Keith Packard via Newlib 7e690708b4 libm: ARM without HW double does not have fast FMA ...

32-bit ARM processors with HW float (but not HW double) may define
__ARM_FEATURE_FMA, but that only means they have fast FMA for 32-bit
floats.

Signed-off-by: Keith Packard <keithp@keithp.com>

2020-10-14 10:54:27 -04:00

..

common

libm: ARM without HW double does not have fast FMA

2020-10-14 10:54:27 -04:00

complex

Remove matherr, and SVID and X/Open math library configurations

2019-01-23 10:46:24 +01:00

fenv

Add patch from Joel Sherrill for i386 and x86_64 fenv support

2020-10-14 10:54:03 -04:00

machine

arm: Fix include to avoid undefined reference

2020-10-14 10:54:25 -04:00

math

libm/math: ensure that expf(-huge) sets FE_UNDERFLOW exception

2020-10-14 10:54:27 -04:00

mathfp

Remove matherr, and SVID and X/Open math library configurations

2019-01-23 10:46:24 +01:00

test

Remove matherr, and SVID and X/Open math library configurations

2019-01-23 10:46:24 +01:00

aclocal.m4

Fix libm regenerated files again.

2016-02-17 17:06:46 -05:00

config.h.in

import newlib-2000-02-17 snapshot

2000-02-17 19:39:52 +00:00

configure

Bump up newlib release to 3.3.0

2020-10-14 10:54:08 -04:00

configure.in

Add default implementation of fenv.h and all methods

2020-10-14 10:53:58 -04:00

libm.in.xml

fenv: Include documentation in generated .info file

2020-10-14 10:53:58 -04:00

libm.texinfo

fenv: Include documentation in generated .info file

2020-10-14 10:53:58 -04:00

Makefile.am

Add default implementation of fenv.h and all methods

2020-10-14 10:53:58 -04:00

Makefile.in

Fix some generated files

2020-10-14 10:54:03 -04:00