Changeset e43d658 in mainline

Timestamp:

2018-08-29T20:05:30Z (6 years ago)

Author:

Jiří Zárevúcky <jiri.zarevucky@…>

Children:

b1b201f

Parents:

aab85d90

git-author:

Jiří Zárevúcky <jiri.zarevucky@…> (2018-08-29 15:32:57)

git-committer:

Jiří Zárevúcky <jiri.zarevucky@…> (2018-08-29 20:05:30)

Message:

Remove arch-specific parts of libmath

These are just performance optimizations and also currently using just the obsolete slow x87 instructions on x86 platforms. These also hide the brokenness of the soft-implementations on the most used platforms. In principle, complicating the build with optimizations of a broken library doesn't make sense.

Location:

uspace/lib/math

Files:

• Makefile (modified) (2 diffs)
• arch/abs32le/Makefile.inc (deleted)
• arch/abs32le/include/libarch/math.h (deleted)
• arch/amd64/Makefile.inc (deleted)
• arch/amd64/include/libarch/math.h (deleted)
• arch/amd64/include/libarch/x87.h (deleted)
• arch/amd64/src/cos.S (deleted)
• arch/amd64/src/sin.S (deleted)
• arch/amd64/src/trunc.S (deleted)
• arch/arm32/Makefile.inc (deleted)
• arch/arm32/include/libarch/math.h (deleted)
• arch/ia32/Makefile.inc (deleted)
• arch/ia32/include/libarch/math.h (deleted)
• arch/ia32/include/libarch/x87.h (deleted)
• arch/ia32/src/cos.S (deleted)
• arch/ia32/src/sin.S (deleted)
• arch/ia32/src/trunc.S (deleted)
• arch/ia64/Makefile.inc (deleted)
• arch/ia64/include/libarch/math.h (deleted)
• arch/mips32/Makefile.inc (deleted)
• arch/mips32/include/libarch/math.h (deleted)
• arch/mips32eb/Makefile.inc (deleted)
• arch/mips32eb/include/libarch/math.h (deleted)
• arch/ppc32/Makefile.inc (deleted)
• arch/ppc32/include/libarch/math.h (deleted)
• arch/riscv64/Makefile.inc (deleted)
• arch/riscv64/include/libarch/math.h (deleted)
• arch/sparc64/Makefile.inc (deleted)
• arch/sparc64/include/libarch/math.h (deleted)
• include/math.h (modified) (1 diff)

uspace/lib/math/Makefile

@@ -30 +30 @@
 ROOT_PATH = $(USPACE_PREFIX)/..
 
-CONFIG_MAKEFILE = $(ROOT_PATH)/Makefile.config
-
 LIBRARY = libmath
 SOVERSION = 0.0
 
-EXTRA_CFLAGS += -Iarch/$(UARCH)/include
-
--include $(CONFIG_MAKEFILE)
--include arch/$(UARCH)/Makefile.inc
-
-GENERIC_SOURCES = \
+SOURCES = \
         generic/acos.c \
         generic/asin.c \
@@ -65 +58 @@
         generic/trunc.c
 
-SOURCES = \
-        $(GENERIC_SOURCES) \
-        $(ARCH_SOURCES)
-
 include $(USPACE_PREFIX)/Makefile.common

uspace/lib/math/include/math.h

@@ -37 +37 @@
 #define LIBMATH_MATH_H_
 
-#include <libarch/math.h>
+#include <acos.h>
+#include <asin.h>
+#include <atan.h>
+#include <atan2.h>
+#include <ceil.h>
+#include <cosh.h>
+#include <exp.h>
+#include <fabs.h>
+#include <floor.h>
+#include <fmod.h>
+#include <frexp.h>
+#include <ldexp.h>
+#include <log.h>
+#include <log10.h>
+#include <log2.h>
+#include <mathtypes.h>
+#include <modf.h>
+#include <pow.h>
+#include <sinh.h>
+#include <sqrt.h>
+#include <tan.h>
+#include <tanh.h>
+#include <trig.h>
+#include <trunc.h>
+
+#define HUGE_VAL FLOAT64_INF
+
+static inline float64_t acos_f64(float64_t val)
+{
+        return float64_acos(val);
+}
+
+static inline float32_t acos_f32(float32_t val)
+{
+        return float32_acos(val);
+}
+
+static inline float64_t asin_f64(float64_t val)
+{
+        return float64_asin(val);
+}
+
+static inline float32_t asin_f32(float32_t val)
+{
+        return float32_asin(val);
+}
+
+static inline float64_t atan_f64(float64_t val)
+{
+        return float64_atan(val);
+}
+
+static inline float32_t atan_f32(float32_t val)
+{
+        return float32_atan(val);
+}
+
+static inline float64_t atan2_f64(float64_t y, float64_t x)
+{
+        return float64_atan2(y, x);
+}
+
+static inline float32_t atan2_f32(float32_t y, float32_t x)
+{
+        return float32_atan2(y, x);
+}
+
+static inline float64_t ceil_f64(float64_t val)
+{
+        return float64_ceil(val);
+}
+
+static inline float32_t ceil_f32(float32_t val)
+{
+        return float32_ceil(val);
+}
+
+static inline float64_t cos_f64(float64_t val)
+{
+        return float64_cos(val);
+}
+
+static inline float32_t cos_f32(float32_t val)
+{
+        return float32_cos(val);
+}
+
+static inline float64_t cosh_f64(float64_t val)
+{
+        return float64_cosh(val);
+}
+
+static inline float32_t cosh_f32(float32_t val)
+{
+        return float32_cosh(val);
+}
+
+static inline float64_t exp_f64(float64_t val)
+{
+        return float64_exp(val);
+}
+
+static inline float32_t exp_f32(float32_t val)
+{
+        return float32_exp(val);
+}
+
+static inline float64_t fabs_f64(float64_t val)
+{
+        return float64_fabs(val);
+}
+
+static inline float32_t fabs_f32(float32_t val)
+{
+        return float32_fabs(val);
+}
+
+static inline float64_t floor_f64(float64_t val)
+{
+        return float64_floor(val);
+}
+
+static inline float32_t floor_f32(float32_t val)
+{
+        return float32_floor(val);
+}
+static inline float64_t fmod_f64(float64_t dividend, float64_t divisor)
+{
+        return float64_fmod(dividend, divisor);
+}
+
+static inline float64_t fmod_f32(float32_t dividend, float32_t divisor)
+{
+        return float32_fmod(dividend, divisor);
+}
+
+static inline float64_t frexp_f64(float64_t x, int *exp)
+{
+        return float64_frexp(x, exp);
+}
+
+static inline float64_t frexp_f32(float32_t x, int *exp)
+{
+        return float32_frexp(x, exp);
+}
+
+static inline float64_t ldexp_f64(float64_t x, int exp)
+{
+        return float64_ldexp(x, exp);
+}
+
+static inline float64_t ldexp_f32(float32_t x, int exp)
+{
+        return float32_ldexp(x, exp);
+}
+
+static inline float64_t log_f64(float64_t val)
+{
+        return float64_log(val);
+}
+
+static inline float32_t log_f32(float32_t val)
+{
+        return float32_log(val);
+}
+
+static inline float64_t log10_f64(float64_t val)
+{
+        return float64_log10(val);
+}
+
+static inline float32_t log10_f32(float32_t val)
+{
+        return float32_log10(val);
+}
+
+static inline float32_t log2_f32(float32_t val)
+{
+        return float32_log2(val);
+}
+
+static inline float64_t log2_f64(float64_t val)
+{
+        return float64_log2(val);
+}
+
+static inline float64_t modf_f64(float64_t value, float64_t *iptr)
+{
+        return float64_modf(value, iptr);
+}
+
+static inline float64_t modf_f32(float32_t value, float32_t *iptr)
+{
+        return float32_modf(value, iptr);
+}
+
+static inline float64_t pow_f64(float64_t x, float64_t y)
+{
+        return float64_pow(x, y);
+}
+
+static inline float32_t pow_f32(float32_t x, float32_t y)
+{
+        return float32_pow(x, y);
+}
+
+static inline float64_t sin_f64(float64_t val)
+{
+        return float64_sin(val);
+}
+
+static inline float32_t sin_f32(float32_t val)
+{
+        return float32_sin(val);
+}
+
+static inline float64_t sinh_f64(float64_t val)
+{
+        return float64_sinh(val);
+}
+
+static inline float32_t sinh_f32(float32_t val)
+{
+        return float32_sinh(val);
+}
+
+static inline float64_t sqrt_f64(float64_t val)
+{
+        return float64_sqrt(val);
+}
+
+static inline float32_t sqrt_f32(float32_t val)
+{
+        return float32_sqrt(val);
+}
+
+static inline float64_t tan_f64(float64_t val)
+{
+        return float64_tan(val);
+}
+
+static inline float32_t tan_f32(float32_t val)
+{
+        return float32_tan(val);
+}
+
+static inline float64_t tanh_f64(float64_t val)
+{
+        return float64_tanh(val);
+}
+
+static inline float32_t tanh_f32(float32_t val)
+{
+        return float32_tanh(val);
+}
+
+static inline float64_t trunc_f64(float64_t val)
+{
+        return float64_trunc(val);
+}
+
+static inline float32_t trunc_f32(float32_t val)
+{
+        return float32_trunc(val);
+}
+
+static inline float64_t acos(float64_t val)
+{
+        return acos_f64(val);
+}
+
+static inline float32_t acosf(float32_t val)
+{
+        return acos_f32(val);
+}
+
+static inline float64_t asin(float64_t val)
+{
+        return asin_f64(val);
+}
+
+static inline float32_t asinf(float32_t val)
+{
+        return asin_f32(val);
+}
+
+static inline float64_t atan(float64_t val)
+{
+        return atan_f64(val);
+}
+
+static inline float32_t atanf(float32_t val)
+{
+        return atan_f32(val);
+}
+
+static inline float64_t atan2(float64_t y, float64_t x)
+{
+        return atan2_f64(y, x);
+}
+
+static inline float32_t atan2f(float32_t y, float32_t x)
+{
+        return atan2_f32(y, x);
+}
+
+static inline float64_t ceil(float64_t val)
+{
+        return ceil_f64(val);
+}
+
+static inline float32_t ceilf(float32_t val)
+{
+        return ceil_f32(val);
+}
+
+static inline float64_t cos(float64_t val)
+{
+        return cos_f64(val);
+}
+
+static inline float32_t cosf(float32_t val)
+{
+        return cos_f32(val);
+}
+
+static inline float64_t cosh(float64_t val)
+{
+        return cosh_f64(val);
+}
+
+static inline float32_t coshf(float32_t val)
+{
+        return cosh_f32(val);
+}
+
+static inline float64_t exp(float64_t val)
+{
+        return exp_f64(val);
+}
+
+static inline float32_t expf(float32_t val)
+{
+        return exp_f32(val);
+}
+
+static inline float64_t fabs(float64_t val)
+{
+        return fabs_f64(val);
+}
+
+static inline float32_t fabsf(float32_t val)
+{
+        return fabs_f32(val);
+}
+
+static inline float64_t floor(float64_t val)
+{
+        return floor_f64(val);
+}
+
+static inline float32_t floorf(float32_t val)
+{
+        return floor_f32(val);
+}
+
+static inline float64_t fmod(float64_t dividend, float64_t divisor)
+{
+        return fmod_f64(dividend, divisor);
+}
+
+static inline float32_t fmodf(float32_t dividend, float32_t divisor)
+{
+        return fmod_f32(dividend, divisor);
+}
+
+static inline float64_t frexp(float64_t x, int *exp)
+{
+        return frexp_f64(x, exp);
+}
+
+static inline float32_t frexpf(float32_t x, int *exp)
+{
+        return frexp_f32(x, exp);
+}
+
+static inline float64_t ldexp(float64_t x, int exp)
+{
+        return ldexp_f64(x, exp);
+}
+
+static inline float32_t ldexpf(float32_t x, int exp)
+{
+        return ldexp_f32(x, exp);
+}
+
+static inline float64_t log(float64_t val)
+{
+        return log_f64(val);
+}
+
+static inline float32_t logf(float32_t val)
+{
+        return log_f32(val);
+}
+
+static inline float64_t log10(float64_t val)
+{
+        return log10_f64(val);
+}
+
+static inline float32_t log10f(float32_t val)
+{
+        return log10_f32(val);
+}
+
+static inline float64_t log2(float64_t val)
+{
+        return log2_f64(val);
+}
+
+static inline float32_t log2f(float32_t val)
+{
+        return log2_f32(val);
+}
+
+static inline float64_t modf(float64_t value, float64_t *iptr)
+{
+        return modf_f64(value, iptr);
+}
+
+static inline float32_t modff(float32_t value, float32_t *iptr)
+{
+        return modf_f32(value, iptr);
+}
+
+static inline float64_t pow(float64_t x, float64_t y)
+{
+        return pow_f64(x, y);
+}
+
+static inline float32_t powf(float32_t x, float32_t y)
+{
+        return pow_f32(x, y);
+}
+
+static inline float64_t sin(float64_t val)
+{
+        return sin_f64(val);
+}
+
+static inline float32_t sinf(float32_t val)
+{
+        return sin_f32(val);
+}
+
+static inline float64_t sinh(float64_t val)
+{
+        return sinh_f64(val);
+}
+
+static inline float32_t sinhf(float32_t val)
+{
+        return sinh_f32(val);
+}
+
+static inline float64_t sqrt(float64_t val)
+{
+        return sqrt_f64(val);
+}
+
+static inline float32_t sqrtf(float32_t val)
+{
+        return sqrt_f32(val);
+}
+
+static inline float64_t tan(float64_t val)
+{
+        return tan_f64(val);
+}
+
+static inline float32_t tanf(float32_t val)
+{
+        return tan_f32(val);
+}
+
+static inline float64_t tanh(float64_t val)
+{
+        return tanh_f64(val);
+}
+
+static inline float32_t tanhf(float32_t val)
+{
+        return tanh_f32(val);
+}
+
+static inline float64_t trunc(float64_t val)
+{
+        return trunc_f64(val);
+}
+
+static inline float32_t truncf(float32_t val)
+{
+        return trunc_f32(val);
+}
 
 #define M_LN10 2.30258509299404568402