math: Fix float conversion regressions with gcc-12 [BZ #28713]

Converting double precision constants to float is now affected by the
runtime dynamic rounding mode instead of being evaluated at compile
time with default rounding mode (except static object initializers).

This can change the computed result and cause performance regression.
The known correctness issues (increased ulp errors) are already fixed,
this patch fixes remaining cases of unnecessary runtime conversions.

Add float M_* macros to math.h as new GNU extension API.  To avoid
conversions the new M_* macros are used and instead of casting double
literals to float, use float literals (only required if the conversion
is inexact).

The patch was tested on aarch64 where the following symbols had new
spurious conversion instructions that got fixed:

  __clog10f
  __gammaf_r_finite@GLIBC_2.17
  __j0f_finite@GLIBC_2.17
  __j1f_finite@GLIBC_2.17
  __jnf_finite@GLIBC_2.17
  __kernel_casinhf
  __lgamma_negf
  __log1pf
  __y0f_finite@GLIBC_2.17
  __y1f_finite@GLIBC_2.17
  cacosf
  cacoshf
  casinhf
  catanf
  catanhf
  clogf
  gammaf_positive

Fixes bug 28713.

Reviewed-by: Paul Zimmermann <Paul.Zimmermann@inria.fr>
This commit is contained in:
Szabolcs Nagy 2021-12-31 09:50:50 +00:00
parent e72ef23ee8
commit 347a5b592c
16 changed files with 54 additions and 32 deletions

5
NEWS
View File

@ -43,6 +43,11 @@ Major new features:
fminimum_mag, fminimum_mag_num and corresponding functions for float, fminimum_mag, fminimum_mag_num and corresponding functions for float,
long double, _FloatN and _FloatNx. long double, _FloatN and _FloatNx.
* <math.h> macros for single-precision float constants are added as a
GNU extension: M_Ef, M_LOG2Ef, M_LOG10Ef, M_LN2f, M_LN10f, M_PIf,
M_PI_2f, M_PI_4f, M_1_PIf, M_2_PIf, M_2_SQRTPIf, M_SQRT2f and
M_SQRT1_2f.
* The __STDC_IEC_60559_BFP__ and __STDC_IEC_60559_COMPLEX__ macros are * The __STDC_IEC_60559_BFP__ and __STDC_IEC_60559_COMPLEX__ macros are
predefined as specified in TS 18661-1:2014. predefined as specified in TS 18661-1:2014.

View File

@ -131,9 +131,10 @@ defined. The default set of features includes these constants.
@xref{Feature Test Macros}. @xref{Feature Test Macros}.
All values are of type @code{double}. As an extension, @theglibc{} All values are of type @code{double}. As an extension, @theglibc{}
also defines these constants with type @code{long double}. The also defines these constants with type @code{long double} and
@code{long double} macros have a lowercase @samp{l} appended to their @code{float}. The @code{long double} macros have a lowercase @samp{l}
names: @code{M_El}, @code{M_PIl}, and so forth. These are only while the @code{float} macros have a lowercase @samp{f} appended to
their names: @code{M_El}, @code{M_PIl}, and so forth. These are only
available if @code{_GNU_SOURCE} is defined. available if @code{_GNU_SOURCE} is defined.
Likewise, @theglibc{} also defines these constants with the types Likewise, @theglibc{} also defines these constants with the types

View File

@ -56,7 +56,7 @@ M_DECL_FUNC (__kernel_casinh) (CFLOAT x, int adj)
} }
res = M_SUF (__clog) (y); res = M_SUF (__clog) (y);
__real__ res += (FLOAT) M_MLIT (M_LN2); __real__ res += M_MLIT (M_LN2);
} }
else if (rx >= M_LIT (0.5) && ix < M_EPSILON / 8) else if (rx >= M_LIT (0.5) && ix < M_EPSILON / 8)
{ {

View File

@ -1158,6 +1158,23 @@ iszero (__T __val)
# define M_SQRT1_2 0.70710678118654752440 /* 1/sqrt(2) */ # define M_SQRT1_2 0.70710678118654752440 /* 1/sqrt(2) */
#endif #endif
/* GNU extension to provide float constants with similar names. */
#ifdef __USE_GNU
# define M_Ef 2.7182818284590452354f /* e */
# define M_LOG2Ef 1.4426950408889634074f /* log_2 e */
# define M_LOG10Ef 0.43429448190325182765f /* log_10 e */
# define M_LN2f 0.69314718055994530942f /* log_e 2 */
# define M_LN10f 2.30258509299404568402f /* log_e 10 */
# define M_PIf 3.14159265358979323846f /* pi */
# define M_PI_2f 1.57079632679489661923f /* pi/2 */
# define M_PI_4f 0.78539816339744830962f /* pi/4 */
# define M_1_PIf 0.31830988618379067154f /* 1/pi */
# define M_2_PIf 0.63661977236758134308f /* 2/pi */
# define M_2_SQRTPIf 1.12837916709551257390f /* 2/sqrt(pi) */
# define M_SQRT2f 1.41421356237309504880f /* sqrt(2) */
# define M_SQRT1_2f 0.70710678118654752440f /* 1/sqrt(2) */
#endif
/* The above constants are not adequate for computation using `long double's. /* The above constants are not adequate for computation using `long double's.
Therefore we provide as an extension constants with similar names as a Therefore we provide as an extension constants with similar names as a
GNU extension. Provide enough digits for the 128-bit IEEE quad. */ GNU extension. Provide enough digits for the 128-bit IEEE quad. */

View File

@ -32,7 +32,7 @@ M_DECL_FUNC (__cacos) (CFLOAT x)
{ {
y = M_SUF (__casin) (x); y = M_SUF (__casin) (x);
__real__ res = (FLOAT) M_MLIT (M_PI_2) - __real__ y; __real__ res = M_MLIT (M_PI_2) - __real__ y;
if (__real__ res == 0) if (__real__ res == 0)
__real__ res = 0; __real__ res = 0;
__imag__ res = -__imag__ y; __imag__ res = -__imag__ y;

View File

@ -106,7 +106,7 @@ M_DECL_FUNC (__catan) (CFLOAT x)
if (M_FABS (__imag__ x) == 1 if (M_FABS (__imag__ x) == 1
&& M_FABS (__real__ x) < M_EPSILON * M_EPSILON) && M_FABS (__real__ x) < M_EPSILON * M_EPSILON)
__imag__ res = (M_COPYSIGN (M_LIT (0.5), __imag__ x) __imag__ res = (M_COPYSIGN (M_LIT (0.5), __imag__ x)
* ((FLOAT) M_MLIT (M_LN2) * (M_MLIT (M_LN2)
- M_LOG (M_FABS (__real__ x)))); - M_LOG (M_FABS (__real__ x))));
else else
{ {

View File

@ -75,7 +75,7 @@ M_DECL_FUNC (__catanh) (CFLOAT x)
if (M_FABS (__real__ x) == 1 if (M_FABS (__real__ x) == 1
&& M_FABS (__imag__ x) < M_EPSILON * M_EPSILON) && M_FABS (__imag__ x) < M_EPSILON * M_EPSILON)
__real__ res = (M_COPYSIGN (M_LIT (0.5), __real__ x) __real__ res = (M_COPYSIGN (M_LIT (0.5), __real__ x)
* ((FLOAT) M_MLIT (M_LN2) * (M_MLIT (M_LN2)
- M_LOG (M_FABS (__imag__ x)))); - M_LOG (M_FABS (__imag__ x))));
else else
{ {

View File

@ -72,7 +72,7 @@ M_DECL_FUNC (__clog10) (CFLOAT x)
if (absx == 1 && scale == 0) if (absx == 1 && scale == 0)
{ {
__real__ result = (M_LOG1P (absy * absy) __real__ result = (M_LOG1P (absy * absy)
* ((FLOAT) M_MLIT (M_LOG10E) / 2)); * (M_MLIT (M_LOG10E) / 2));
math_check_force_underflow_nonneg (__real__ result); math_check_force_underflow_nonneg (__real__ result);
} }
else if (absx > 1 && absx < 2 && absy < 1 && scale == 0) else if (absx > 1 && absx < 2 && absy < 1 && scale == 0)
@ -80,7 +80,7 @@ M_DECL_FUNC (__clog10) (CFLOAT x)
FLOAT d2m1 = (absx - 1) * (absx + 1); FLOAT d2m1 = (absx - 1) * (absx + 1);
if (absy >= M_EPSILON) if (absy >= M_EPSILON)
d2m1 += absy * absy; d2m1 += absy * absy;
__real__ result = M_LOG1P (d2m1) * ((FLOAT) M_MLIT (M_LOG10E) / 2); __real__ result = M_LOG1P (d2m1) * (M_MLIT (M_LOG10E) / 2);
} }
else if (absx < 1 else if (absx < 1
&& absx >= M_LIT (0.5) && absx >= M_LIT (0.5)
@ -88,7 +88,7 @@ M_DECL_FUNC (__clog10) (CFLOAT x)
&& scale == 0) && scale == 0)
{ {
FLOAT d2m1 = (absx - 1) * (absx + 1); FLOAT d2m1 = (absx - 1) * (absx + 1);
__real__ result = M_LOG1P (d2m1) * ((FLOAT) M_MLIT (M_LOG10E) / 2); __real__ result = M_LOG1P (d2m1) * (M_MLIT (M_LOG10E) / 2);
} }
else if (absx < 1 else if (absx < 1
&& absx >= M_LIT (0.5) && absx >= M_LIT (0.5)
@ -96,7 +96,7 @@ M_DECL_FUNC (__clog10) (CFLOAT x)
&& absx * absx + absy * absy >= M_LIT (0.5)) && absx * absx + absy * absy >= M_LIT (0.5))
{ {
FLOAT d2m1 = M_SUF (__x2y2m1) (absx, absy); FLOAT d2m1 = M_SUF (__x2y2m1) (absx, absy);
__real__ result = M_LOG1P (d2m1) * ((FLOAT) M_MLIT (M_LOG10E) / 2); __real__ result = M_LOG1P (d2m1) * (M_MLIT (M_LOG10E) / 2);
} }
else else
{ {

View File

@ -32,7 +32,7 @@ M_DECL_FUNC (__clog) (CFLOAT x)
if (__glibc_unlikely (rcls == FP_ZERO && icls == FP_ZERO)) if (__glibc_unlikely (rcls == FP_ZERO && icls == FP_ZERO))
{ {
/* Real and imaginary part are 0.0. */ /* Real and imaginary part are 0.0. */
__imag__ result = signbit (__real__ x) ? (FLOAT) M_MLIT (M_PI) : 0; __imag__ result = signbit (__real__ x) ? M_MLIT (M_PI) : 0;
__imag__ result = M_COPYSIGN (__imag__ result, __imag__ x); __imag__ result = M_COPYSIGN (__imag__ result, __imag__ x);
/* Yes, the following line raises an exception. */ /* Yes, the following line raises an exception. */
__real__ result = -1 / M_FABS (__real__ x); __real__ result = -1 / M_FABS (__real__ x);
@ -94,7 +94,7 @@ M_DECL_FUNC (__clog) (CFLOAT x)
else else
{ {
FLOAT d = M_HYPOT (absx, absy); FLOAT d = M_HYPOT (absx, absy);
__real__ result = M_LOG (d) - scale * (FLOAT) M_MLIT (M_LN2); __real__ result = M_LOG (d) - scale * M_MLIT (M_LN2);
} }
__imag__ result = M_ATAN2 (__imag__ x, __real__ x); __imag__ result = M_ATAN2 (__imag__ x, __real__ x);

View File

@ -27,9 +27,8 @@
#define M_STRTO_NAN __strtof_nan #define M_STRTO_NAN __strtof_nan
#define M_USE_BUILTIN(c) USE_ ##c ##F_BUILTIN #define M_USE_BUILTIN(c) USE_ ##c ##F_BUILTIN
/* Standard/GNU macro literals do not exist for the float type. Use /* GNU extension float constant macros. */
the double macro constants. */ #define M_MLIT(c) c ## f
#define M_MLIT(c) c
#include <libm-alias-float.h> #include <libm-alias-float.h>
#include <math-nan-payload-float.h> #include <math-nan-payload-float.h>

View File

@ -85,7 +85,7 @@ gammaf_positive (float x, int *exp2_adj)
float x_adj_frac = x_adj - x_adj_int; float x_adj_frac = x_adj - x_adj_int;
int x_adj_log2; int x_adj_log2;
float x_adj_mant = __frexpf (x_adj, &x_adj_log2); float x_adj_mant = __frexpf (x_adj, &x_adj_log2);
if (x_adj_mant < (float) M_SQRT1_2) if (x_adj_mant < M_SQRT1_2f)
{ {
x_adj_log2--; x_adj_log2--;
x_adj_mant *= 2.0f; x_adj_mant *= 2.0f;
@ -94,7 +94,7 @@ gammaf_positive (float x, int *exp2_adj)
float ret = (__ieee754_powf (x_adj_mant, x_adj) float ret = (__ieee754_powf (x_adj_mant, x_adj)
* __ieee754_exp2f (x_adj_log2 * x_adj_frac) * __ieee754_exp2f (x_adj_log2 * x_adj_frac)
* __ieee754_expf (-x_adj) * __ieee754_expf (-x_adj)
* sqrtf (2 * (float) M_PI / x_adj) * sqrtf (2 * M_PIf / x_adj)
/ prod); / prod);
exp_adj += x_eps * __ieee754_logf (x_adj); exp_adj += x_eps * __ieee754_logf (x_adj);
float bsum = gamma_coeff[NCOEFF - 1]; float bsum = gamma_coeff[NCOEFF - 1];
@ -176,11 +176,11 @@ __ieee754_gammaf_r (float x, int *signgamp)
if (frac > 0.5f) if (frac > 0.5f)
frac = 1.0f - frac; frac = 1.0f - frac;
float sinpix = (frac <= 0.25f float sinpix = (frac <= 0.25f
? __sinf ((float) M_PI * frac) ? __sinf (M_PIf * frac)
: __cosf ((float) M_PI * (0.5f - frac))); : __cosf (M_PIf * (0.5f - frac)));
int exp2_adj; int exp2_adj;
float tret = (float) M_PI / (-x * sinpix float tret = M_PIf / (-x * sinpix
* gammaf_positive (-x, &exp2_adj)); * gammaf_positive (-x, &exp2_adj));
ret = __scalbnf (tret, -exp2_adj); ret = __scalbnf (tret, -exp2_adj);
math_check_force_underflow_nonneg (ret); math_check_force_underflow_nonneg (ret);
} }

View File

@ -233,7 +233,7 @@ j0f_near_root (float x, float z)
float index_f; float index_f;
int index; int index;
index_f = roundf ((x - FIRST_ZERO_J0) / (float) M_PI); index_f = roundf ((x - FIRST_ZERO_J0) / M_PIf);
/* j0f_asympt fails to give an error <= 9 ulps for x=0x1.324e92p+7 /* j0f_asympt fails to give an error <= 9 ulps for x=0x1.324e92p+7
(index 48) thus we can't reduce SMALL_SIZE below 49. */ (index 48) thus we can't reduce SMALL_SIZE below 49. */
if (index_f >= SMALL_SIZE) if (index_f >= SMALL_SIZE)
@ -514,7 +514,7 @@ y0f_near_root (float x, float z)
float index_f; float index_f;
int index; int index;
index_f = roundf ((x - FIRST_ZERO_Y0) / (float) M_PI); index_f = roundf ((x - FIRST_ZERO_Y0) / M_PIf);
if (index_f >= SMALL_SIZE) if (index_f >= SMALL_SIZE)
return y0f_asympt (x); return y0f_asympt (x);
index = (int) index_f; index = (int) index_f;

View File

@ -243,7 +243,7 @@ j1f_near_root (float x, float z)
x = -x; x = -x;
sign = -1.0f; sign = -1.0f;
} }
index_f = roundf ((x - FIRST_ZERO_J1) / (float) M_PI); index_f = roundf ((x - FIRST_ZERO_J1) / M_PIf);
if (index_f >= SMALL_SIZE) if (index_f >= SMALL_SIZE)
return sign * j1f_asympt (x); return sign * j1f_asympt (x);
index = (int) index_f; index = (int) index_f;
@ -525,7 +525,7 @@ y1f_near_root (float x, float z)
float index_f; float index_f;
int index; int index;
index_f = roundf ((x - FIRST_ZERO_Y1) / (float) M_PI); index_f = roundf ((x - FIRST_ZERO_Y1) / M_PIf);
if (index_f >= SMALL_SIZE) if (index_f >= SMALL_SIZE)
return y1f_asympt (x); return y1f_asympt (x);
index = (int) index_f; index = (int) index_f;

View File

@ -134,7 +134,7 @@ __ieee754_jnf(int n, float x)
tmp = n; tmp = n;
v = two/x; v = two/x;
tmp = tmp*__ieee754_logf(fabsf(v*tmp)); tmp = tmp*__ieee754_logf(fabsf(v*tmp));
if(tmp<(float)8.8721679688e+01) { if(tmp<8.8721679688e+01f) {
for(i=n-1,di=(float)(i+i);i>0;i--){ for(i=n-1,di=(float)(i+i);i>0;i--){
temp = b; temp = b;
b *= di; b *= di;

View File

@ -165,9 +165,9 @@ static float
lg_sinpi (float x) lg_sinpi (float x)
{ {
if (x <= 0.25f) if (x <= 0.25f)
return __sinf ((float) M_PI * x); return __sinf (M_PIf * x);
else else
return __cosf ((float) M_PI * (0.5f - x)); return __cosf (M_PIf * (0.5f - x));
} }
/* Compute cos (pi * X) for -0.25 <= X <= 0.5. */ /* Compute cos (pi * X) for -0.25 <= X <= 0.5. */
@ -176,9 +176,9 @@ static float
lg_cospi (float x) lg_cospi (float x)
{ {
if (x <= 0.25f) if (x <= 0.25f)
return __cosf ((float) M_PI * x); return __cosf (M_PIf * x);
else else
return __sinf ((float) M_PI * (0.5f - x)); return __sinf (M_PIf * (0.5f - x));
} }
/* Compute cot (pi * X) for -0.25 <= X <= 0.5. */ /* Compute cot (pi * X) for -0.25 <= X <= 0.5. */

View File

@ -92,7 +92,7 @@ __log1pf(float x)
if(k==0) return zero; if(k==0) return zero;
else {c += k*ln2_lo; return k*ln2_hi+c;} else {c += k*ln2_lo; return k*ln2_hi+c;}
} }
R = hfsq*((float)1.0-(float)0.66666666666666666*f); R = hfsq*(1.0f-0.66666666666666666f*f);
if(k==0) return f-R; else if(k==0) return f-R; else
return k*ln2_hi-((R-(k*ln2_lo+c))-f); return k*ln2_hi-((R-(k*ln2_lo+c))-f);
} }