https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80117
--- Comment #13 from Eric Botcazou <ebotcazou at gcc dot gnu.org> ---
> So what's a word then? Obviously, UNITS_PER_WORD does not fit.
I think that UNITS_PER_WORD is correct for Word_Size, the problem lies more in
the definition of Memory_Size and Address:
Memory_Size
An implementation-defined value [that is intended to reflect the memory
size of the configuration in storage elements.]
Address is of a definite, nonlimited type. Address represents machine
addresses capable of addressing individual storage elements.
so Memory_Size would need to be 2 ** 32 for ILP32 ABIs on 64-bit architectures
but I have no idea as to whether this would work with the current
implementation, hence the trick of using a pure 32-bit configuration instead
for them.
