On Tue, Sep 17, 2013 at 04:58:10PM +0100, Maciej W. Rozycki wrote: > Replace hardcoded CP0 PRId and CP1 FPIR register access masks throughout. > The change does not touch places that use shifted or partial masks. > > Signed-off-by: Maciej W. Rozycki <macro@xxxxxxxxxxxxxx> > --- > Ralf, > > Please apply. I think the places ignored by this change should be > further reviewed, especially the shifted masks that can likely remove the > shifts and rely on compiler optimisation instead. I decided to make this > change as straightforward as possible to avoid accidental breakage in code > I have no way to test. Also partial masks are probably better handled > with macros rather than hardcoded constants scattered throughout. I can > see steps have been taken towards this already (PRID_REV_ENCODE_*). Looks ok, queud for 3.13. But while it's cleaner, I think the idiom read_c0_prid() & some_MASK is so common that maybe something like #define read_c0_prid_imp() (read_c0_prid() & PRID_IMP_MASK) #define read_c0_prid_rev() (read_c0_prid() & PRID_REV_MASK) #define read_c0_prid_comp() (read_c0_prid() & PRID_COMP_MASK) should be introduced as a next step. Ralf
