[Mesa-dev] [PATCH 34/64] genxml: Add macros and #includes for gens 4-6

[Chad Versace]

On Mon 20 Jun 2016, Jason Ekstrand wrote:
> On Mon, Jun 20, 2016 at 12:08 PM, Chad Versace <chad.versace at intel.com>
> wrote:
> 
> > On Sat 11 Jun 2016, Jason Ekstrand wrote:
> > > ---
> > >  src/intel/genxml/genX_pack.h  | 10 +++++++++-
> > >  src/intel/genxml/gen_macros.h | 15 ++++++++++++++-
> > >  2 files changed, 23 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/src/intel/genxml/genX_pack.h b/src/intel/genxml/genX_pack.h
> > > index 7967c29..0c25c4e 100644
> > > --- a/src/intel/genxml/genX_pack.h
> > > +++ b/src/intel/genxml/genX_pack.h
> > > @@ -27,7 +27,15 @@
> > >  #  error "The GEN_VERSIONx10 macro must be defined"
> > >  #endif
> > >
> > > -#if (GEN_VERSIONx10 == 70)
> > > +#if (GEN_VERSIONx10 == 40)
> > > +#  include "genxml/gen4_pack.h"
> > > +#elif (GEN_VERSIONx10 == 45)
> > > +#  include "genxml/gen45_pack.h"
> > > +#elif (GEN_VERSIONx10 == 50)
> > > +#  include "genxml/gen5_pack.h"
> > > +#elif (GEN_VERSIONx10 == 60)
> > > +#  include "genxml/gen6_pack.h"
> > > +#elif (GEN_VERSIONx10 == 70)
> > >  #  include "genxml/gen7_pack.h"
> > >  #elif (GEN_VERSIONx10 == 75)
> > >  #  include "genxml/gen75_pack.h"
> > > diff --git a/src/intel/genxml/gen_macros.h
> > b/src/intel/genxml/gen_macros.h
> > > index 868bc22..1d591fa 100644
> > > --- a/src/intel/genxml/gen_macros.h
> > > +++ b/src/intel/genxml/gen_macros.h
> > > @@ -57,9 +57,22 @@
> > >
> > >  #define GEN_GEN ((GEN_VERSIONx10) / 10)
> > >  #define GEN_IS_HASWELL ((GEN_VERSIONx10) == 75)
> > > +#define GEN_IS_G4X ((GEN_VERSIONx10) == 45)
> >
> > I expected the new macro name to be GEN_IS_G45. Why is it G4X?
> >
> 
> i started with IS_G45 and then realized that the brw_device_info field is
> is_g4x so I decided to be consistent.  Chatting around the cube a bit, it
> sounds like g44 and g43 were also things.  Ken would be happy to rename it
> all to is_g45.  I don't think anyone cares that much.

I prefer is_g45, because that matches the prm name. All the PRM pages
for that chipset say G45 in the page header. And I suspect people, when
discussing it, say "g45" and not "g4x".

However, it's not important at all. Either name works for me.