Use __RENAME for long double functions.
We can cut a lot of stuff out of the NDK's libandroid_support with this,
and reduce unnecessary relocations for all LP32 code. LP64 code should
be unaffected.
Bug: https://issuetracker.google.com/64450768
Bug: https://github.com/android-ndk/ndk/issues/507
Test: ran tests, plus manual readelf on the _test.o files
Change-Id: I3de6015921195304ea9c829ef31665cd34664066
diff --git a/libc/bionic/strtold.cpp b/libc/bionic/strtold.cpp
index c55dd61..6b673e6 100644
--- a/libc/bionic/strtold.cpp
+++ b/libc/bionic/strtold.cpp
@@ -26,6 +26,8 @@
* SUCH DAMAGE.
*/
+#define __BIONIC_LP32_USE_LONG_DOUBLE
+
#include <float.h>
#include <stdlib.h>
diff --git a/libc/bionic/wcstod.cpp b/libc/bionic/wcstod.cpp
index f7bd433..41df854 100644
--- a/libc/bionic/wcstod.cpp
+++ b/libc/bionic/wcstod.cpp
@@ -26,6 +26,8 @@
* SUCH DAMAGE.
*/
+#define __BIONIC_LP32_USE_LONG_DOUBLE
+
#include <wchar.h>
#include <stdlib.h>
diff --git a/libc/include/complex.h b/libc/include/complex.h
index 79fe07c..3fe8f35 100644
--- a/libc/include/complex.h
+++ b/libc/include/complex.h
@@ -61,99 +61,99 @@
/* 7.3.5.1 The cacos functions */
double complex cacos(double complex __z) __INTRODUCED_IN(23);
float complex cacosf(float complex __z) __INTRODUCED_IN(23);
-long double complex cacosl(long double complex __z) __INTRODUCED_IN(26);
+long double complex cacosl(long double complex __z) __RENAME_LDBL(cacos, 23, 26);
/* 7.3.5.2 The casin functions */
double complex casin(double complex __z) __INTRODUCED_IN(23);
float complex casinf(float complex __z) __INTRODUCED_IN(23);
-long double complex casinl(long double complex __z) __INTRODUCED_IN(26);
+long double complex casinl(long double complex __z) __RENAME_LDBL(casin, 23, 26);
/* 7.3.5.1 The catan functions */
double complex catan(double complex __z) __INTRODUCED_IN(23);
float complex catanf(float complex __z) __INTRODUCED_IN(23);
-long double complex catanl(long double complex __z) __INTRODUCED_IN(26);
+long double complex catanl(long double complex __z) __RENAME_LDBL(catan, 23, 26);
/* 7.3.5.1 The ccos functions */
double complex ccos(double complex __z) __INTRODUCED_IN(23);
float complex ccosf(float complex __z) __INTRODUCED_IN(23);
-long double complex ccosl(long double complex __z) __INTRODUCED_IN(26);
+long double complex ccosl(long double complex __z) __RENAME_LDBL(ccos, 23, 26);
/* 7.3.5.1 The csin functions */
double complex csin(double complex __z) __INTRODUCED_IN(23);
float complex csinf(float complex __z) __INTRODUCED_IN(23);
-long double complex csinl(long double complex __z) __INTRODUCED_IN(26);
+long double complex csinl(long double complex __z) __RENAME_LDBL(csin, 23, 26);
/* 7.3.5.1 The ctan functions */
double complex ctan(double complex __z) __INTRODUCED_IN(23);
float complex ctanf(float complex __z) __INTRODUCED_IN(23);
-long double complex ctanl(long double complex __z) __INTRODUCED_IN(26);
+long double complex ctanl(long double complex __z) __RENAME_LDBL(ctan, 23, 26);
/* 7.3.6 Hyperbolic functions */
/* 7.3.6.1 The cacosh functions */
double complex cacosh(double complex __z) __INTRODUCED_IN(23);
float complex cacoshf(float complex __z) __INTRODUCED_IN(23);
-long double complex cacoshl(long double complex __z) __INTRODUCED_IN(26);
+long double complex cacoshl(long double complex __z) __RENAME_LDBL(cacosh, 23, 26);
/* 7.3.6.2 The casinh functions */
double complex casinh(double complex __z) __INTRODUCED_IN(23);
float complex casinhf(float complex __z) __INTRODUCED_IN(23);
-long double complex casinhl(long double complex __z) __INTRODUCED_IN(26);
+long double complex casinhl(long double complex __z) __RENAME_LDBL(casinh, 23, 26);
/* 7.3.6.3 The catanh functions */
double complex catanh(double complex __z) __INTRODUCED_IN(23);
float complex catanhf(float complex __z) __INTRODUCED_IN(23);
-long double complex catanhl(long double complex __z) __INTRODUCED_IN(26);
+long double complex catanhl(long double complex __z) __RENAME_LDBL(catanh, 23, 26);
/* 7.3.6.4 The ccosh functions */
double complex ccosh(double complex __z) __INTRODUCED_IN(23);
float complex ccoshf(float complex __z) __INTRODUCED_IN(23);
-long double complex ccoshl(long double complex __z) __INTRODUCED_IN(26);
+long double complex ccoshl(long double complex __z) __RENAME_LDBL(ccosh, 23, 26);
/* 7.3.6.5 The csinh functions */
double complex csinh(double complex __z) __INTRODUCED_IN(23);
float complex csinhf(float complex __z) __INTRODUCED_IN(23);
-long double complex csinhl(long double complex __z) __INTRODUCED_IN(26);
+long double complex csinhl(long double complex __z) __RENAME_LDBL(csinh, 23, 26);
/* 7.3.6.6 The ctanh functions */
double complex ctanh(double complex __z) __INTRODUCED_IN(23);
float complex ctanhf(float complex __z) __INTRODUCED_IN(23);
-long double complex ctanhl(long double complex __z) __INTRODUCED_IN(26);
+long double complex ctanhl(long double complex __z) __RENAME_LDBL(ctanh, 23, 26);
/* 7.3.7 Exponential and logarithmic functions */
/* 7.3.7.1 The cexp functions */
double complex cexp(double complex __z) __INTRODUCED_IN(23);
float complex cexpf(float complex __z) __INTRODUCED_IN(23);
-long double complex cexpl(long double complex __z) __INTRODUCED_IN(26);
+long double complex cexpl(long double complex __z) __RENAME_LDBL(cexp, 23, 26);
/* 7.3.7.2 The clog functions */
double complex clog(double complex __z) __INTRODUCED_IN(26);
float complex clogf(float complex __z) __INTRODUCED_IN(26);
-long double complex clogl(long double complex __z) __INTRODUCED_IN(26);
+long double complex clogl(long double complex __z) __RENAME_LDBL(clog, 26, 26);
/* 7.3.8 Power and absolute-value functions */
/* 7.3.8.1 The cabs functions */
double cabs(double complex __z) __INTRODUCED_IN(23);
float cabsf(float complex __z) __INTRODUCED_IN(23);
-long double cabsl(long double complex __z) __INTRODUCED_IN_32(21) __INTRODUCED_IN_64(23);
+long double cabsl(long double complex __z) __INTRODUCED_IN_32(21) __INTRODUCED_IN_64(23) /*__RENAME_LDBL(cabs)*/;
/* 7.3.8.2 The cpow functions */
double complex cpow(double complex __x, double complex __z) __INTRODUCED_IN(26);
float complex cpowf(float complex __x, float complex __z) __INTRODUCED_IN(26);
-long double complex cpowl(long double complex __x, long double complex __z) __INTRODUCED_IN(26);
+long double complex cpowl(long double complex __x, long double complex __z) __RENAME_LDBL(cpow, 26, 26);
/* 7.3.8.3 The csqrt functions */
double complex csqrt(double complex __z) __INTRODUCED_IN(23);
float complex csqrtf(float complex __z) __INTRODUCED_IN(23);
-long double complex csqrtl(long double complex __z) __INTRODUCED_IN_32(21) __INTRODUCED_IN_64(23);
+long double complex csqrtl(long double complex __z) __INTRODUCED_IN_32(21) __INTRODUCED_IN_64(23) /*__RENAME_LDBL(csqrt)*/;
/* 7.3.9 Manipulation functions */
/* 7.3.9.1 The carg functions */
double carg(double complex __z) __INTRODUCED_IN(23);
float cargf(float complex __z) __INTRODUCED_IN(23);
-long double cargl(long double complex __z) __INTRODUCED_IN(23);
+long double cargl(long double complex __z) __RENAME_LDBL(carg, 23, 23);
/* 7.3.9.2 The cimag functions */
double cimag(double complex __z) __INTRODUCED_IN(23);
float cimagf(float complex __z) __INTRODUCED_IN(23);
-long double cimagl(long double complex __z) __INTRODUCED_IN(23);
+long double cimagl(long double complex __z) __RENAME_LDBL(cimag, 23, 23);
/* 7.3.9.3 The conj functions */
double complex conj(double complex __z) __INTRODUCED_IN(23);
float complex conjf(float complex __z) __INTRODUCED_IN(23);
-long double complex conjl(long double complex __z) __INTRODUCED_IN(23);
+long double complex conjl(long double complex __z) __RENAME_LDBL(conj, 23, 23);
/* 7.3.9.4 The cproj functions */
double complex cproj(double complex __z) __INTRODUCED_IN(23);
float complex cprojf(float complex __z) __INTRODUCED_IN(23);
-long double complex cprojl(long double complex __z) __INTRODUCED_IN_32(21) __INTRODUCED_IN_64(23);
+long double complex cprojl(long double complex __z) __INTRODUCED_IN_32(21) __INTRODUCED_IN_64(23) /*__RENAME_LDBL(cproj)*/;
/* 7.3.9.5 The creal functions */
double creal(double complex __z) __INTRODUCED_IN(23);
float crealf(float complex __z) __INTRODUCED_IN(23);
-long double creall(long double complex __z) __INTRODUCED_IN(23);
+long double creall(long double complex __z) __RENAME_LDBL(creal, 23, 23);
__END_DECLS
diff --git a/libc/include/math.h b/libc/include/math.h
index 248ed3c..3d0ec49 100644
--- a/libc/include/math.h
+++ b/libc/include/math.h
@@ -22,19 +22,26 @@
__BEGIN_DECLS
-#define HUGE_VAL __builtin_huge_val()
+/* C11. */
-#define FP_ILOGB0 (-INT_MAX)
-#define FP_ILOGBNAN INT_MAX
+typedef double __double_t;
+typedef __double_t double_t;
+typedef float __float_t;
+typedef __float_t float_t;
-#define HUGE_VALF __builtin_huge_valf()
-#define HUGE_VALL __builtin_huge_vall()
-#define INFINITY __builtin_inff()
-#define NAN __builtin_nanf("")
+#define HUGE_VAL __builtin_huge_val()
+#define HUGE_VALF __builtin_huge_valf()
+#define HUGE_VALL __builtin_huge_vall()
-#define MATH_ERRNO 1
-#define MATH_ERREXCEPT 2
-#define math_errhandling MATH_ERREXCEPT
+#define INFINITY __builtin_inff()
+
+#define NAN __builtin_nanf("")
+
+#define FP_INFINITE 0x01
+#define FP_NAN 0x02
+#define FP_NORMAL 0x04
+#define FP_SUBNORMAL 0x08
+#define FP_ZERO 0x10
#if defined(__FP_FAST_FMA)
#define FP_FAST_FMA 1
@@ -46,20 +53,257 @@
#define FP_FAST_FMAL 1
#endif
-/* Symbolic constants to classify floating point numbers. */
-#define FP_INFINITE 0x01
-#define FP_NAN 0x02
-#define FP_NORMAL 0x04
-#define FP_SUBNORMAL 0x08
-#define FP_ZERO 0x10
-#define fpclassify(x) \
- __builtin_fpclassify(FP_NAN, FP_INFINITE, FP_NORMAL, FP_SUBNORMAL, FP_ZERO, x)
+#define FP_ILOGB0 (-INT_MAX)
+#define FP_ILOGBNAN INT_MAX
+
+#define MATH_ERRNO 1
+#define MATH_ERREXCEPT 2
+#define math_errhandling MATH_ERREXCEPT
+
+#define fpclassify(x) __builtin_fpclassify(FP_NAN, FP_INFINITE, FP_NORMAL, FP_SUBNORMAL, FP_ZERO, x)
#define isfinite(x) __builtin_isfinite(x)
+
#define isinf(x) __builtin_isinf(x)
+
#define isnan(x) __builtin_isnan(x)
+
#define isnormal(x) __builtin_isnormal(x)
+#define signbit(x) \
+ ((sizeof(x) == sizeof(float)) ? __builtin_signbitf(x) \
+ : (sizeof(x) == sizeof(double)) ? __builtin_signbit(x) \
+ : __builtin_signbitl(x))
+
+double acos(double __x);
+float acosf(float __x);
+long double acosl(long double __x) __RENAME_LDBL(acos, 3, 21);
+
+double asin(double __x);
+float asinf(float __x);
+long double asinl(long double __x) __RENAME_LDBL(asin, 3, 21);
+
+double atan(double __x);
+float atanf(float __x);
+long double atanl(long double __x) __RENAME_LDBL(atan, 3, 21);
+
+double atan2(double __y, double __x);
+float atan2f(float __y, float __x);
+long double atan2l(long double __y, long double __x) __RENAME_LDBL(atan2, 3, 21);
+
+double cos(double __x);
+float cosf(float __x);
+long double cosl(long double __x) __RENAME_LDBL(cos, 3, 21);
+
+double sin(double __x);
+float sinf(float __x);
+long double sinl(long double __x) __RENAME_LDBL(sin, 3, 21);
+
+double tan(double __x);
+float tanf(float __x);
+long double tanl(long double __x) __RENAME_LDBL(tan, 3, 21);
+
+double acosh(double __x);
+float acoshf(float __x);
+long double acoshl(long double __x) __RENAME_LDBL(acosh, 3, 21);
+
+double asinh(double __x);
+float asinhf(float __x);
+long double asinhl(long double __x) __RENAME_LDBL(asinh, 3, 21);
+
+double atanh(double __x);
+float atanhf(float __x);
+long double atanhl(long double __x) __RENAME_LDBL(atanh, 3, 21);
+
+double cosh(double __x);
+float coshf(float __x);
+long double coshl(long double __x) __RENAME_LDBL(cosh, 3, 21);
+
+double sinh(double __x);
+float sinhf(float __x);
+long double sinhl(long double __x) __RENAME_LDBL(sinh, 3, 21);
+
+double tanh(double __x);
+float tanhf(float __x);
+long double tanhl(long double __x) __RENAME_LDBL(tanh, 3, 21);
+
+double exp(double __x);
+float expf(float __x);
+long double expl(long double __x) __RENAME_LDBL(exp, 3, 21);
+
+double exp2(double __x);
+float exp2f(float __x);
+long double exp2l(long double __x) __RENAME_LDBL(exp2, 3, 21);
+
+double expm1(double __x);
+float expm1f(float __x);
+long double expm1l(long double __x) __RENAME_LDBL(expm1, 3, 21);
+
+double frexp(double __x, int* __exponent);
+float frexpf(float __x, int* __exponent);
+long double frexpl(long double __x, int* __exponent) __RENAME_LDBL(frexp, 3, 21);
+
+int ilogb(double __x) __attribute_const__;
+int ilogbf(float __x) __attribute_const__;
+int ilogbl(long double __x) __attribute_const__ __RENAME_LDBL(ilogb, 3, 3);
+
+double ldexp(double __x, int __exponent);
+float ldexpf(float __x, int __exponent);
+long double ldexpl(long double __x, int __exponent) __RENAME_LDBL(ldexp, 3, 3);
+
+double log(double __x);
+float logf(float __x);
+long double logl(long double __x) __RENAME_LDBL(log, 3, 21);
+
+double log10(double __x);
+float log10f(float __x);
+long double log10l(long double __x) __RENAME_LDBL(log10, 3, 21);
+
+double log1p(double __x);
+float log1pf(float __x);
+long double log1pl(long double __x) __RENAME_LDBL(log1p, 3, 21);
+
+double log2(double __x) __INTRODUCED_IN(18);
+float log2f(float __x) __INTRODUCED_IN(18);
+long double log2l(long double __x) __RENAME_LDBL(log2, 18, 18);
+
+double logb(double __x);
+float logbf(float __x);
+long double logbl(long double __x) __RENAME_LDBL(logb, 3, 18);
+
+double modf(double __x, double* __integral_part);
+float modff(float __x, float* __integral_part);
+long double modfl(long double __x, long double* __integral_part) __RENAME_LDBL(modf, 3, 21);
+
+double scalbn(double __x, int __exponent);
+float scalbnf(float __x, int __exponent);
+long double scalbnl(long double __x, int __exponent) __RENAME_LDBL(scalbn, 3, 3);
+
+/* TODO: once the NDK only supports >= 18, use __RENAME_LDBL here too. */
+double scalbln(double __x, long __exponent) __INTRODUCED_IN_X86(18) __VERSIONER_NO_GUARD;
+float scalblnf(float __x, long __exponent) __INTRODUCED_IN_X86(18) __VERSIONER_NO_GUARD;
+long double scalblnl(long double __x, long __exponent) __INTRODUCED_IN_X86(18) __VERSIONER_NO_GUARD;
+
+double cbrt(double __x);
+float cbrtf(float __x);
+long double cbrtl(long double __x) __RENAME_LDBL(cbrt, 3, 21);
+
+double fabs(double __x) __attribute_const__;
+float fabsf(float __x) __attribute_const__;
+long double fabsl(long double __x) __attribute_const__ __RENAME_LDBL(fabs, 3, 3);
+
+double hypot(double __x, double __y);
+float hypotf(float __x, float __y);
+long double hypotl(long double __x, long double __y) __RENAME_LDBL(hypot, 3, 21);
+
+double pow(double __x, double __y);
+float powf(float __x, float __y);
+long double powl(long double __x, long double __y) __RENAME_LDBL(pow, 3, 21);
+
+double sqrt(double __x);
+float sqrtf(float __x);
+long double sqrtl(long double __x) __RENAME_LDBL(sqrt, 3, 21);
+
+double erf(double __x);
+float erff(float __x);
+long double erfl(long double __x) __RENAME_LDBL(erf, 3, 21);
+
+double erfc(double __x);
+float erfcf(float __x);
+long double erfcl(long double __x) __RENAME_LDBL(erfc, 3, 21);
+
+double lgamma(double __x);
+float lgammaf(float __x);
+long double lgammal(long double __x) __RENAME_LDBL(lgamma, 3, 21);
+
+double tgamma(double __x);
+float tgammaf(float __x) __INTRODUCED_IN_ARM(13) __INTRODUCED_IN_MIPS(13) __INTRODUCED_IN_X86(9);
+long double tgammal(long double __x) __RENAME_LDBL(tgamma, 3, 21);
+
+double ceil(double __x);
+float ceilf(float __x);
+long double ceill(long double __x) __RENAME_LDBL(ceil, 3, 3);
+
+double floor(double __x);
+float floorf(float __x);
+long double floorl(long double __x) __RENAME_LDBL(floor, 3, 3);
+
+double nearbyint(double __x);
+float nearbyintf(float __x);
+long double nearbyintl(long double __x) __RENAME_LDBL(nearbyint, 3, 21);
+
+double rint(double __x);
+float rintf(float __x);
+long double rintl(long double __x) __RENAME_LDBL(rint, 3, 21);
+
+long lrint(double __x);
+long lrintf(float __x);
+long lrintl(long double __x) __RENAME_LDBL(lrint, 3, 21);
+
+long long llrint(double __x);
+long long llrintf(float __x);
+long long llrintl(long double __x) __RENAME_LDBL(llrint, 3, 21);
+
+double round(double __x);
+float roundf(float __x);
+long double roundl(long double __x) __RENAME_LDBL(roundl, 3, 3);
+
+long lround(double __x);
+long lroundf(float __x);
+long lroundl(long double __x) __RENAME_LDBL(lround, 3, 3);
+
+long long llround(double __x);
+long long llroundf(float __x);
+long long llroundl(long double __x) __RENAME_LDBL(llround, 3, 3);
+
+double trunc(double __x);
+float truncf(float __x);
+long double truncl(long double __x) __RENAME_LDBL(trunc, 3, 3);
+
+double fmod(double __x, double __y);
+float fmodf(float __x, float __y);
+long double fmodl(long double __x, long double __y) __RENAME_LDBL(fmod, 3, 21);
+
+double remainder(double __x, double __y);
+float remainderf(float __x, float __y);
+long double remainderl(long double __x, long double __y) __RENAME_LDBL(remainder, 3, 21);
+
+double remquo(double __x, double __y, int* __quotient_bits);
+float remquof(float __x, float __y, int* __quotient_bits);
+long double remquol(long double __x, long double __y, int* __quotient_bits) __RENAME_LDBL(remquo, 3, 21);
+
+double copysign(double __value, double __sign) __attribute_const__;
+float copysignf(float __value, float __sign) __attribute_const__;
+long double copysignl(long double __value, long double __sign) __attribute_const__ __RENAME_LDBL(copysign, 3, 3);
+
+double nan(const char* __kind) __attribute_const__ __INTRODUCED_IN_ARM(13) __INTRODUCED_IN_MIPS(13) __INTRODUCED_IN_X86(9);
+float nanf(const char* __kind) __attribute_const__ __INTRODUCED_IN_ARM(13) __INTRODUCED_IN_MIPS(13) __INTRODUCED_IN_X86(9);
+long double nanl(const char* __kind) __attribute_const__ __RENAME_LDBL(nan, 13, 13);
+
+double nextafter(double __x, double __y);
+float nextafterf(float __x, float __y);
+long double nextafterl(long double __x, long double __y) __RENAME_LDBL(nextafter, 3, 21) __VERSIONER_NO_GUARD;
+
+double nexttoward(double __x, long double __y) __INTRODUCED_IN(18) __VERSIONER_NO_GUARD;
+float nexttowardf(float __x, long double __y);
+long double nexttowardl(long double __x, long double __y) __RENAME_LDBL(nexttoward, 18, 18) __VERSIONER_NO_GUARD;
+
+double fdim(double __x, double __y);
+float fdimf(float __x, float __y);
+long double fdiml(long double __x, long double __y) __RENAME_LDBL(fdim, 3, 3);
+
+double fmax(double __x, double __y) __attribute_const__;
+float fmaxf(float __x, float __y) __attribute_const__;
+long double fmaxl(long double __x, long double __y) __attribute_const__ __RENAME_LDBL(fmax, 3, 3);
+
+double fmin(double __x, double __y) __attribute_const__;
+float fminf(float __x, float __y) __attribute_const__;
+long double fminl(long double __x, long double __y) __attribute_const__ __RENAME_LDBL(fmin, 3, 3);
+
+double fma(double __x, double __y, double __z);
+float fmaf(float __x, float __y, float __z);
+long double fmal(long double __x, long double __y, long double __z) __RENAME_LDBL(fma, 3, 21) __VERSIONER_NO_GUARD;
+
#define isgreater(x, y) __builtin_isgreater((x), (y))
#define isgreaterequal(x, y) __builtin_isgreaterequal((x), (y))
#define isless(x, y) __builtin_isless((x), (y))
@@ -67,92 +311,6 @@
#define islessgreater(x, y) __builtin_islessgreater((x), (y))
#define isunordered(x, y) __builtin_isunordered((x), (y))
-#define signbit(x) \
- ((sizeof(x) == sizeof(float)) ? __builtin_signbitf(x) \
- : (sizeof(x) == sizeof(double)) ? __builtin_signbit(x) \
- : __builtin_signbitl(x))
-
-typedef double __double_t;
-typedef __double_t double_t;
-typedef float __float_t;
-typedef __float_t float_t;
-
-#if defined(__USE_BSD)
-#define HUGE MAXFLOAT
-#endif
-
-extern int signgam;
-
-/*
- * Most of these functions depend on the rounding mode and have the side
- * effect of raising floating-point exceptions, so they are not declared
- * as __attribute_const__. In C99, FENV_ACCESS affects the purity of these functions.
- */
-
-int __fpclassifyd(double __x) __attribute_const__;
-int __fpclassifyf(float __x) __attribute_const__;
-int __fpclassifyl(long double __x) __attribute_const__;
-int __isfinitef(float __x) __attribute_const__;
-int __isfinite(double __x) __attribute_const__;
-int __isfinitel(long double __x) __attribute_const__;
-int __isinff(float __x) __attribute_const__;
-int __isinfl(long double __x) __attribute_const__;
-int __isnanf(float __x) __attribute_const__ __INTRODUCED_IN(21);
-int __isnanl(long double __x) __attribute_const__;
-int __isnormalf(float __x) __attribute_const__;
-int __isnormal(double __x) __attribute_const__;
-int __isnormall(long double __x) __attribute_const__;
-int __signbit(double __x) __attribute_const__;
-int __signbitf(float __x) __attribute_const__;
-int __signbitl(long double __x) __attribute_const__;
-
-double acos(double __x);
-double asin(double __x);
-double atan(double __x);
-double atan2(double __y, double __x);
-double cos(double __x);
-double sin(double __x);
-double tan(double __x);
-
-double cosh(double __x);
-double sinh(double __x);
-double tanh(double __x);
-
-double exp(double __x);
-double frexp(double __x, int* __exponent); /* fundamentally !__attribute_const__ */
-double ldexp(double __x, int __exponent);
-double log(double __x);
-double log10(double __x);
-double modf(double __x, double* __integral_part); /* fundamentally !__attribute_const__ */
-
-double pow(double __x, double __y);
-double sqrt(double __x);
-
-double ceil(double __x);
-double fabs(double __x) __attribute_const__;
-double floor(double __x);
-double fmod(double __x, double __y);
-
-double acosh(double __x);
-double asinh(double __x);
-double atanh(double __x);
-double cbrt(double __x);
-double erf(double __x);
-double erfc(double __x);
-double exp2(double __x);
-double expm1(double __x);
-double fma(double __x, double __y, double __z);
-double hypot(double __x, double __y);
-int ilogb(double __x) __attribute_const__;
-double lgamma(double __x);
-long long llrint(double __x);
-long long llround(double __x);
-double log1p(double __x);
-double log2(double __x) __INTRODUCED_IN(18);
-double logb(double __x);
-long lrint(double __x);
-long lround(double __x);
-
/*
* https://code.google.com/p/android/issues/detail?id=271629
* To be fully compliant with C++, we need to not define these (C doesn't
@@ -167,150 +325,9 @@
int (isinf)(double __x) __attribute_const__ __INTRODUCED_IN(21);
int (isnan)(double __x) __attribute_const__;
-double nan(const char* __kind) __attribute_const__ __INTRODUCED_IN_ARM(13) __INTRODUCED_IN_MIPS(13)
- __INTRODUCED_IN_X86(9);
+/* POSIX extensions. */
-double nextafter(double __x, double __y);
-double remainder(double __x, double __y);
-double remquo(double __x, double __y, int* __quotient_bits);
-double rint(double __x);
-
-double copysign(double __value, double __sign) __attribute_const__;
-double fdim(double __x, double __y);
-double fmax(double __x, double __y) __attribute_const__;
-double fmin(double __x, double __y) __attribute_const__;
-double nearbyint(double __x);
-double round(double __x);
-double scalbln(double __x, long __exponent) __INTRODUCED_IN_X86(18) __VERSIONER_NO_GUARD;
-double scalbn(double __x, int __exponent);
-double tgamma(double __x);
-double trunc(double __x);
-
-float acosf(float __x);
-float asinf(float __x);
-float atanf(float __x);
-float atan2f(float __y, float __x);
-float cosf(float __x);
-float sinf(float __x);
-float tanf(float __x);
-
-float coshf(float __x);
-float sinhf(float __x);
-float tanhf(float __x);
-
-float exp2f(float __x);
-float expf(float __x);
-float expm1f(float __x);
-float frexpf(float __x, int* __exponent); /* fundamentally !__attribute_const__ */
-int ilogbf(float __x) __attribute_const__;
-float ldexpf(float __x, int __exponent);
-float log10f(float __x);
-float log1pf(float __x);
-float log2f(float __x) __INTRODUCED_IN(18);
-float logf(float __x);
-float modff(float __x, float* __integral_part); /* fundamentally !__attribute_const__ */
-
-float powf(float __x, float __y);
-float sqrtf(float __x);
-
-float ceilf(float __x);
-float fabsf(float __x) __attribute_const__;
-float floorf(float __x);
-float fmodf(float __x, float __y);
-float roundf(float __x);
-
-float erff(float __x);
-float erfcf(float __x);
-float hypotf(float __x, float __y);
-float lgammaf(float __x);
-float tgammaf(float __x) __INTRODUCED_IN_ARM(13) __INTRODUCED_IN_MIPS(13) __INTRODUCED_IN_X86(9);
-
-float acoshf(float __x);
-float asinhf(float __x);
-float atanhf(float __x);
-float cbrtf(float __x);
-float logbf(float __x);
-float copysignf(float __value, float __sign) __attribute_const__;
-long long llrintf(float __x);
-long long llroundf(float __x);
-long lrintf(float __x);
-long lroundf(float __x);
-float nanf(const char* __kind) __attribute_const__ __INTRODUCED_IN_ARM(13) __INTRODUCED_IN_MIPS(13)
- __INTRODUCED_IN_X86(9);
-float nearbyintf(float __x);
-float nextafterf(float __x, float __y);
-float remainderf(float __x, float __y);
-float remquof(float __x, float __y, int* __quotient_bits);
-float rintf(float __x);
-float scalblnf(float __x, long __exponent) __INTRODUCED_IN_X86(18) __VERSIONER_NO_GUARD;
-float scalbnf(float __x, int __exponent);
-float truncf(float __x);
-
-float fdimf(float __x, float __y);
-float fmaf(float __x, float __y, float __z);
-float fmaxf(float __x, float __y) __attribute_const__;
-float fminf(float __x, float __y) __attribute_const__;
-
-long double acoshl(long double __x) __INTRODUCED_IN(21);
-long double acosl(long double __x) __INTRODUCED_IN(21);
-long double asinhl(long double __x) __INTRODUCED_IN(21);
-long double asinl(long double __x) __INTRODUCED_IN(21);
-long double atan2l(long double __y, long double __x) __INTRODUCED_IN(21);
-long double atanhl(long double __x) __INTRODUCED_IN(21);
-long double atanl(long double __x) __INTRODUCED_IN(21);
-long double cbrtl(long double __x) __INTRODUCED_IN(21);
-long double ceill(long double __x);
-long double copysignl(long double __value, long double __sign) __attribute_const__;
-long double coshl(long double __x) __INTRODUCED_IN(21);
-long double cosl(long double __x) __INTRODUCED_IN(21);
-long double erfcl(long double __x) __INTRODUCED_IN(21);
-long double erfl(long double __x) __INTRODUCED_IN(21);
-long double exp2l(long double __x) __INTRODUCED_IN(21);
-long double expl(long double __x) __INTRODUCED_IN(21);
-long double expm1l(long double __x) __INTRODUCED_IN(21);
-long double fabsl(long double __x) __attribute_const__;
-long double fdiml(long double __x, long double __y);
-long double floorl(long double __x);
-long double fmal(long double __x, long double __y, long double __z) __INTRODUCED_IN(21) __VERSIONER_NO_GUARD;
-long double fmaxl(long double __x, long double __y) __attribute_const__;
-long double fminl(long double __x, long double __y) __attribute_const__;
-long double fmodl(long double __x, long double __y) __INTRODUCED_IN(21);
-long double frexpl(long double __x, int* __exponent)
- __INTRODUCED_IN(21) __VERSIONER_NO_GUARD; /* fundamentally !__attribute_const__ */
-long double hypotl(long double __x, long double __y) __INTRODUCED_IN(21);
-int ilogbl(long double __x) __attribute_const__;
-long double ldexpl(long double __x, int __exponent);
-long double lgammal(long double __x) __INTRODUCED_IN(21);
-long long llrintl(long double __x) __INTRODUCED_IN(21);
-long long llroundl(long double __x);
-long double log10l(long double __x) __INTRODUCED_IN(21);
-long double log1pl(long double __x) __INTRODUCED_IN(21);
-long double log2l(long double __x) __INTRODUCED_IN(18);
-long double logbl(long double __x) __INTRODUCED_IN(18);
-long double logl(long double __x) __INTRODUCED_IN(21);
-long lrintl(long double __x) __INTRODUCED_IN(21);
-long lroundl(long double __x);
-long double modfl(long double __x, long double* __integral_part) __INTRODUCED_IN(21); /* fundamentally !__attribute_const__ */
-long double nanl(const char* __kind) __attribute_const__ __INTRODUCED_IN(13);
-long double nearbyintl(long double __x) __INTRODUCED_IN(21);
-long double nextafterl(long double __x, long double __y) __INTRODUCED_IN(21) __VERSIONER_NO_GUARD;
-double nexttoward(double __x, long double __y) __INTRODUCED_IN(18) __VERSIONER_NO_GUARD;
-float nexttowardf(float __x, long double __y);
-long double nexttowardl(long double __x, long double __y) __INTRODUCED_IN(18) __VERSIONER_NO_GUARD;
-long double powl(long double __x, long double __y) __INTRODUCED_IN(21);
-long double remainderl(long double __x, long double __y) __INTRODUCED_IN(21);
-long double remquol(long double __x, long double __y, int* __quotient_bits) __INTRODUCED_IN(21);
-long double rintl(long double __x) __INTRODUCED_IN(21);
-long double roundl(long double __x);
-long double scalblnl(long double __x, long __exponent) __INTRODUCED_IN_X86(18) __VERSIONER_NO_GUARD;
-long double scalbnl(long double __x, int __exponent);
-long double sinhl(long double __x) __INTRODUCED_IN(21);
-long double sinl(long double __x) __INTRODUCED_IN(21);
-long double sqrtl(long double __x) __INTRODUCED_IN(21);
-long double tanhl(long double __x) __INTRODUCED_IN(21);
-long double tanl(long double __x) __INTRODUCED_IN(21);
-long double tgammal(long double __x) __INTRODUCED_IN(21);
-long double truncl(long double __x);
+extern int signgam;
double j0(double __x);
double j1(double __x);
@@ -335,6 +352,38 @@
#define MAXFLOAT ((float)3.40282346638528860e+38)
+/* Legacy cruft from before we had builtin implementations of the standard macros. */
+
+int __fpclassifyd(double __x) __attribute_const__;
+int __fpclassifyf(float __x) __attribute_const__;
+int __fpclassifyl(long double __x) __attribute_const__;
+
+int __isfinitef(float __x) __attribute_const__;
+int __isfinite(double __x) __attribute_const__;
+int __isfinitel(long double __x) __attribute_const__;
+
+int __isinff(float __x) __attribute_const__;
+int __isinfl(long double __x) __attribute_const__;
+
+int __isnanf(float __x) __attribute_const__ __INTRODUCED_IN(21);
+int __isnanl(long double __x) __attribute_const__;
+
+int __isnormal(double __x) __attribute_const__;
+int __isnormalf(float __x) __attribute_const__;
+int __isnormall(long double __x) __attribute_const__;
+
+int __signbit(double __x) __attribute_const__;
+int __signbitf(float __x) __attribute_const__;
+int __signbitl(long double __x) __attribute_const__;
+
+/* BSD extensions. */
+
+#if defined(__USE_BSD)
+#define HUGE MAXFLOAT
+#endif
+
+/* Extensions in both BSD and GNU. */
+
#if defined(__USE_BSD) || defined(__USE_GNU)
double gamma(double __x);
double scalb(double __x, double __exponent);
@@ -361,6 +410,8 @@
float significandf(float __x);
#endif
+/* GNU extensions. */
+
#if defined(__USE_GNU)
#define M_El 2.718281828459045235360287471352662498L /* e */
#define M_LOG2El 1.442695040888963407359924681001892137L /* log 2e */
diff --git a/libc/include/stdlib.h b/libc/include/stdlib.h
index 4182f2d..86336a0 100644
--- a/libc/include/stdlib.h
+++ b/libc/include/stdlib.h
@@ -75,7 +75,7 @@
int posix_memalign(void** __memptr, size_t __alignment, size_t __size) __INTRODUCED_IN(16);
double strtod(const char* __s, char** __end_ptr);
-long double strtold(const char* __s, char** __end_ptr) __INTRODUCED_IN(21);
+long double strtold(const char* __s, char** __end_ptr) __RENAME_LDBL(strtod, 3, 21);
unsigned long strtoul_l(const char* __s, char** __end_ptr, int __base, locale_t __l) __INTRODUCED_IN(26);
diff --git a/libc/include/sys/cdefs.h b/libc/include/sys/cdefs.h
index 58eebc5..063b51d 100644
--- a/libc/include/sys/cdefs.h
+++ b/libc/include/sys/cdefs.h
@@ -37,6 +37,11 @@
#ifndef _SYS_CDEFS_H_
#define _SYS_CDEFS_H_
+#include <android/api-level.h>
+#include <android/versioning.h>
+
+#define __BIONIC__ 1
+
/*
* Testing against Clang-specific extensions.
*/
@@ -219,8 +224,21 @@
#define __RENAME_IF_FILE_OFFSET64(func)
#endif
-#define __BIONIC__ 1
-#include <android/api-level.h>
+/*
+ * For LP32, `long double` == `double`. Historically many `long double` functions were incorrect
+ * on x86, missing on most architectures, and even if they are present and correct, linking to
+ * them just bloats your ELF file by adding extra relocations. The __BIONIC_LP32_USE_LONG_DOUBLE
+ * macro lets us test the headers both ways (and adds an escape valve).
+ *
+ * Note that some functions have their __RENAME_LDBL commented out as a sign that although we could
+ * use __RENAME_LDBL it would actually cause the function to be introduced later because the
+ * `long double` variant appeared before the `double` variant.
+ */
+#if defined(__LP64__) || defined(__BIONIC_LP32_USE_LONG_DOUBLE)
+#define __RENAME_LDBL(rewrite,rewrite_api_level,regular_api_level) __INTRODUCED_IN(regular_api_level)
+#else
+#define __RENAME_LDBL(rewrite,rewrite_api_level,regular_api_level) __RENAME(rewrite) __INTRODUCED_IN(rewrite_api_level)
+#endif
/* glibc compatibility. */
#if defined(__LP64__)
@@ -334,8 +352,6 @@
/* Used to rename functions so that the compiler emits a call to 'x' rather than the function this was applied to. */
#define __RENAME(x) __asm__(#x)
-#include <android/versioning.h>
-
#if __has_builtin(__builtin_umul_overflow) || __GNUC__ >= 5
#if defined(__LP64__)
#define __size_mul_overflow(a, b, result) __builtin_umull_overflow(a, b, result)
diff --git a/libc/include/wchar.h b/libc/include/wchar.h
index 1dda4bd..1cf498c83 100644
--- a/libc/include/wchar.h
+++ b/libc/include/wchar.h
@@ -98,7 +98,7 @@
wchar_t* wcstok(wchar_t* __s, const wchar_t* __delimiter, wchar_t** __ptr);
long wcstol(const wchar_t* __s, wchar_t** __end_ptr, int __base);
long long wcstoll(const wchar_t* __s, wchar_t** __end_ptr, int __base) __INTRODUCED_IN(21);
-long double wcstold(const wchar_t* __s, wchar_t** __end_ptr) __INTRODUCED_IN(21);
+long double wcstold(const wchar_t* __s, wchar_t** __end_ptr) __RENAME_LDBL(wcstod, 3, 21);
unsigned long wcstoul(const wchar_t* __s, wchar_t** __end_ptr, int __base);
unsigned long long wcstoull(const wchar_t* __s, wchar_t** __end_ptr, int __base) __INTRODUCED_IN(21);
int wcswidth(const wchar_t* __s, size_t __n);