On May 9, 2013, at 2:06 AM, Richard Sandiford <rsand...@linux.vnet.ibm.com> wrote: >> Hmmm... psi/si in memory are tieable, psi/si in registers >> aren't... what's the right thing to do here? > > I think we assume everything is tieable in memory. MODES_TIEABLE_P says > whether we should assume that a pseudo register in one mode can be > accessed in another mode without triggering some kind of copy operation.
I hate TIEABLE. :-) I'd love to see it engineered out.
