On Thu, Nov 05, 2015 at 12:01:51PM +0000, Chris Wilson wrote: > On Thu, Nov 05, 2015 at 01:44:08PM +0200, Ville Syrjälä wrote: > > On Thu, Nov 05, 2015 at 11:03:38AM +0000, Chris Wilson wrote: > > > On Wed, Nov 04, 2015 at 11:20:07PM +0200, ville.syrjala@xxxxxxxxxxxxxxx wrote: > > > > From: Ville Syrjälä <ville.syrjala@xxxxxxxxxxxxxxx> > > > > > > > > When register type safety happens, we can't just try to emit the > > > > register itself to the ring. Instead we'll need to extract the > > > > offset from it first. Add some convenience functions that will do > > > > that. > > > > > > > > v2: Convert MOCS setup too > > > > > > > > Signed-off-by: Ville Syrjälä <ville.syrjala@xxxxxxxxxxxxxxx> > > > > > > The only insane thing about this patch is the stupid ring emission API. > > > > One extra idea just popped to my mind. Should I maybe make the > > emit_reg() take the value too and emit both the reg offset and value? > > They always come in pairs after all. > > Uncertain. I think no. You need to do LRI and SRM separately (so that we > don't get confused between the register value to load and the memory > address to write to). But the more important factor for me, is that I > don't want to hide the individual calls to emit() - as they need to be > easily reviewed and checked against the count given to ring_begin(). Indeed, that's a good reason for keeping it separate. -- Ville Syrjälä Intel OTC _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/intel-gfx
