Re: [Intel-gfx] [RFC 1/5] drm/i915/dp: Program source OUI on eDP panels

[Lyude Paul]

On Wed, 2020-09-16 at 10:43 +0300, Jani Nikula wrote:
> On Tue, 15 Sep 2020, Rodrigo Vivi <rodrigo.vivi@xxxxxxxxx> wrote:
> > On Tue, Sep 15, 2020 at 01:29:35PM -0400, Lyude Paul wrote:
> > > Since we're about to start adding support for Intel's magic HDR
> > > backlight interface over DPCD, we need to ensure we're properly
> > > programming this field so that Intel specific sink services are exposed.
> > > Otherwise, 0x300-0x3ff will just read zeroes.
> > > 
> > > We also take care not to reprogram the source OUI if it already matches
> > > what we expect. This is just to be careful so that we don't accidentally
> > > take the panel out of any backlight control modes we found it in.
> 
> (For whatever reason I didn't receive the original message.)
> 
> > > Signed-off-by: Lyude Paul <lyude@xxxxxxxxxx>
> > > Cc: thaytan@xxxxxxxxxxxx
> > > Cc: Vasily Khoruzhick <anarsoul@xxxxxxxxx>
> > > ---
> > >  drivers/gpu/drm/i915/display/intel_dp.c | 32 +++++++++++++++++++++++++
> > >  1 file changed, 32 insertions(+)
> > > 
> > > diff --git a/drivers/gpu/drm/i915/display/intel_dp.c
> > > b/drivers/gpu/drm/i915/display/intel_dp.c
> > > index 4bd10456ad188..b591672ec4eab 100644
> > > --- a/drivers/gpu/drm/i915/display/intel_dp.c
> > > +++ b/drivers/gpu/drm/i915/display/intel_dp.c
> > > @@ -3428,6 +3428,7 @@ void intel_dp_sink_set_decompression_state(struct
> > > intel_dp *intel_dp,
> > >  void intel_dp_sink_dpms(struct intel_dp *intel_dp, int mode)
> > >  {
> > >  	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
> > > +	u8 edp_oui[] = { 0x00, 0xaa, 0x01 };
> > 
> > what are these values?
> 
> An OUI lookup confirms these are Intel OUI.

Thanks for the confirmation!

> 
> > >  	int ret, i;
> > >  
> > >  	/* Should have a valid DPCD by this point */
> > > @@ -3443,6 +3444,14 @@ void intel_dp_sink_dpms(struct intel_dp *intel_dp,
> > > int mode)
> > >  	} else {
> > >  		struct intel_lspcon *lspcon = dp_to_lspcon(intel_dp);
> > >  
> > > +		/* Write the source OUI as early as possible */
> > > +		if (intel_dp_is_edp(intel_dp)) {
> > > +			ret = drm_dp_dpcd_write(&intel_dp->aux, DP_SOURCE_OUI,
> > > edp_oui,
> > > +						sizeof(edp_oui));
> > > +			if (ret < 0)
> > > +				drm_err(&i915->drm, "Failed to write eDP source
> > > OUI\n");
> > > +		}
> > > +
> > >  		/*
> > >  		 * When turning on, we need to retry for 1ms to give the sink
> > >  		 * time to wake up.
> > > @@ -4530,6 +4539,23 @@ static void intel_dp_get_dsc_sink_cap(struct
> > > intel_dp *intel_dp)
> > >  	}
> > >  }
> > >  
> > > +static void
> > > +intel_edp_init_source_oui(struct intel_dp *intel_dp)
> > > +{
> > > +	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
> > > +	u8 oui[] = { 0x00, 0xaa, 0x01 };
> > > +	u8 buf[3] = { 0 };
> > > +
> > > +	if (drm_dp_dpcd_read(&intel_dp->aux, DP_SOURCE_OUI, buf, sizeof(buf)) <
> > > 0)
> > > +		drm_err(&i915->drm, "Failed to read source OUI\n");
> > > +
> > > +	if (memcmp(oui, buf, sizeof(oui)) == 0)
> > > +		return;
> > > +
> > > +	if (drm_dp_dpcd_write(&intel_dp->aux, DP_SOURCE_OUI, oui, sizeof(oui)) <
> > > 0)
> > > +		drm_err(&i915->drm, "Failed to write source OUI\n");
> > > +}
> 
> Maybe add this function with a parameter to force write or write only if
> necessary, and call from both places that set source OUI?
> 
> > > +
> > >  static bool
> > >  intel_edp_init_dpcd(struct intel_dp *intel_dp)
> > >  {
> > > @@ -4607,6 +4633,12 @@ intel_edp_init_dpcd(struct intel_dp *intel_dp)
> > >  	if (INTEL_GEN(dev_priv) >= 10 || IS_GEMINILAKE(dev_priv))
> > >  		intel_dp_get_dsc_sink_cap(intel_dp);
> > >  
> > > +	/*
> > > +	 * Program our source OUI so we can make various Intel-specific AUX
> > > +	 * services available (such as HDR backlight controls)
> > > +	 */
> > > +	intel_edp_init_source_oui(intel_dp);
> > 
> > I believe we should restrict this to the supported platforms: cfl, whl, cml,
> > icl, tgl
> > no?
> 
> Mmh, this just exposes sink behaviour that I think can be supported by
> any platform. I don't understand the notion of "supported platforms"
> here.

Probably because the spec sheets that we have on this seem to suggest that this
is new for particular platforms, and Intel seems to also additionally move a bit
away from some of the interfaces exposed here onto actual VESA standards
starting with icl and tgl.

I would be fine with adding this, but I'm not really sure it's needed here
either unless we want to stop using Intel backlight control interfaces for later
hardware generations at some point in the future.

> 
> > > +
> > >  	return true;
> > >  }
> > >  
> > > -- 
> > > 2.26.2
> > > 
> > > _______________________________________________
> > > dri-devel mailing list
> > > dri-devel@xxxxxxxxxxxxxxxxxxxxx
> > > https://lists.freedesktop.org/mailman/listinfo/dri-devel
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
> > https://lists.freedesktop.org/mailman/listinfo/intel-gfx
-- 
Sincerely,
      Lyude Paul (she/her)
      Software Engineer at Red Hat