Richard Henderson 5c4f79ac62 target-hppa: Add softfloat specializations ... Like the original MIPS, HPPA has the MSB of an SNaN set. However, it has different rules for silencing an SNaN: (1) msb is cleared and (2) msb-1 must be set if the fraction is now zero, and (implementation defined) may be set always. I haven't checked real hardware but chose the set always alternative because it's easy and within spec. Backports commit 005fa38d86257d471ac461c066a5409a9f5ebb02 from qemu		2018-03-01 23:42:09 -05:00
..
softfloat-macros.h	fpu: Use plain 'int' rather than 'int_fast16_t' for exponents	2018-02-20 18:57:32 -05:00
softfloat-specialize.h	target-hppa: Add softfloat specializations	2018-03-01 23:42:09 -05:00
softfloat.c	fpu: add mechanism to check for invalid long double formats	2018-02-26 02:27:40 -05:00