> (I'm not sure what you mean by "problems with PowerPC FPU semantics
> vs expected fp{get,set}sticky semantics", can you explain?)

Hm... I meant "the ppc requires that you clear the exception or it
will keep generating the trap, but fpgetsticky seems to expect it to
stick around after handling the exception", but it looks like I wasn't
paying attention and actually fpgetsticky only needs to work if the
exception was masked (at least that's all that's needed to pass
regress/lib/libc/ieeefp/except). So I think my problem is solved...

> Anyway, I can provide some code for this, but am currently

I've actually got most of it written.

-- Dan