[Bug target/107590] __atomic_test_and_set broken on PowerPC

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107590

--- Comment #6 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Iain Sandoe from comment #5)
> I'm away from my usual sources of information but I'd suggest exploring the
> possibility that someone has assumed that either the spinlock or a bool is
> 8bits; As far as my memory serves both are 32b on power darwin.

Oh yes I forgot about bool being 32bit due to an ABI choice early on with
Apple's GCC 2.95 because of not having a bool type but defining it to int.
So there might be still a mismatch there still ...