On Fri, Sep 07, 2018 at 04:30:36PM +0100, Chris Wilson wrote:
> Quoting Ville Syrjala (2018-09-07 16:24:04)
> > diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
> > index 8162025114f5..9e16bdcffc84 100644
> > --- a/drivers/gpu/drm/i915/intel_drv.h
> > +++ b/drivers/gpu/drm/i915/intel_drv.h
> > @@ -504,11 +504,7 @@ struct intel_plane_state {
> > struct {
> > u32 offset;
> > int x, y;
> > - } main;
> > - struct {
> > - u32 offset;
> > - int x, y;
> > - } aux;
> > + } color_plane[2];
>
> Is (main, aux) significant to keep around as an enum?
The main/aux names are specific to our hardware design whereas we
often just index this stuff using the color plane index as used
by drm where the main/aux terminology doesn't apply. So we would
have some code indexing this with just a raw int and some using
the enum. Not sure how confusing that would end up being.
Also with icl nv12 support we'll actually just end up using two
hardware planes where neither one will get its aux surface registers
programmed. So if we still stick to a single shared plane state for
both hardware planes the main/aux terminology no longer really applies.
But the details of this are still up in the air so not sure how it'll
end up looking.
So I guess I'd probably just go with no enum and make it clear that
this stuff refers to the drm color plane index instead of the hardware
main and aux surfaces, even though currently they do end up being the
same thing.
--
Ville Syrjälä
Intel
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
