[Intel-gfx] [PATCH] drm/i915/icl: apply Display WA #1178 to fix type C dongles

[Ville Syrjälä]

On Wed, Oct 03, 2018 at 04:43:53PM -0700, Lucas De Marchi wrote:
> On Wed, Oct 03, 2018 at 04:30:36PM -0700, Rodrigo Vivi wrote:
> > On Mon, Sep 24, 2018 at 09:54:35AM -0700, Lucas De Marchi wrote:
> > > Display WA #1178 is meant to fix Aux channel voltage swing too low with
> > > some type C dongles. Although it is for type C, of ICL it only applies
> > > to combo phy and not to eDP. This means we need to apply the WA only on
> > > Aux B.
> > > 
> > > Cc: Rodrigo Vivi <rodrigo.vivi at intel.com>
> > > Cc: Paulo Zanoni <paulo.r.zanoni at intel.com>
> > > Cc: Ville Syrjälä <ville.syrjala at linux.intel.com>
> > > Cc: Imre Deak <imre.deak at intel.com>
> > > Signed-off-by: Lucas De Marchi <lucas.demarchi at intel.com>
> > > ---
> > >  drivers/gpu/drm/i915/i915_reg.h         | 4 ++++
> > >  drivers/gpu/drm/i915/intel_runtime_pm.c | 7 +++++++
> > >  2 files changed, 11 insertions(+)
> > > 
> > > diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
> > > index e7e6ca7f9665..1e92112d23de 100644
> > > --- a/drivers/gpu/drm/i915/i915_reg.h
> > > +++ b/drivers/gpu/drm/i915/i915_reg.h
> > > @@ -8928,6 +8928,10 @@ enum skl_power_gate {
> > >  #define   CNL_AUX_ANAOVRD1_ENABLE	(1 << 16)
> > >  #define   CNL_AUX_ANAOVRD1_LDO_BYPASS	(1 << 23)
> > >  
> > > +#define ICL_AUX_ANAOVRD1_B		_MMIO(0x6C398)
> > > +#define   ICL_AUX_ANAOVRD1_LDO_BYPASS	(1 << 7)
> > > +#define   ICL_AUX_ANAOVRD1_ENABLE	(1 << 0)
> > > +
> > >  /* HDCP Key Registers */
> > >  #define HDCP_KEY_CONF			_MMIO(0x66c00)
> > >  #define  HDCP_AKSV_SEND_TRIGGER		BIT(31)
> > > diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c
> > > index 0fdabce647ab..a97d2f762b77 100644
> > > --- a/drivers/gpu/drm/i915/intel_runtime_pm.c
> > > +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c
> > > @@ -436,6 +436,13 @@ icl_combo_phy_aux_power_well_enable(struct drm_i915_private *dev_priv,
> > >  	I915_WRITE(ICL_PORT_CL_DW12(port), val | ICL_LANE_ENABLE_AUX);
> > >  
> > >  	hsw_wait_for_power_well_enable(dev_priv, power_well);
> > > +
> > > +	/* Display WA #1178: icl */
> > > +	if (IS_ICELAKE(dev_priv) && pw_idx == ICL_PW_CTL_IDX_AUX_B) {
> > 
> > Spec tells:
> > CNL: This programming does not apply to Aux A.
> > ILC: Aux A: Set 0x162398 bit 0 and bit 7 = 1
> > with extra note of This programming only applies for external ports on the combo PHY,
> > not on type C PHY, and not for eDP.
> > 
> > So I believe we need to add Aux A here and add a check for !is_edp
> 
> ha! But spec 20584 for Icelake says Combo Port A can only be used for eDP or DSI.
> Or am I missing anything here?

Hmm. I wonder if that's correct. Earlier platforms did allow normal DP
(apart from audio/mst/dp++/etc.) on port A (despute the transcoder being
called EDP). I'm not sure what the critical difference between DP and eDP
would be really. Maybe file an issue in the spec to clarify?

> 
> Lucas De Marchi
> 
> > 
> > > +		val = I915_READ(ICL_AUX_ANAOVRD1_B);
> > > +		val |= ICL_AUX_ANAOVRD1_ENABLE | ICL_AUX_ANAOVRD1_LDO_BYPASS;
> > > +		I915_WRITE(ICL_AUX_ANAOVRD1_B, val);
> > > +	}
> > >  }
> > >  
> > >  static void
> > > -- 
> > > 2.17.1
> > > 
> > > _______________________________________________
> > > Intel-gfx mailing list
> > > Intel-gfx at lists.freedesktop.org
> > > https://lists.freedesktop.org/mailman/listinfo/intel-gfx
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx at lists.freedesktop.org
> > https://lists.freedesktop.org/mailman/listinfo/intel-gfx
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx

-- 
Ville Syrjälä
Intel