06-12  Miles Bader  <miles@gnu.ai.mit.edu>

	* argp/argp-parse.c (parser_parse_opt): Detect and report unhandled
	options here.
	(parser_parse_arg): Handle ARGP_KEY_ARGS here.
	Adjust NEXT pointer back if we fail to parse anything.
	(parser_parse_next): Simplify arg code.  Leave state NEXT frobbing
	to parser_parse_arg.

1997-06-11  Miles Bader  <miles@gnu.ai.mit.edu>

	* argp/argp.h (ARGP_KEY_ARGS, ARGP_KEY_FINI): New macros.
	* argp/argp-parse.c (parser_finalize): Do another pass over the
	parsers with ARGP_KEY_FINI.

1997-06-18  Miles Bader  <miles@gnu.ai.mit.edu>

	* string/Makefile (routines): Add argz-replace.

1997-06-16 00:16  Miles Bader  <miles@gnu.ai.mit.edu>

	* manual/string.texi (Argz Functions): Document argz_replace.

1997-06-12  Miles Bader  <miles@gnu.ai.mit.edu>

	* string/argz.h (__argz_replace, argz_replace): New declarations.
	* string/argz-replace.c: New file.

1997-06-11  Miles Bader  <miles@gnu.ai.mit.edu>

	* manual/argp.texi (Argp Special Keys): Document ARGP_KEY_ARGS.

1997-06-16 23:08  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* Makeconfig (libc-map): Remove definition.
	* Makerules (libc-map): Define it here, using the full name.
	(load-map-file): Remove case for empty $(..).
	($(common-objpfx)libc.so): Revert last change.

1997-06-17 22:18  Mark Kettenis  <kettenis@phys.uva.nl>

	* login/programs/utmpd.c (main): Improve signal handling.

	* login/programs/request.c (do_setutent, do_updwtmp):
	Allow arbitrary length filenames.
	* login/programs/utmpd.h (setutent_request, updwtmp_request):
	Get rid of fixed length file field.
	* login/utmp_daemon.c (do_setutent, do_updwtmp):
	Allow arbitrary	length filenames.

	* login/programs/request.c (do_pututline):
	Don't fail if connection->position is -1 on entry.

1997-06-15 16:32  Mark Kettenis  <kettenis@phys.uva.nl>

	* login/utmp_file.c (updwtmp_file): Use the same method for
	appending an entry as in pututline_file.

1997-06-11 18:59  Mark Kettenis  <kettenis@phys.uva.nl>

	* login/utmp_file.c (getutent_r_file):
	Use read lock instead of write lock.
	(getutline_r_file, internal_getut_r): Lock utmp file.
	(updwtmp_file): Use fcntl to lock file instead of flock.

1997-06-18 00:11  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* stdio-common/vfscanf.c (inchar, ungetc): Don't count EOF as
	character read in/put back.
	* stdio-common/tstscanf.c: Add test case for this.

1997-06-17 22:17  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* libc.map: Add more symbols.

1997-06-18 12:01  Ulrich Drepper  <drepper@cygnus.com>

	* manual/Makefile (install): Make sure `dir' file exists if we use
	install-info.

1997-06-17 19:32  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* manual/Makefile (info): Depend on dir-add.info.

1997-06-17 17:19  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* elf/ldd.bash.in: Fix spacing in message.

1997-06-17 14:28  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* elf/do-rel.h (elf_dynamic_do_rel): Always use version
	information if available.

1997-06-17 11:34  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* sunrpc/Makefile ($(rpcsvc:%.x=$(objpfx)rpcsvc/%.h)): Make
	command non-empty to force make to recheck modification time.
	($(rpcsvc:%.x=$(objpfx)x%.c)): Likewise.

1997-06-17 00:26  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* sysdeps/stub/e_acoshl.c: Set errno to ENOSYS.
	* sysdeps/stub/e_acosl.c: Likewise.
	* sysdeps/stub/e_asinl.c: Likewise.
	* sysdeps/stub/e_atan2l.c: Likewise.
	* sysdeps/stub/e_expl.c: Likewise.
	* sysdeps/stub/e_fmodl.c: Likewise.
	* sysdeps/stub/e_j0l.c: Likewise.
	* sysdeps/stub/e_j1l.c: Likewise.
	* sysdeps/stub/e_jnl.c: Likewise.
	* sysdeps/stub/e_lgammal_r.c: Likewise.
	* sysdeps/stub/e_log10l.c: Likewise.
	* sysdeps/stub/e_logl.c: Likewise.
	* sysdeps/stub/e_powl.c: Likewise.
	* sysdeps/stub/e_rem_pio2l.c: Likewise.
	* sysdeps/stub/e_sqrtl.c: Likewise.
	* sysdeps/stub/k_cosl.c: Likewise.
	* sysdeps/stub/k_rem_pio2l.c: Likewise.
	* sysdeps/stub/k_sinl.c: Likewise.
	* sysdeps/stub/k_tanl.c: Likewise.
	* sysdeps/stub/s_atanl.c: Likewise.
	* sysdeps/stub/s_erfl.c: Likewise.
	* sysdeps/stub/s_exp2.c: Likewise.
	* sysdeps/stub/s_exp2f.c: Likewise.
	* sysdeps/stub/s_exp2l.c: Likewise.
	* sysdeps/stub/s_expm1l.c: Likewise.
	* sysdeps/stub/s_log1pl.c: Likewise.
	* sysdeps/stub/s_log2l.c: Likewise.

1997-06-18 11:46  Ulrich Drepper  <drepper@cygnus.com>

	* sysdeps/libm-ieee754/w_gamma.c: If _LIB_VERSION is _SVID_ compute
	result as before last change.
	* sysdeps/libm-ieee754/w_gammaf.c: Likewise.
	* sysdeps/libm-ieee754/w_gammal.c: Likewise.

1997-06-16 23:37  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* sysdeps/libm-ieee754/s_remquo.c: Fix off-by-one when computing
	quotient.
	* sysdeps/libm-ieee754/s_remquof.c: Likewise.
	* sysdeps/libm-ieee754/s_remquol.c: Likewise.

	* sysdeps/m68k/fpu/s_remquo.c: Remove FIXME and special case for
	quotient.

	* sysdeps/libm-ieee754/w_gamma.c: Add missing call to exp
	function.  Don't use global signgam.
	* sysdeps/libm-ieee754/w_gammaf.c: Likewise.
	* sysdeps/libm-ieee754/w_gammal.c: Likewise.

	* math/Makefile (libm-calls): Remove w_gamma_r.
	* sysdeps/libm-ieee754/w_gamma_r.c: Remove file.
	* sysdeps/libm-ieee754/w_gammaf_r.c: Remove file.
	* sysdeps/libm-ieee754/w_gammal_r.c: Remove file.

	* math/libm-test.c (atanh_test): Declare x only if needed.
	(signbit_test): Fix typo.
	(gamma_test): Check whether function is implemented.  Add
	epsilons.
	(lgamma_test): Likewise.
	(fmod_test): Add epsilons.
	(exp2_test): Use right function for existence test.

1997-06-07 09:20  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* sysdeps/powerpc/Dist: Add fenv_const.c.
	* sysdeps/unix/sysv/linux/Dist: Add net/if_slip.h.
	* sysdeps/unix/sysv/linux/powerpc/Dist: Add init-first.h and
	syscall.h.
	* sysdeps/unix/sysv/linux/sparc/Dist: Add init-first.h.
	* string/Makefile (distribute): Add tst-svc.expect.
	* nis/Makefile (distribute): Add nis_intern.h and Banner.
	* elf/Makefile (distribute): Add dl-hash.h.
	* Rules (subdir_echo-distinfo): Add headers from $(distribute).

	* login/Makefile (others): Add utmpdump.
	* sysdeps/unix/sysv/linux/Makefile (sysdep_headers): Add
	net/if_slip.h.

	* manual/Makefile (dir-add.texi): Also look in indirectly included
	files.

1997-06-16 23:15  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* sysdeps/unix/sysv/linux/configure.in: Remove commands to
	generate stdio_lim.h.
	* sysdeps/unix/sysv/linux/mk-stdiolim.c: Remove.
	* sysdeps/unix/sysv/linux/Makefile: Add rules to generate
	stdio_lim.h here.
	(common-generated): Add generated files.
	(inhibit-stdio_lim): Define.
	* sysdeps/posix/Makefile [$(inhibit-stdio_lim)=yes]: Disable rules
	to generate stdio_lim.h.

	* sysdeps/unix/sysv/linux/Makefile: Suppress inclusion of
	dependecy files if no_deps is set.
	($(objpfx)syscall-%.d): Add header file as target to dependency
	generation.

1997-06-14 19:19  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* sysdeps/unix/sysv/linux/arm/Dist: New file.

1997-06-14 17:59  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* extra-lib.mk (others): Depend on versioned shared library, not
	the unversioned one.
	* Makerules (build-shlib): Don't make the version link here.
	($(common-objpfx)libc.so$(libc.so-version)): New rule for libc
	version link.

1997-06-16 03:07  Ulrich Drepper  <drepper@cygnus.com>

	* sysdeps/libm-ieee754/k_standard.c: Undo change of Tue Aug 6
	01:13:56 1996.

	* argp/argp-help.c (argp_args_usage): Supply correct argp to
	filter_doc.
	* argp/argp-help.c (hol_add_cluster): Initialize CL->depth.
	* argp/argp-help.c (_help): Supply STATE to argp_args_usage.
	* argp/argp.h (ARGP_KEY_HELP_ARGS_DOC): New macro.
	* argp/argp-fmtstream.c: Add casts to prevent warnings.
	* argp/argp.h (OPTION_NO_USAGE): New macro.
	* argp/argp-help.c (usage_long_opt, usage_argful_short_opt,
	* argp/argp-fmtstream.c (__argp_fmtstream_update): Account for case
	* argp/argp-help.c <stddef.h>: New include.
	* argp/argp.h (argp_state_help, __argp_state_help, argp_usage,
	* argp/argp.h (argp_program_bug_address): Make const.
	* argp/argp-parse.c (argp_default_parser): Set STATE->name for OPT_PROGNAME.
	* argp/argp-help.c (__argp_error, __argp_failure, __argp_state_help):
	* argp/argp-parse.c (parser_init): Set PARSER->state.flags.
1997-06-19 19:11:22 +00:00

148 lines
3.8 KiB
ArmAsm

/* ix87 specific implementation of arcsinh.
Copyright (C) 1996, 1997 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Ulrich Drepper <drepper@cygnus.com>, 1996.
The GNU C Library is free software; you can redistribute it and/or
modify it under the terms of the GNU Library General Public License as
published by the Free Software Foundation; either version 2 of the
License, or (at your option) any later version.
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
Library General Public License for more details.
You should have received a copy of the GNU Library General Public
License along with the GNU C Library; see the file COPYING.LIB. If not,
write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
Boston, MA 02111-1307, USA. */
#include <machine/asm.h>
#ifdef __ELF__
.section .rodata
#else
.text
#endif
.align ALIGNARG(4)
ASM_TYPE_DIRECTIVE(huge,@object)
huge: .tfloat 1e+4930
ASM_SIZE_DIRECTIVE(huge)
.align ALIGNARG(4)
/* Please note that we use double value for 1.0. This number
has an exact representation and so we don't get accuracy
problems. The advantage is that the code is simpler. */
ASM_TYPE_DIRECTIVE(one,@object)
one: .double 1.0
ASM_SIZE_DIRECTIVE(one)
/* It is not important that this constant is precise. It is only
a value which is known to be on the safe side for using the
fyl2xp1 instruction. */
ASM_TYPE_DIRECTIVE(limit,@object)
limit: .double 0.29
ASM_SIZE_DIRECTIVE(limit)
#ifdef PIC
#define MO(op) op##@GOTOFF(%edx)
#else
#define MO(op) op
#endif
.text
ENTRY(__asinhl)
movl 12(%esp), %ecx
movl $0x7fff, %eax
andl %ecx, %eax
andl $0x8000, %ecx
movl %eax, %edx
orl $0xffff8000, %edx
incl %edx
jz 7f // x in ±Inf or NaN
xorl %ecx, 12(%esp)
fldt 4(%esp) // |x|
cmpl $0x3fde, %eax
jb 2f // |x| < 2^-34
fldln2 // log(2) : |x|
cmpl $0x4020, %eax
fxch // |x| : log(2)
ja 3f // |x| > 2^34
#ifdef PIC
call 1f
1: popl %edx
addl $_GLOBAL_OFFSET_TABLE_+[.-1b], %edx
#endif
cmpl $0x4000, %eax
ja 5f // |x| > 2
// 2^-34 <= |x| <= 2 => y = sign(x)*log1p(|x|+|x|^2/(1+sqrt(1+|x|^2)))
fld %st // |x| : |x| : log(2)
fmul %st(1) // |x|^2 : |x| : log(2)
fld %st // |x|^2 : |x|^2 : |x| : log(2)
faddl MO(one) // 1+|x|^2 : |x|^2 : |x| : log(2)
fsqrt // sqrt(1+|x|^2) : |x|^2 : |x| : log(2)
faddl MO(one) // 1+sqrt(1+|x|^2) : |x|^2 : |x| : log(2)
fdivrp // |x|^2/(1+sqrt(1+|x|^2)) : |x| : log(2)
faddp // |x|+|x|^2/(1+sqrt(1+|x|^2)) : log(2)
fcoml MO(limit)
fnstsw
sahf
ja 6f
fyl2xp1
jecxz 4f
fchs
4: ret
7: fldt 4(%esp)
ret
6: faddl MO(one)
fyl2x
jecxz 4f
fchs
4: ret
// |x| < 2^-34 => y = x (inexact iff |x| != 0.0)
.align ALIGNARG(4)
2:
#ifdef PIC
call 1f
1: popl %edx
addl $_GLOBAL_OFFSET_TABLE_+[.-1b], %edx
#endif
jecxz 4f
fchs // x
4: fld %st // x : x
fldt MO(huge) // huge : x : x
faddp // huge+x : x
fstp %st(0) // x
ret
// |x| > 2^34 => y = sign(x) * (log(|x|) + log(2))
.align ALIGNARG(4)
3: fyl2x // log(|x|)
fldln2 // log(2) : log(|x|)
faddp // log(|x|)+log(2)
jecxz 4f
fchs
4: ret
// |x| > 2 => y = sign(x) * log(2*|x| + 1/(|x|+sqrt(x*x+1)))
.align ALIGNARG(4)
5: fld %st // |x| : |x| : log(2)
fadd %st, %st(1) // |x| : 2*|x| : log(2)
fld %st // |x| : |x| : 2*|x| : log(2)
fmul %st(1) // |x|^2 : |x| : 2*|x| : log(2)
faddl MO(one) // 1+|x|^2 : |x| : 2*|x| : log(2)
fsqrt // sqrt(1+|x|^2) : |x| : 2*|x| : log(2)
faddp // |x|+sqrt(1+|x|^2) : 2*|x| : log(2)
fdivrl MO(one) // 1/(|x|+sqrt(1+|x|^2)) : 2*|x| : log(2)
faddp // 2*|x|+1/(|x|+sqrt(1+|x|^2)) : log(2)
fyl2x // log(2*|x|+1/(|x|+sqrt(1+|x|^2)))
jecxz 4f
fchs
4: ret
END(__asinhl)
weak_alias (__asinhl, asinhl)