Szabolcs Nagy 8c44a06024 fix scalbn when result is in the subnormal range ...

in nearest rounding mode scalbn could introduce double rounding error
when an intermediate value and the final result were both in the
subnormal range e.g.

  scalbn(0x1.7ffffffffffffp-1, -1073)

returned 0x1p-1073 instead of 0x1p-1074, because the intermediate
computation got rounded to 0x1.8p-1023.

with the fix an intermediate value can only be in the subnormal range
if the final result is 0 which is correct even after double rounding.
(there still can be two roundings so signals may be raised twice, but
that's only observable with trapping exceptions which is not supported.)

2017-04-21 17:31:00 -04:00

..

aarch64

aarch64: add single instruction math functions

2017-03-21 12:39:18 -04:00

arm

work around regression building for armhf with clang (compiler bug)

2016-02-19 01:20:07 +00:00

i386

remove the last of possible-textrels from i386 asm

2015-04-18 20:45:39 -04:00

x32

fix regression in x86_64 math asm with old binutils

2015-04-23 06:21:49 -04:00

x86_64

fix regression in x86_64 math asm with old binutils

2015-04-23 06:21:49 -04:00

__cos.c

math: use double_t for temporaries to avoid stores on i386

2013-05-15 23:08:52 +00:00

__cosdf.c

math: use double_t for temporaries to avoid stores on i386

2013-05-15 23:08:52 +00:00

__cosl.c

math: long double trigonometric cleanup (cosl, sinl, sincosl, tanl)

2013-09-05 11:30:08 +00:00

__expo2.c

clean up __expo2.c, use a slightly better k constant

2012-03-13 16:38:22 +01:00

__expo2f.c

clean up __expo2.c, use a slightly better k constant

2012-03-13 16:38:22 +01:00

__fpclassify.c

math: cosmetic cleanup (use explicit union instead of fshape and dshape)

2013-09-05 11:30:09 +00:00

__fpclassifyf.c

math: cosmetic cleanup (use explicit union instead of fshape and dshape)

2013-09-05 11:30:09 +00:00

__fpclassifyl.c

math: fix __fpclassifyl(-0.0) for IEEE binary128

2015-02-08 17:41:56 +01:00

__invtrigl.c

math: fix 128bit long double inverse trigonometric functions

2016-08-30 16:00:38 -04:00

__invtrigl.h

math: long double inverse trigonometric cleanup (acosl, asinl, atanl, atan2l)

2013-09-05 11:30:07 +00:00

__polevll.c

prune math code on archs with binary64 long double

2014-09-08 15:51:42 +02:00

__rem_pio2_large.c

math: clean up __rem_pio2

2013-11-24 01:06:38 +00:00

__rem_pio2.c

math: explicitly promote expressions to excess-precision types

2015-11-21 21:41:42 -05:00

__rem_pio2f.c

math: explicitly promote expressions to excess-precision types

2015-11-21 21:41:42 -05:00

__rem_pio2l.c

math: use the rounding idiom consistently

2014-10-31 11:35:40 -04:00

__signbit.c

remove a fixme comment

2012-03-20 20:08:35 +01:00

__signbitf.c

remove a fixme comment

2012-03-20 20:08:35 +01:00

__signbitl.c

always provide __fpclassifyl and __signbitl definitions

2014-10-08 10:17:19 -04:00

__sin.c

math: use double_t for temporaries to avoid stores on i386

2013-05-15 23:08:52 +00:00

__sindf.c

math: use double_t for temporaries to avoid stores on i386

2013-05-15 23:08:52 +00:00

__sinl.c

math: long double trigonometric cleanup (cosl, sinl, sincosl, tanl)

2013-09-05 11:30:08 +00:00

__tan.c

math: tan cleanups

2013-05-18 12:34:00 +00:00

__tandf.c

math: tan cleanups

2013-05-18 12:34:00 +00:00

__tanl.c

math: long double trigonometric cleanup (cosl, sinl, sincosl, tanl)

2013-09-05 11:30:08 +00:00

acos.c

math: use double_t for temporaries to avoid stores on i386

2013-05-15 23:08:52 +00:00

acosf.c

math: use double_t for temporaries to avoid stores on i386

2013-05-15 23:08:52 +00:00

acosh.c

math: use sqrtl if FLT_EVAL_METHOD==2 in acosh and acoshf

2013-10-07 18:41:03 +00:00

acoshf.c

math: use sqrtl if FLT_EVAL_METHOD==2 in acosh and acoshf

2013-10-07 18:41:03 +00:00

acoshl.c

math: add dummy implementations of 128 bit long double functions

2015-03-11 18:54:53 -04:00

acosl.c

math: long double inverse trigonometric cleanup (acosl, asinl, atanl, atan2l)

2013-09-05 11:30:07 +00:00

asin.c

math: fix asin, atan, log1p, tanh to raise underflow on subnormal

2013-08-15 10:14:46 +00:00

asinf.c

math: fix asin, atan, log1p, tanh to raise underflow on subnormal

2013-08-15 10:14:46 +00:00

asinh.c

math: use 0x1p-120f and 0x1p120f for tiny and huge values

2012-12-16 20:28:43 +01:00

asinhf.c

math: rewrite inverse hyperbolic functions to be simpler/smaller

2012-12-11 23:06:20 +01:00

asinhl.c

math: add dummy implementations of 128 bit long double functions

2015-03-11 18:54:53 -04:00

asinl.c

math: long double inverse trigonometric cleanup (acosl, asinl, atanl, atan2l)

2013-09-05 11:30:07 +00:00

atan2.c

math: clean up atan2.c

2013-08-15 14:05:19 +00:00

atan2f.c

math: clean up atan2.c

2013-08-15 14:05:19 +00:00

atan2l.c

math: long double inverse trigonometric cleanup (acosl, asinl, atanl, atan2l)

2013-09-05 11:30:07 +00:00

atan.c

math: fix asin, atan, log1p, tanh to raise underflow on subnormal

2013-08-15 10:14:46 +00:00

atanf.c

math: fix asin, atan, log1p, tanh to raise underflow on subnormal

2013-08-15 10:14:46 +00:00

atanh.c

math: fix atanh (overflow and underflow issues)

2013-09-05 16:57:46 +00:00

atanhf.c

math: fix atanh (overflow and underflow issues)

2013-09-05 16:57:46 +00:00

atanhl.c

math: add dummy implementations of 128 bit long double functions

2015-03-11 18:54:53 -04:00

atanl.c

math: long double inverse trigonometric cleanup (acosl, asinl, atanl, atan2l)

2013-09-05 11:30:07 +00:00

cbrt.c

math: cbrt cleanup and long double fix

2013-09-05 11:30:08 +00:00

cbrtf.c

math: cbrt cleanup and long double fix

2013-09-05 11:30:08 +00:00

cbrtl.c

math: cbrt cleanup and long double fix

2013-09-05 11:30:08 +00:00

ceil.c

math: use the rounding idiom consistently

2014-10-31 11:35:40 -04:00

ceilf.c

math: rewrite rounding functions (ceil, floor, trunc, round, rint)

2013-09-05 11:30:07 +00:00

ceill.c

math: use the rounding idiom consistently

2014-10-31 11:35:40 -04:00

copysign.c

math: cosmetic cleanup (use explicit union instead of fshape and dshape)

2013-09-05 11:30:09 +00:00

copysignf.c

math: cosmetic cleanup (use explicit union instead of fshape and dshape)

2013-09-05 11:30:09 +00:00

copysignl.c

long double cleanup, initial commit

2013-09-05 11:30:07 +00:00

cos.c

math: sin cos cleanup

2013-05-18 14:40:22 +00:00

cosf.c

math: sin cos cleanup

2013-05-18 14:40:22 +00:00

cosh.c

math: finished cosh.c cleanup

2012-12-16 19:23:51 +01:00

coshf.c

math: finished cosh.c cleanup

2012-12-16 19:23:51 +01:00

coshl.c

math: add dummy implementations of 128 bit long double functions

2015-03-11 18:54:53 -04:00

cosl.c

math: long double trigonometric cleanup (cosl, sinl, sincosl, tanl)

2013-09-05 11:30:08 +00:00

erf.c

math: use double_t for temporaries to avoid stores on i386

2013-05-15 23:08:52 +00:00

erff.c

math: use double_t for temporaries to avoid stores on i386

2013-05-15 23:08:52 +00:00

erfl.c

math: add dummy implementations of 128 bit long double functions

2015-03-11 18:54:53 -04:00

exp2.c

math: remove STRICT_ASSIGN macro

2013-09-06 18:35:55 +00:00

exp2f.c

math: fix expf(-NAN) and exp2f(-NAN) to return -NAN instead of 0

2016-03-04 17:58:49 -05:00

exp2l.c

math: add ld128 exp2l based on the freebsd implementation

2015-03-11 18:54:50 -04:00

exp10.c

math: fix exp10 not to raise invalid exception on NaN

2014-09-18 17:52:50 +02:00

exp10f.c

math: fix exp10 not to raise invalid exception on NaN

2014-09-18 17:52:50 +02:00

exp10l.c

math: fix exp10 not to raise invalid exception on NaN

2014-09-18 17:52:50 +02:00

exp.c

math: remove STRICT_ASSIGN macro

2013-09-06 18:35:55 +00:00

expf.c

math: fix expf(-NAN) and exp2f(-NAN) to return -NAN instead of 0

2016-03-04 17:58:49 -05:00

expl.c

math: add dummy implementations of 128 bit long double functions

2015-03-11 18:54:53 -04:00

expm1.c

math: remove STRICT_ASSIGN macro

2013-09-06 18:35:55 +00:00

expm1f.c

math: remove STRICT_ASSIGN macro

2013-09-06 18:35:55 +00:00

expm1l.c

math: add dummy implementations of 128 bit long double functions

2015-03-11 18:54:53 -04:00

fabs.c

math: cosmetic cleanup (use explicit union instead of fshape and dshape)

2013-09-05 11:30:09 +00:00

fabsf.c

math: cosmetic cleanup (use explicit union instead of fshape and dshape)

2013-09-05 11:30:09 +00:00

fabsl.c

long double cleanup, initial commit

2013-09-05 11:30:07 +00:00

fdim.c

remove libm.h includes when math.h and float.h are enough

2012-03-13 19:51:14 +01:00

fdimf.c

remove libm.h includes when math.h and float.h are enough

2012-03-13 19:51:14 +01:00

fdiml.c

remove libm.h includes when math.h and float.h are enough

2012-03-13 19:51:14 +01:00

finite.c

math: add (obsolete) bsd drem and finite functions

2013-11-21 01:16:49 +00:00

finitef.c

math: add (obsolete) bsd drem and finite functions

2013-11-21 01:16:49 +00:00

floor.c

math: use the rounding idiom consistently

2014-10-31 11:35:40 -04:00

floorf.c

math: rewrite rounding functions (ceil, floor, trunc, round, rint)

2013-09-05 11:30:07 +00:00

floorl.c

math: use the rounding idiom consistently

2014-10-31 11:35:40 -04:00

fma.c

add parenthesis in fma.c to clarify intent and silence warnings

2015-02-08 17:40:54 +01:00

fmaf.c

math: fix rare underflow issue in fma

2013-10-07 18:46:43 +00:00

fmal.c

math: fix rare underflow issue in fma

2013-10-07 18:46:43 +00:00

fmax.c

remove libm.h includes when math.h and float.h are enough

2012-03-13 19:51:14 +01:00

fmaxf.c

remove libm.h includes when math.h and float.h are enough

2012-03-13 19:51:14 +01:00

fmaxl.c

remove libm.h includes when math.h and float.h are enough

2012-03-13 19:51:14 +01:00

fmin.c

remove libm.h includes when math.h and float.h are enough

2012-03-13 19:51:14 +01:00

fminf.c

remove libm.h includes when math.h and float.h are enough

2012-03-13 19:51:14 +01:00

fminl.c

remove libm.h includes when math.h and float.h are enough

2012-03-13 19:51:14 +01:00

fmod.c

math: rewrite remainder functions (remainder, remquo, fmod, modf)

2013-09-05 11:30:07 +00:00

fmodf.c

math: rewrite remainder functions (remainder, remquo, fmod, modf)

2013-09-05 11:30:07 +00:00

fmodl.c

math: fix fmodl for IEEE binary128

2015-02-09 01:16:35 +01:00

frexp.c

first commit of the new libm!

2012-03-13 01:17:53 -04:00

frexpf.c

first commit of the new libm!

2012-03-13 01:17:53 -04:00

frexpl.c

math: long double fix (use ldshape union)

2013-09-05 11:30:08 +00:00

hypot.c

math: explicitly promote expressions to excess-precision types

2015-11-21 21:41:42 -05:00

hypotf.c

math: rewrite hypot

2013-09-05 11:30:07 +00:00

hypotl.c

math: rewrite hypot

2013-09-05 11:30:07 +00:00

ilogb.c

math: ilogb cleanup

2013-09-05 11:30:07 +00:00

ilogbf.c

math: ilogb cleanup

2013-09-05 11:30:07 +00:00

ilogbl.c

math: ilogb cleanup

2013-09-05 11:30:07 +00:00

j0.c

math: use double_t for temporaries to avoid stores on i386

2013-05-15 23:08:52 +00:00

j0f.c

fix threshold constants in j0f, y0f, j1f, y1f

2017-03-15 00:15:28 -04:00

j1.c

math: use double_t for temporaries to avoid stores on i386

2013-05-15 23:08:52 +00:00

j1f.c

fix threshold constants in j0f, y0f, j1f, y1f

2017-03-15 00:15:28 -04:00

jn.c

math: bessel cleanup (jn.c and jnf.c)

2013-01-01 22:20:45 +01:00

jnf.c

math: bessel cleanup (jn.c and jnf.c)

2013-01-01 22:20:45 +01:00

ldexp.c

remove libm.h includes when math.h and float.h are enough

2012-03-13 19:51:14 +01:00

ldexpf.c

remove libm.h includes when math.h and float.h are enough

2012-03-13 19:51:14 +01:00

ldexpl.c

remove libm.h includes when math.h and float.h are enough

2012-03-13 19:51:14 +01:00

lgamma_r.c

math: lgamma cleanup (simpler sin(pi*x) for the negative case)

2013-11-21 01:01:57 +00:00

lgamma.c

math: remove comment about aliasing lgamma as gamma

2012-03-27 22:12:20 +02:00

lgammaf_r.c

math: lgamma cleanup (simpler sin(pi*x) for the negative case)

2013-11-21 01:01:57 +00:00

lgammaf.c

math: remove comment about aliasing lgamma as gamma

2012-03-27 22:12:20 +02:00

lgammal.c

math: add dummy implementations of 128 bit long double functions

2015-03-11 18:54:53 -04:00

llrint.c

make lrint and llrint functions work without fenv support

2012-03-18 20:40:43 +01:00

llrintf.c

make lrint and llrint functions work without fenv support

2012-03-18 20:40:43 +01:00

llrintl.c

math: use '#pragma STDC FENV_ACCESS ON' when fenv is accessed

2012-11-13 13:34:45 +01:00

llround.c

simplify lround and llround functions

2012-03-18 20:52:33 +01:00

llroundf.c

simplify lround and llround functions

2012-03-18 20:52:33 +01:00

llroundl.c

simplify lround and llround functions

2012-03-18 20:52:33 +01:00

log1p.c

math: extensive log*.c cleanup

2013-10-28 01:16:14 +00:00

log1pf.c

math: extensive log*.c cleanup

2013-10-28 01:16:14 +00:00

log1pl.c

math: add dummy implementations of 128 bit long double functions

2015-03-11 18:54:53 -04:00

log2.c

math: extensive log*.c cleanup

2013-10-28 01:16:14 +00:00

log2f.c

math: extensive log*.c cleanup

2013-10-28 01:16:14 +00:00

log2l.c

math: add dummy implementations of 128 bit long double functions

2015-03-11 18:54:53 -04:00

log10.c

math: extensive log*.c cleanup

2013-10-28 01:16:14 +00:00

log10f.c

math: extensive log*.c cleanup

2013-10-28 01:16:14 +00:00

log10l.c

math: add dummy implementations of 128 bit long double functions

2015-03-11 18:54:53 -04:00

log.c

math: extensive log*.c cleanup

2013-10-28 01:16:14 +00:00

logb.c

math: fix logb(-0.0) in downward rounding mode

2013-09-05 11:30:07 +00:00

logbf.c

math: fix logb(-0.0) in downward rounding mode

2013-09-05 11:30:07 +00:00

logbl.c

math: fix logb(-0.0) in downward rounding mode

2013-09-05 11:30:07 +00:00

logf.c

math: extensive log*.c cleanup

2013-10-28 01:16:14 +00:00

logl.c

math: add dummy implementations of 128 bit long double functions

2015-03-11 18:54:53 -04:00

lrint.c

math: use '#pragma STDC FENV_ACCESS ON' when fenv is accessed

2012-11-13 13:34:45 +01:00

lrintf.c

make lrint and llrint functions work without fenv support

2012-03-18 20:40:43 +01:00

lrintl.c

math: use '#pragma STDC FENV_ACCESS ON' when fenv is accessed

2012-11-13 13:34:45 +01:00

lround.c

simplify lround and llround functions

2012-03-18 20:52:33 +01:00

lroundf.c

simplify lround and llround functions

2012-03-18 20:52:33 +01:00

lroundl.c

simplify lround and llround functions

2012-03-18 20:52:33 +01:00

modf.c

math: rewrite remainder functions (remainder, remquo, fmod, modf)

2013-09-05 11:30:07 +00:00

modff.c

math: rewrite remainder functions (remainder, remquo, fmod, modf)

2013-09-05 11:30:07 +00:00

modfl.c

math: use the rounding idiom consistently

2014-10-31 11:35:40 -04:00

nan.c

implement nan, nanf, nanl

2012-03-13 01:55:25 -04:00

nanf.c

implement nan, nanf, nanl

2012-03-13 01:55:25 -04:00

nanl.c

implement nan, nanf, nanl

2012-03-13 01:55:25 -04:00

nearbyint.c

math: use '#pragma STDC FENV_ACCESS ON' when fenv is accessed

2012-11-13 13:34:45 +01:00

nearbyintf.c

math: use '#pragma STDC FENV_ACCESS ON' when fenv is accessed

2012-11-13 13:34:45 +01:00

nearbyintl.c

math: use '#pragma STDC FENV_ACCESS ON' when fenv is accessed

2012-11-13 13:34:45 +01:00

nextafter.c

math: cosmetic cleanup (use explicit union instead of fshape and dshape)

2013-09-05 11:30:09 +00:00

nextafterf.c

math: cosmetic cleanup (use explicit union instead of fshape and dshape)

2013-09-05 11:30:09 +00:00

nextafterl.c

long double cleanup, initial commit

2013-09-05 11:30:07 +00:00

nexttoward.c

math: cosmetic cleanup (use explicit union instead of fshape and dshape)

2013-09-05 11:30:09 +00:00

nexttowardf.c

math: cosmetic cleanup (use explicit union instead of fshape and dshape)

2013-09-05 11:30:09 +00:00

nexttowardl.c

remove libm.h includes when math.h and float.h are enough

2012-03-13 19:51:14 +01:00

pow.c

math: fix pow signed shift ub

2016-10-20 01:32:27 -04:00

powf.c

math: fix pow(+-0,-inf) not to raise divbyzero flag

2015-04-18 00:18:52 -04:00

powl.c

math: fix pow(+-0,-inf) not to raise divbyzero flag

2015-04-18 00:18:52 -04:00

remainder.c

math: add (obsolete) bsd drem and finite functions

2013-11-21 01:16:49 +00:00

remainderf.c

math: add (obsolete) bsd drem and finite functions

2013-11-21 01:16:49 +00:00

remainderl.c

remove libm.h includes when math.h and float.h are enough

2012-03-13 19:51:14 +01:00

remquo.c

math: rewrite remainder functions (remainder, remquo, fmod, modf)

2013-09-05 11:30:07 +00:00

remquof.c

math: rewrite remainder functions (remainder, remquo, fmod, modf)

2013-09-05 11:30:07 +00:00

remquol.c

math: rewrite remainder functions (remainder, remquo, fmod, modf)

2013-09-05 11:30:07 +00:00

rint.c

fix rint.c and rintf.c when FLT_EVAL_METHOD!=0

2014-10-31 11:35:40 -04:00

rintf.c

fix rint.c and rintf.c when FLT_EVAL_METHOD!=0

2014-10-31 11:35:40 -04:00

rintl.c

math: use the rounding idiom consistently

2014-10-31 11:35:40 -04:00

round.c

math: use the rounding idiom consistently

2014-10-31 11:35:40 -04:00

roundf.c

math: use the rounding idiom consistently

2014-10-31 11:35:40 -04:00

roundl.c

math: use the rounding idiom consistently

2014-10-31 11:35:40 -04:00

scalb.c

math: define _GNU_SOURCE when implementing non-standard math functions

2013-12-12 03:42:11 +00:00

scalbf.c

math: define _GNU_SOURCE when implementing non-standard math functions

2013-12-12 03:42:11 +00:00

scalbln.c

remove libm.h includes when math.h and float.h are enough

2012-03-13 19:51:14 +01:00

scalblnf.c

remove libm.h includes when math.h and float.h are enough

2012-03-13 19:51:14 +01:00

scalblnl.c

remove libm.h includes when math.h and float.h are enough

2012-03-13 19:51:14 +01:00

scalbn.c

fix scalbn when result is in the subnormal range

2017-04-21 17:31:00 -04:00

scalbnf.c

fix scalbn when result is in the subnormal range

2017-04-21 17:31:00 -04:00

scalbnl.c

fix scalbn when result is in the subnormal range

2017-04-21 17:31:00 -04:00

signgam.c

math: remove libc.h include from libm.h

2013-09-05 14:03:10 +00:00

significand.c

add significand[fl] math functions

2012-08-13 14:51:43 -04:00

significandf.c

add significand[fl] math functions

2012-08-13 14:51:43 -04:00

sin.c

math: sin cos cleanup

2013-05-18 14:40:22 +00:00

sincos.c

math: define _GNU_SOURCE when implementing non-standard math functions

2013-12-12 03:42:11 +00:00

sincosf.c

math: define _GNU_SOURCE when implementing non-standard math functions

2013-12-12 03:42:11 +00:00

sincosl.c

math: fix aliasing violation in long double wrappers

2014-04-11 18:07:08 +02:00

sinf.c

math: sin cos cleanup

2013-05-18 14:40:22 +00:00

sinh.c

math: fix asin, atan, log1p, tanh to raise underflow on subnormal

2013-08-15 10:14:46 +00:00

sinhf.c

math: sinh.c cleanup similar to the cosh one

2012-12-16 19:49:55 +01:00

sinhl.c

math: add dummy implementations of 128 bit long double functions

2015-03-11 18:54:53 -04:00

sinl.c

math: long double trigonometric cleanup (cosl, sinl, sincosl, tanl)

2013-09-05 11:30:08 +00:00

sqrt.c

code cleanup of named constants

2012-03-19 23:41:19 +01:00

sqrtf.c

code cleanup of named constants

2012-03-19 23:41:19 +01:00

sqrtl.c

math: long double fix (use ldshape union)

2013-09-05 11:30:08 +00:00

tan.c

math: tan cleanups

2013-05-18 12:34:00 +00:00

tanf.c

math: tan cleanups

2013-05-18 12:34:00 +00:00

tanh.c

math: fix asin, atan, log1p, tanh to raise underflow on subnormal

2013-08-15 10:14:46 +00:00

tanhf.c

math: fix asin, atan, log1p, tanh to raise underflow on subnormal

2013-08-15 10:14:46 +00:00

tanhl.c

math: add dummy implementations of 128 bit long double functions

2015-03-11 18:54:53 -04:00

tanl.c

math: long double trigonometric cleanup (cosl, sinl, sincosl, tanl)

2013-09-05 11:30:08 +00:00

tgamma.c

math: lgamma cleanup (simpler sin(pi*x) for the negative case)

2013-11-21 01:01:57 +00:00

tgammaf.c

math: add a non-dummy tgamma implementation

2012-12-12 01:43:43 +01:00

tgammal.c

math: add dummy implementations of 128 bit long double functions

2015-03-11 18:54:53 -04:00

trunc.c

math: rewrite rounding functions (ceil, floor, trunc, round, rint)

2013-09-05 11:30:07 +00:00

truncf.c

math: rewrite rounding functions (ceil, floor, trunc, round, rint)

2013-09-05 11:30:07 +00:00

truncl.c

math: use the rounding idiom consistently

2014-10-31 11:35:40 -04:00