Re: [Intel-xe] [PATCH 3/4] drm/xe/display: Consider has_display to enable it

["Souza, Jose" <jose.souza@intel.com>]

On Thu, 2023-05-11 at 10:12 -0700, Lucas De Marchi wrote:
> On Thu, May 11, 2023 at 12:47:26PM +0000, Jose Souza wrote:
> > On Wed, 2023-05-10 at 12:54 -0700, Lucas De Marchi wrote:
> > > Stop the dance of enabling the display-related driver_features to later
> > > disable them on display info init if the platform doesn't have display
> > > IP. Besides being needless work, it wasn't covering the ATS-M case that
> > > is the same platform as DG2, but without display.
> > 
> > Xe should set pipe_mask = 0 and call i915 functions that will handle no display cases.
> 
> in xe, enable_display is the runtime config to be the equivalent of
> DRM_XE_DISPLAY=n. It is *not* to meant disabling the display.
> 
> history why this ever came to be was:
> 
> 1) display integration back in the day was less than ideal (still is),
>     and developers couldn't test things ignoring display

Xe CI is now testing display, in my opinion this option to disable display should be removed and display support always built.

> 2) have a way to tell the driver "don't ever touch display IP" for
>     bring-up situations.

pipe_mask = 0/HAS_DISPLAY() should take care of it.
developers could force pipe_mask = 0 for giving platform bring-up.

> 
> For (1) we may turn that into "disable display" now, but not for (2).
> 
> I'll take a look on how much work it would be to migrate to a
> disable-display scenario rather than the simple "Fix ats-m" being done
> here.
> 
> Lucas De Marchi