Optimize libm

libm is now somewhat integrated with gcc's -ffinite-math-only option
and lots of the wrapper functions have been optimized.

1 files changed, 315 insertions, 0 deletions

diff --git a/math/bits/math-finite.h b/math/bits/math-finite.h
new file mode 100644
index 0000000000..c6b952736d
--- /dev/null
+++ b/math/bits/math-finite.h
@@ -0,0 +1,315 @@
+/* Entry points to finite-math-only compiler runs.
+   Copyright (C) 2011 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   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.
+
+   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, write to the Free
+   Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+   02111-1307 USA.  */
+
+#ifndef _MATH_H
+# error "Never use <bits/math-finite.h> directly; include <math.h> instead."
+#endif
+
+/* acos.  */
+extern double acos (double) __asm__ ("__acos_finite");
+extern float acosf (float) __asm__ ("__acosf_finite");
+#ifdef __MATH_DECLARE_LDOUBLE
+extern long double acosl (long double) __asm__ ("__acosl_finite");
+#endif
+
+#if defined __USE_MISC || defined __USE_XOPEN_EXTENDED || defined __USE_ISOC99
+/* acosh.  */
+extern double acosh (double) __asm__ ("__acosh_finite");
+extern float acoshf (float) __asm__ ("__acoshf_finite");
+# ifdef __MATH_DECLARE_LDOUBLE
+extern long double acoshl (long double) __asm__ ("__acoshl_finite");
+# endif
+#endif
+
+/* asin.  */
+extern double asin (double) __asm__ ("__asin_finite");
+extern float asinf (float) __asm__ ("__asinf_finite");
+#ifdef __MATH_DECLARE_LDOUBLE
+extern long double asinl (long double) __asm__ ("__asinl_finite");
+#endif
+
+/* atan2.  */
+extern double atan2 (double, double) __asm__ ("__atan2_finite");
+extern float atan2f (float, float) __asm__ ("__atan2f_finite");
+#ifdef __MATH_DECLARE_LDOUBLE
+extern long double atan2l (long double, long double) __asm__ ("__atan2l_finite");
+#endif
+
+#if defined __USE_MISC || defined __USE_XOPEN_EXTENDED || defined __USE_ISOC99
+/* atanh.  */
+extern double atanh (double) __asm__ ("__atanh_finite");
+extern float atanhf (float) __asm__ ("__atanhf_finite");
+# ifdef __MATH_DECLARE_LDOUBLE
+extern long double atanhl (long double) __asm__ ("__atanhl_finite");
+# endif
+#endif
+
+/* cosh.  */
+extern double cosh (double) __asm__ ("__cosh_finite");
+extern float coshf (float) __asm__ ("__coshf_finite");
+#ifdef __MATH_DECLARE_LDOUBLE
+extern long double coshl (long double) __asm__ ("__coshl_finite");
+#endif
+
+#ifdef __USE_GNU
+/* exp10.  */
+extern double exp10 (double) __asm__ ("__exp10_finite");
+extern float exp10f (float) __asm__ ("__exp10f_finite");
+# ifdef __MATH_DECLARE_LDOUBLE
+extern long double exp10l (long double) __asm__ ("__exp10l_finite");
+# endif
+
+/* pow10.  */
+extern double pow10 (double) __asm__ ("__exp10_finite");
+extern float pow10f (float) __asm__ ("__exp10f_finite");
+# ifdef __MATH_DECLARE_LDOUBLE
+extern long double pow10l (long double) __asm__ ("__exp10l_finite");
+# endif
+#endif
+
+#ifdef __USE_ISOC99
+/* exp2.  */
+extern double exp2 (double) __asm__ ("__exp2_finite");
+extern float exp2f (float) __asm__ ("__exp2f_finite");
+# ifdef __MATH_DECLARE_LDOUBLE
+extern long double exp2l (long double) __asm__ ("__exp2l_finite");
+# endif
+#endif
+
+/* fmod.  */
+extern double fmod (double, double) __asm__ ("__fmod_finite");
+extern float fmodf (float, float) __asm__ ("__fmodf_finite");
+#ifdef __MATH_DECLARE_LDOUBLE
+extern long double fmodl (long double, long double) __asm__ ("__fmodl_finite");
+#endif
+
+#ifdef __USE_ISOC99
+/* hypot.  */
+extern double hypot (double, double) __asm__ ("__hypot_finite");
+extern float hypotf (float, float) __asm__ ("__hypotf_finite");
+# ifdef __MATH_DECLARE_LDOUBLE
+extern long double hypotl (long double, long double) __asm__ ("__hypotl_finite");
+# endif
+#endif
+
+#if defined __USE_MISC || defined __USE_XOPEN
+/* j0.  */
+extern double j0 (double) __asm__ ("__j0_finite");
+extern float j0f (float) __asm__ ("__j0f_finite");
+# ifdef __MATH_DECLARE_LDOUBLE
+extern long double j0l (long double) __asm__ ("__j0l_finite");
+# endif
+
+/* y0.  */
+extern double y0 (double) __asm__ ("__y0_finite");
+extern float y0f (float) __asm__ ("__y0f_finite");
+# ifdef __MATH_DECLARE_LDOUBLE
+extern long double y0l (long double) __asm__ ("__y0l_finite");
+# endif
+
+/* j1.  */
+extern double j1 (double) __asm__ ("__j1_finite");
+extern float j1f (float) __asm__ ("__j1f_finite");
+# ifdef __MATH_DECLARE_LDOUBLE
+extern long double j1l (long double) __asm__ ("__j1l_finite");
+# endif
+
+/* y1.  */
+extern double y1 (double) __asm__ ("__y1_finite");
+extern float y1f (float) __asm__ ("__y1f_finite");
+# ifdef __MATH_DECLARE_LDOUBLE
+extern long double y1l (long double) __asm__ ("__y1l_finite");
+# endif
+
+/* jn.  */
+extern double jn (int, double) __asm__ ("__jn_finite");
+extern float jnf (int, float) __asm__ ("__jnf_finite");
+# ifdef __MATH_DECLARE_LDOUBLE
+extern long double jnl (int, long double) __asm__ ("__jnl_finite");
+# endif
+
+/* yn.  */
+extern double yn (int, double) __asm__ ("__yn_finite");
+extern float ynf (int, float) __asm__ ("__ynf_finite");
+# ifdef __MATH_DECLARE_LDOUBLE
+extern long double ynl (int, long double) __asm__ ("__ynl_finite");
+# endif
+#endif
+
+#ifdef __USE_MISC
+/* lgamma_r.  */
+extern double lgamma_r (double, int *) __asm__ ("__lgamma_r_finite");
+extern float lgammaf_r (float, int *) __asm__ ("__lgammaf_r_finite");
+# ifdef __MATH_DECLARE_LDOUBLE
+extern long double lgammal_r (long double, int *) __asm__ ("__lgammal_r_finite");
+# endif
+#endif
+
+#if defined __USE_MISC || defined __USE_XOPEN || defined __USE_ISOC99
+/* lgamma.  */
+__extern_always_inline double lgamma (double __d)
+{
+# ifdef __USE_ISOC99
+  int __local_signgam = 0;
+  return lgamma_r (__d, &__local_signgam);
+# else
+  return lgamma_r (__d, &signgam);
+# endif
+}
+__extern_always_inline float lgammaf (float __d)
+{
+# ifdef __USE_ISOC99
+  int __local_signgam = 0;
+  return lgammaf_r (__d, &__local_signgam);
+# else
+  return lgammaf_r (__d, &signgam);
+# endif
+}
+# ifdef __MATH_DECLARE_LDOUBLE
+__extern_always_inline long double lgammal (long double __d)
+{
+# ifdef __USE_ISOC99
+  int __local_signgam = 0;
+  return lgammal_r (__d, &__local_signgam);
+# else
+  return lgammal_r (__d, &signgam);
+# endif
+}
+# endif
+#endif
+
+#if defined __USE_MISC || defined __USE_XOPEN
+/* gamma.  */
+__extern_always_inline double gamma (double __d)
+{
+# ifdef __USE_ISOC99
+  int __local_signgam = 0;
+  return lgamma_r (__d, &__local_signgam);
+# else
+  return lgamma_r (__d, &signgam);
+# endif
+}
+__extern_always_inline float gammaf (float __d)
+{
+# ifdef __USE_ISOC99
+  int __local_signgam = 0;
+  return lgammaf_r (__d, &__local_signgam);
+# else
+  return lgammaf_r (__d, &signgam);
+# endif
+}
+# ifdef __MATH_DECLARE_LDOUBLE
+__extern_always_inline long double gammal (long double __d)
+{
+#  ifdef __USE_ISOC99
+  int __local_signgam = 0;
+  return lgammal_r (__d, &__local_signgam);
+#  else
+  return lgammal_r (__d, &signgam);
+#  endif
+}
+# endif
+#endif
+
+/* log.  */
+extern double log (double) __asm__ ("__log_finite");
+extern float logf (float) __asm__ ("__logf_finite");
+#ifdef __MATH_DECLARE_LDOUBLE
+extern long double logl (long double) __asm__ ("__logl_finite");
+#endif
+
+/* log10.  */
+extern double log10 (double) __asm__ ("__log10_finite");
+extern float log10f (float) __asm__ ("__log10f_finite");
+#ifdef __MATH_DECLARE_LDOUBLE
+extern long double log10l (long double) __asm__ ("__log10l_finite");
+#endif
+
+#ifdef __USE_ISOC99
+/* log2.  */
+extern double log2 (double) __asm__ ("__log2_finite");
+extern float log2f (float) __asm__ ("__log2f_finite");
+# ifdef __MATH_DECLARE_LDOUBLE
+extern long double log2l (long double) __asm__ ("__log2l_finite");
+# endif
+#endif
+
+/* pow.  */
+extern double pow (double, double) __asm__ ("__pow_finite");
+extern float powf (float, float) __asm__ ("__powf_finite");
+#ifdef __MATH_DECLARE_LDOUBLE
+extern long double powl (long double, long double) __asm__ ("__powl_finite");
+#endif
+
+/* remainder.  */
+extern double remainder (double, double) __asm__ ("__remainder_finite");
+extern float remainderf (float, float) __asm__ ("__remainderf_finite");
+#ifdef __MATH_DECLARE_LDOUBLE
+extern long double remainderl (long double, long double) __asm__ ("__remainderl_finite");
+#endif
+
+#if defined __USE_MISC || defined __USE_XOPEN_EXTENDED
+/* scalb.  */
+extern double scalb (double, double) __asm__ ("__scalb_finite");
+extern float scalbf (float, float) __asm__ ("__scalbf_finite");
+# ifdef __MATH_DECLARE_LDOUBLE
+extern long double scalbl (long double, long double) __asm__ ("__scalbl_finite");
+# endif
+#endif
+
+/* sinh.  */
+extern double sinh (double) __asm__ ("__sinh_finite");
+extern float sinhf (float) __asm__ ("__sinhf_finite");
+#ifdef __MATH_DECLARE_LDOUBLE
+extern long double sinhl (long double) __asm__ ("__sinhl_finite");
+#endif
+
+/* sqrt.  */
+extern double sqrt (double) __asm__ ("__sqrt_finite");
+extern float sqrtf (float) __asm__ ("__sqrtf_finite");
+#ifdef __MATH_DECLARE_LDOUBLE
+extern long double sqrtl (long double) __asm__ ("__sqrtl_finite");
+#endif
+
+#ifdef __USE_ISOC99
+/* tgamma.  */
+extern double __gamma_r_finite (double, int *);
+__extern_always_inline double tgamma (double __d)
+{
+  int __local_signgam = 0;
+  double __res = __gamma_r_finite (__d, &__local_signgam);
+  return __local_signgam < 0 ? -__res : __res;
+}
+extern float __gammaf_r_finite (float, int *);
+__extern_always_inline float tgammaf (float __d)
+{
+  int __local_signgam = 0;
+  float __res = __gammaf_r_finite (__d, &__local_signgam);
+  return __local_signgam < 0 ? -__res : __res;
+}
+# ifdef __MATH_DECLARE_LDOUBLE
+extern long double __gammal_r_finite (long double, int *);
+__extern_always_inline long double tgammal (long double __d)
+{
+  int __local_signgam = 0;
+  long double __res = __gammal_r_finite (__d, &__local_signgam);
+  return __local_signgam < 0 ? -__res : __res;
+}
+# endif
+#endif