[Intel-gfx] [PATCH 3/4] drm/i915: Fix platform default aux ch for skl

[Ville Syrjälä]

On Fri, Feb 17, 2023 at 06:35:51PM +0530, Nautiyal, Ankit K wrote:
> 
> On 2/17/2023 5:32 PM, Ville Syrjälä wrote:
> > On Fri, Feb 17, 2023 at 03:15:59PM +0530, Nautiyal, Ankit K wrote:
> >> On 2/17/2023 4:43 AM, Ville Syrjala wrote:
> >>> From: Ville Syrjälä <ville.syrjala at linux.intel.com>
> >>>
> >>> SKL/derivatives have DDI E but no AUX E, so we need to pick
> >>> another aux ch as the platform default. DDI E is more or less
> >>> the other half of DDI A, so we pick AUX A.
> >>>
> >>> In all other cases we should have a corresponding aux ch for
> >>> each DDI.
> >>>
> >>> Signed-off-by: Ville Syrjälä <ville.syrjala at linux.intel.com>
> >>> ---
> >>>    drivers/gpu/drm/i915/display/intel_dp_aux.c | 14 ++++++++++++--
> >>>    1 file changed, 12 insertions(+), 2 deletions(-)
> >>>
> >>> diff --git a/drivers/gpu/drm/i915/display/intel_dp_aux.c b/drivers/gpu/drm/i915/display/intel_dp_aux.c
> >>> index 57eb3ff187fa..96967e21c94c 100644
> >>> --- a/drivers/gpu/drm/i915/display/intel_dp_aux.c
> >>> +++ b/drivers/gpu/drm/i915/display/intel_dp_aux.c
> >>> @@ -739,10 +739,20 @@ void intel_dp_aux_init(struct intel_dp *intel_dp)
> >>>    	cpu_latency_qos_add_request(&intel_dp->pm_qos, PM_QOS_DEFAULT_VALUE);
> >>>    }
> >>>    
> >>> +static enum aux_ch default_aux_ch(struct intel_encoder *encoder)
> >>> +{
> >>> +	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
> >>> +
> >>> +	/* SKL has DDI E but no AUX E */
> >>> +	if (DISPLAY_VER(i915) == 9 && encoder->port == PORT_E)
> >>> +		return AUX_CH_A;
> >>> +
> >> I guess we might need to fix intel_dp_aux_init also. Wont we have
> >> dig_port->aux_ch and intel_dp->aux_ch pointing to different things?
> > There is no intel_dp->aux_ch.
> >
> >> For example for SKL and DDI E, intel_dp->aux.name will be derived from
> >> dig_port->aux_ch i.e. AUX_CH_A but intel_dp->aux_ctl/data reg will still
> >> point to DP_AUX_CH_CTL/DATA_E.
> >>
> >> Am I missing something?
> > Everything is based on dig_port->aux_ch.
> 
> Yeah my bad. I think I misread and got confused. We indeed are using 
> dig_port->aux_ch, for getting aux_ctl/data reg.
> 
> I couldnt get the exact bspec mention for SKL not having AUX E, but SKL 
> Display connection Bspec:4217 throws some light, if I am not fully off.
> 
> Is this applicable to Canonlake too  (Bspec:14004) from Display 
> connection it does seem to have DDI E but no DP_AUX E?

There is no Cannonlake.

-- 
Ville Syrjälä
Intel