[Intel-gfx] [v7 01/10] drm/i915/display: Add HDR Capability detection for LSPCON

Shankar, Uma uma.shankar at intel.com
Thu Oct 8 11:29:04 UTC 2020



> -----Original Message-----
> From: Ville Syrjälä <ville.syrjala at linux.intel.com>
> Sent: Thursday, October 8, 2020 4:25 PM
> To: Shankar, Uma <uma.shankar at intel.com>
> Cc: intel-gfx at lists.freedesktop.org
> Subject: Re: [v7 01/10] drm/i915/display: Add HDR Capability detection for
> LSPCON
> 
> On Tue, Oct 06, 2020 at 06:36:45PM +0530, Uma Shankar wrote:
> > LSPCON firmware exposes HDR capability through LPCON_CAPABILITIES DPCD
> > register. LSPCON implementations capable of supporting HDR set
> > HDR_CAPABILITY bit in LSPCON_CAPABILITIES to 1. This patch reads the
> > same, detects the HDR capability and adds this to intel_lspcon struct.
> >
> > v2: Addressed Jani Nikula's review comment and fixed the HDR
> >     capability detection logic
> >
> > Signed-off-by: Uma Shankar <uma.shankar at intel.com>
> > ---
> >  .../drm/i915/display/intel_display_types.h    |  1 +
> >  drivers/gpu/drm/i915/display/intel_lspcon.c   | 30 +++++++++++++++++++
> >  2 files changed, 31 insertions(+)
> >
> > diff --git a/drivers/gpu/drm/i915/display/intel_display_types.h
> > b/drivers/gpu/drm/i915/display/intel_display_types.h
> > index d5dc18cb8c39..fb8cfc0981d6 100644
> > --- a/drivers/gpu/drm/i915/display/intel_display_types.h
> > +++ b/drivers/gpu/drm/i915/display/intel_display_types.h
> > @@ -1398,6 +1398,7 @@ struct intel_lspcon {
> >  	bool active;
> >  	enum drm_lspcon_mode mode;
> >  	enum lspcon_vendor vendor;
> > +	bool hdr_supported;
> >  };
> >
> >  struct intel_digital_port {
> > diff --git a/drivers/gpu/drm/i915/display/intel_lspcon.c
> > b/drivers/gpu/drm/i915/display/intel_lspcon.c
> > index ee95fc353a56..f92962195698 100644
> > --- a/drivers/gpu/drm/i915/display/intel_lspcon.c
> > +++ b/drivers/gpu/drm/i915/display/intel_lspcon.c
> > @@ -35,6 +35,8 @@
> >  #define LSPCON_VENDOR_PARADE_OUI 0x001CF8  #define
> > LSPCON_VENDOR_MCA_OUI 0x0060AD
> >
> > +#define DPCD_MCA_LSPCON_HDR_STATUS	0x70003
> > +
> >  /* AUX addresses to write MCA AVI IF */  #define
> > LSPCON_MCA_AVI_IF_WRITE_OFFSET 0x5C0  #define
> LSPCON_MCA_AVI_IF_CTRL
> > 0x5DF @@ -104,6 +106,32 @@ static bool lspcon_detect_vendor(struct
> > intel_lspcon *lspcon)
> >  	return true;
> >  }
> >
> > +static void lspcon_detect_hdr_capability(struct intel_lspcon *lspcon)
> > +{
> > +	struct intel_digital_port *intel_dig_port =
> > +		container_of(lspcon, struct intel_digital_port, lspcon);
> 
> s/intel_dig_port/dig_port/ to conform with commit 7801f3b792b0
> ("drm/i915/display: prefer dig_port to reference intel_digital_port")
> 
> > +	struct drm_device *dev = intel_dig_port->base.base.dev;
> > +	struct intel_dp *dp = lspcon_to_intel_dp(lspcon);
> > +	u8 hdr_caps;
> > +	int ret;
> > +
> > +	/* Enable HDR for MCA based LSPCON devices */
> > +	if (lspcon->vendor == LSPCON_VENDOR_MCA)
> > +		ret = drm_dp_dpcd_read(&dp->aux,
> DPCD_MCA_LSPCON_HDR_STATUS,
> > +				       &hdr_caps, 1);
> > +	else
> > +		return;
> > +
> > +	if (ret < 0) {
> > +		drm_dbg_kms(dev, "hdr capability detection failed\n");
> > +		lspcon->hdr_supported = false;
> > +		return;
> > +	} else if (hdr_caps & 0x1) {
> > +		drm_dbg_kms(dev, "lspcon capable of HDR\n");
> > +		lspcon->hdr_supported = true;
> > +	}
> > +}
> > +
> >  static enum drm_lspcon_mode lspcon_get_current_mode(struct
> > intel_lspcon *lspcon)  {
> >  	enum drm_lspcon_mode current_mode;
> > @@ -554,6 +582,8 @@ static bool lspcon_init(struct intel_digital_port
> *dig_port)
> >  		return false;
> >  	}
> >
> > +	lspcon_detect_hdr_capability(lspcon);
> > +
> 
> This is now too late since we do this after registering the connector.
> Need to move this to the init stage, but lspcon detection requires hpd detection
> logic to be enabled, so once I get the hpd init order sorted we need to do this
> after intel_hpd_init() but before the connector is registered. Hmm, maybe we
> can actually do it from connector's
> .late_register() hook?

I am actually calling it from ddi_init itself in patch 3 assuming its just the one hardware which is
having issues (not sure if this a limitation for LSPCON generically). So if detection is successful
register HDR else let it not get enabled. It doesn't sound good though but was not getting any better
ideas ☹.

I think late_register may work out for us.

Regards,
Uma Shankar

> 
> >  	connector->ycbcr_420_allowed = true;
> >  	lspcon->active = true;
> >  	DRM_DEBUG_KMS("Success: LSPCON init\n");
> > --
> > 2.26.2
> 
> --
> Ville Syrjälä
> Intel


More information about the Intel-gfx mailing list