On 08.05.2013, at 12:07, Torbjorn Granlund wrote: > Alexander Graf <ag...@suse.de> writes: > > Ok, so the real problem here is that NARROW_MODE is not set, but is > used to differentiate whether to use the 32bit cmp only or not. > > Eh? > > Richard, there are 2 ways out of this: > > 1) get rid of NARROW_MODE and always check ctx->sf > > No! > > The cmp insn with L set should NOT be affected by SF. That's the entire > point of my change.
You're right. I got confused there :). > > I reviewed the other uses of NARROW_MODE and didn't spot any errors. > (But I must confess that I would need to red the PPC manuals better inn > order to tell for sure.) > > 2) add a new 32bit only insns flag and create separate functions for 32bit > cmp calls > > Aurelien's patch looked promising, if one removes the exception casting. His exception casting is actually correct. You can use qemu-(system-)ppc64, but run it with a CPU definition that is 32bit only, like a G3. These old CPUs did not know the instruction with L yet, so they do throw an illegal instruction exception, which we have to model. Alex
