[kaffe] mipsel JIT3
Kevin D. Kissell
kevink at mips.com
Sat Mar 6 16:51:02 PST 2004
> >> the correct fix is probobly to fix register.c to not assume that
> >> regno is an index into reginfo.
> Timothy> I'm hesitant to mess with the register allocator too much,
> Timothy> especially since it might break the other backends...
I have a very hard time believing that any backend could actually
*depend* on it being broken in this way. And this behavior is,
I believe, relatively recent, dating from 1.1.0 (the 1.0.7/1.0.8 JIT3
was more than 90% functional for MIPS), and I didn't think that
much major back-end work has been done since.
> I think it should be easy (FLW) to rework the register allocation
> without damaging the way register-sparse machines work by assuring
> that the behavior visible from those sides is the same. Besides, my
> current belief is that this is just the confusion between the
> SlotData.regno and kregs.regno.
Yeah, I noticed that name collision last night and wondered if it wasn't
causing trouble. Frankly, if I were a maintainer, I'd take the time
to change one or the other to be "regnum" or some such. It
could save many hours of maintenence time down the road...
More information about the kaffe