> -----Original Message-----
> From: Intel-gfx <intel-gfx-bounces@xxxxxxxxxxxxxxxxxxxxx> On Behalf Of Ville Syrjälä
> Sent: Friday, June 9, 2023 1:06 AM
> To: Jani Nikula <jani.nikula@xxxxxxxxxxxxxxx>
> Cc: intel-gfx@xxxxxxxxxxxxxxxxxxxxx
> Subject: Re: [PATCH v2 2/7] drm/i915: Introduce device info port_mask
>
> On Fri, Jun 02, 2023 at 05:11:45PM +0300, Jani Nikula wrote:
> > On Wed, 31 May 2023, Ville Syrjala <ville.syrjala@xxxxxxxxxxxxxxx> wrote:
> > > @@ -556,6 +590,16 @@ static const struct intel_display_device_info
> > > gen11_display = {
> > >
> > > static const struct intel_display_device_info tgl_display = {
> > > XE_D_DISPLAY,
> > > +
> > > + .__runtime_defaults.port_mask = BIT(PORT_A) | BIT(PORT_B) |
> > > +BIT(PORT_C) |
> >
> > Where does port C come from?
>
> From the spec.
>
> > It's not in intel_setup_outputs()
> > currently.
>
> Hmm. Based on the history there seems to be some kind of screwup in the combo
> PHY code wrt. ports that are not actually present in the SKU. The spec says we
> should rely on hpd to figure out what ports are actually present. So looks like the
> combo PHY code needs quite a bit of redesign :( I guess I'll leave this out until then.
>
This info on what ports are physically present is advertised via VBT currently.
But based on spec, we can declare the capability at setup and intel_ddi_init should
check VBT and proceed.
> > > + BIT(PORT_TC1) | BIT(PORT_TC2) | BIT(PORT_TC3) | BIT(PORT_TC4)
> |
> > > +BIT(PORT_TC5) | BIT(PORT_TC5),
> >
> > TC5 duplicated, TC6 missing.
Yeah some typo here.
With the above addressed, this change looks good to me.
Reviewed-by: Uma Shankar <uma.shankar@xxxxxxxxx>
> >
> > BR,
> > Jani.
> >
> > --
> > Jani Nikula, Intel Open Source Graphics Center
>
> --
> Ville Syrjälä
> Intel
