"Paulo J. Matos" <pa...@matos-sorge.com> writes: > On 26/09/11 17:23, Ian Lance Taylor wrote: >> >> The function added_clobbers_hard_reg_p is a generated function. So >> another approach would be some sort of attribute which directs the >> generator (genemit) to ignore certain hard registers. >> > > This definitely sounds like the best approach for my specific case. I > understand that for such a change to be integrated in GCC there might be > other implications due to the vast number of different backends > supported by GCC.
I don't understand why FLAGS_REGNUM is special. I'm not sure which port you're working on, but presumably the point of modelling FLAGS_REGNUM is that it can be live across instructions. The problem described by the comment you quote then applies: we can't know (because the caller isn't guaranteed to have checked) that FLAGS_REGNUM isn't already live. Richard