m68020 is used because that is what gcc uses by default also in other m68k targets and that CPU ~ "^m68020" check is based on ISA, not CPU name and I wanted to change as little as possible against upstream.
I have also added mapping mcpu.5475=mcpu.5475/m68881 so -mcpu=5475 -mhard-float wont default into /usr/lib/lib*.a.
I'm not going to publish this anytime soon, I plan to backport also other m68k changes from gcc >= 8, too. Plus, if approved, we need to update mintlib & friends (see also freemint/fdlibm#4).
Making m68030 soft-float and m68060 hard-float by default required some creative thinking. This MR enables following multilib paths:
(+
mshort
variants)m68020
is used because that is what gcc uses by default also in other m68k targets and thatCPU ~ "^m68020"
check is based on ISA, not CPU name and I wanted to change as little as possible against upstream.I have also added mapping
mcpu.5475=mcpu.5475/m68881
so-mcpu=5475 -mhard-float
wont default into/usr/lib/lib*.a
.I'm not going to publish this anytime soon, I plan to backport also other m68k changes from gcc >= 8, too. Plus, if approved, we need to update mintlib & friends (see also freemint/fdlibm#4).
Fixes #1.