Open glaubitz opened 6 years ago
Try:
2b745c4c03 softfloat: disable floatx80_invalid_encoding() for m68k
It's not the good behavior: we should raise an exception instead of ignoring it, but it works better than behaving like intel FPU
Yes, this patch fixes this particular problem.
I had to add an additional declaration at the top of translate.c
though in order to make it build:
+static inline bool floatx80_invalid_encoding(floatx80 a);
+
Trying to build an R package on Debian when using
qemu
from git upstream fails with:Building the same package using
qemu-m68k
works just fine. So I guess it's one of the FPU patches that's missing.Can you suggest which?