[Intel-xe] [PATCH v2 rebased 4/6] drm/xe/display: Enable modular fia in TGL

Souza, Jose jose.souza at intel.com
Mon Apr 10 17:29:16 UTC 2023


On Mon, 2023-04-10 at 13:20 -0400, Rodrigo Vivi wrote:
> On Mon, Apr 10, 2023 at 05:15:11PM +0000, Souza, Jose wrote:
> > On Mon, 2023-04-10 at 12:50 -0400, Rodrigo Vivi wrote:
> > > On Mon, Apr 10, 2023 at 04:24:09PM +0000, Souza, Jose wrote:
> > > > On Mon, 2023-04-10 at 11:38 -0400, Rodrigo Vivi wrote:
> > > > > On Thu, Apr 06, 2023 at 07:31:31AM -0700, José Roberto de Souza wrote:
> > > > > > TGL is the first platform tha has it.
> > > > > > Enabling it for DG1 don't cause any issues as DG1 don't have any real
> > > > > > TC ports, so intel_tc_port_init() will not be called for DG1.
> > > > > 
> > > > > so, why do we need to enable then?
> > > > 
> > > > TGL had it disabled.
> > > 
> > > so it looks like we need to refactor this info stuff...
> > > because looking to the desc it was there on TGL:
> > > 
> > > .display.has_modular_fia = 1,
> > 
> > Where?
> > I only see it set for adl-P in xe_display.c and on i915_pci.c for tgl, adl-p and mtl.
> 
> doh! I looked to the wrong place.
> What confused me is that this has_modular_fia doesn't exist anymore on i915.
> So we probably need something different here... or backport something that is
> already in drm-intel and then a fixup! on top... or something like that...

Huum, did not saw that this was gone in drm-tip.
It would need several backports to drop has_modular_fia in Xe, TC had a huge refactor.

Can we have this change marked as temporary patch until next Xe rebase?


> 
> > 
> > 
> > > 
> > > > This is needed to properly initialize TC in some skus.
> > > 
> > > This is a key info that was missing in the commit msg, please include it
> > > in a next revision. Although I'm now confused with the info set in the
> > > descrition and here...
> > 
> > Will do.
> > 
> > > 
> > > Why is that desc one not working?
> > > 
> > > > 
> > > > > 
> > > > > > 
> > > > > > Signed-off-by: José Roberto de Souza <jose.souza at intel.com>
> > > > > > ---
> > > > > >  drivers/gpu/drm/xe/xe_display.c | 5 ++++-
> > > > > >  1 file changed, 4 insertions(+), 1 deletion(-)
> > > > > > 
> > > > > > diff --git a/drivers/gpu/drm/xe/xe_display.c b/drivers/gpu/drm/xe/xe_display.c
> > > > > > index 5fab5b9fbde0c..0001a9d02495d 100644
> > > > > > --- a/drivers/gpu/drm/xe/xe_display.c
> > > > > > +++ b/drivers/gpu/drm/xe/xe_display.c
> > > > > > @@ -486,7 +486,10 @@ void xe_display_info_init(struct xe_device *xe)
> > > > > >  	switch (xe->info.platform) {
> > > > > >  	case XE_TIGERLAKE:
> > > > > >  	case XE_DG1:
> > > > > > -		xe->info.display = (struct xe_device_display_info) { GEN12_DISPLAY };
> > > > > > +		xe->info.display = (struct xe_device_display_info) {
> > > > > > +			GEN12_DISPLAY,
> > > > > > +			.has_modular_fia = 1,
> > > > > > +		};
> > > > > >  		break;
> > > > > >  	case XE_ALDERLAKE_S:
> > > > > >  		xe->info.display = (struct xe_device_display_info) {
> > > > > > -- 
> > > > > > 2.40.0
> > > > > > 
> > > > 
> > 



More information about the Intel-xe mailing list