[Intel-gfx] [PATCH v7] drm/i915: implement EXTENDED_RECEIVER_CAPABILITY_FIELD_PRESENT

[Manasi Navare]

On Wed, Nov 28, 2018 at 11:09:46AM +0200, Jani Nikula wrote:
> On Tue, 27 Nov 2018, Manasi Navare <manasi.d.navare at intel.com> wrote:
> > From: Matt Atwood <matthew.s.atwood at intel.com>
> >
> > According to DP spec (2.9.3.1 of DP 1.4) if
> > EXTENDED_RECEIVER_CAPABILITY_FIELD_PRESENT is set the addresses in DPCD
> > 02200h through 0220Fh shall contain the DPRX's true capability. These
> > values will match 00000h through 0000Fh, except for DPCD_REV,
> > MAX_LINK_RATE, DOWN_STREAM_PORT_PRESENT.
> >
> > Read from DPCD once for all 3 values as this is an expensive operation.
> > Spec mentions that all of address space 02200h through 0220Fh should
> > contain the right information however currently only 3 values can
> > differ.
> >
> > There is no address space in the intel_dp->dpcd struct for addresses
> > 02200h through 0220Fh, and since so much of the data is a identical,
> > simply overwrite the values stored in 00000h through 0000Fh with the
> > values that can be overwritten from addresses 02200h through 0220Fh.
> >
> > This patch helps with backward compatibility for devices pre DP1.3.
> >
> > v2: read only dpcd values which can be affected, remove incorrect check,
> > split into drm include changes into separate patch, commit message,
> > verbose debugging statements during overwrite.
> > v3: white space fixes
> > v4: make path dependent on DPCD revision > 1.2
> > v5: split into function, removed DPCD rev check
> > v6: add debugging prints for early exit conditions
> > v7 (From Manasi):
> > * Memcpy, memcmp and debig logging based on sizeof(dpcd_ext) (Jani N)
> > * Exit early (Jani N)
> >
> > Cc: Jani Nikula <jani.nikula at linux.intel.com>
> > Cc: Ville Syrjala <ville.syrjala at linux.intel.com>
> > Signed-off-by: Matt Atwood <matthew.s.atwood at intel.com>
> > Tested-by: Manasi Navare <manasi.d.navare at intel.com>
> > Acked-by: Manasi Navare <manasi.d.navare at intel.com>
> > Reviewed-by: Rodrigo Vivi <rodrigo.vivi at intel.com>
> > ---
> >  drivers/gpu/drm/i915/intel_dp.c | 41 +++++++++++++++++++++++++++++++++
> >  1 file changed, 41 insertions(+)
> >
> > diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
> > index 70ae3d57316b..a9eb14a4ab27 100644
> > --- a/drivers/gpu/drm/i915/intel_dp.c
> > +++ b/drivers/gpu/drm/i915/intel_dp.c
> > @@ -3802,6 +3802,45 @@ intel_dp_link_down(struct intel_encoder *encoder,
> >  	}
> >  }
> >  
> > +static void
> > +intel_dp_extended_receiver_capabilities(struct intel_dp *intel_dp)
> > +{
> > +	u8 dpcd_ext[6];
> > +
> > +	/*
> > +	 * Prior to DP1.3 the bit represented by
> > +	 * DP_EXTENDED_RECEIVER_CAP_FIELD_PRESENT was reserved.
> > +	 * if it is set DP_DPCD_REV at 0000h could be at a value less than
> > +	 * the true capability of the panel. The only way to check is to
> > +	 * then compare 0000h and 2200h.
> > +	 */
> > +	if (!(intel_dp->dpcd[DP_TRAINING_AUX_RD_INTERVAL] &
> > +	      DP_EXTENDED_RECEIVER_CAP_FIELD_PRESENT))
> > +		return;
> > +
> > +	DRM_DEBUG_KMS("DPCD: Reading extended receiver capabilities\n");
> 
> Superfluous debug logging.

Will get rid of this

> 
> > +
> > +	if (drm_dp_dpcd_read(&intel_dp->aux, DP_DP13_DPCD_REV,
> > +			     &dpcd_ext, sizeof(dpcd_ext)) != sizeof(dpcd_ext)) {
> > +		DRM_ERROR("DPCD failed read at extended capabilities\n");
> 
> Most of our dpcd failures are logged using DRM_DEBUG_KMS. The ones that
> log DRM_ERROR seem to be very recent additions deviating from the debug
> loggin practice. There isn't much the user can do, really.

Here this change from DEBUG_KMS to DRM_ERROR was as per Rodrigo's comment
on the initial patch. (https://patchwork.freedesktop.org/patch/240452/) 
Also IMO it should be an error since it will give unexpected results as we were
unable to get the true extended capabilities.

> 
> > +		return;
> > +	}
> > +	if (intel_dp->dpcd[DP_DPCD_REV] > dpcd_ext[DP_DPCD_REV]) {
> > +		DRM_DEBUG_KMS("DPCD extended DPCD rev less than base DPCD rev\n");
> 
> Okay, seems like a rare event.

Again this check and logging comes from Rodrigo's review comments on the initial patch.

https://patchwork.freedesktop.org/patch/240452/

> 
> > +		return;
> > +	}
> > +	if (!memcmp(intel_dp->dpcd, dpcd_ext, sizeof(dpcd_ext))) {
> > +		DRM_DEBUG_KMS("Extended Receiver Cap DPCD match the base DPCD\n");
> 
> I don't think this debug logging is needed.

Sure will get rid of this.

> 
> > +		return;
> > +	}
> > +
> > +	DRM_DEBUG_KMS("Base DPCD: %*ph\n", (int)sizeof(dpcd_ext), intel_dp->dpcd);
> 
> Using sizeof(dpcd_ext) when printing something else is a red flag. You
> could log the whole dpcd here.
> 
> 	DRM_DEBUG_KMS("DPCD: %*ph (base)\n", (int) sizeof(intel_dp->dpcd), intel_dp->dpcd);

Yes will do this.

> 
> > +	DRM_DEBUG_KMS("Extended Receiver Cap DPCD: %*ph\n",
> > +		      (int)sizeof(dpcd_ext), dpcd_ext);
> 
> The caller will log the *updated* DPCD right after this returns, you
> don't need to log dpcd_ext.

Okay agreed, I will get rid of this debug print since the caller already prints it.


> 
> > +	memcpy(intel_dp->dpcd, dpcd_ext, sizeof(dpcd_ext));
> > +}
> > +
> > +
> 
> Superfluous newline.

Will remove.

> 
> >  bool
> >  intel_dp_read_dpcd(struct intel_dp *intel_dp)
> >  {
> > @@ -3809,6 +3848,8 @@ intel_dp_read_dpcd(struct intel_dp *intel_dp)
> >  			     sizeof(intel_dp->dpcd)) < 0)
> >  		return false; /* aux transfer failed */
> >  
> > +	intel_dp_extended_receiver_capabilities(intel_dp);
> > +
> >  	DRM_DEBUG_KMS("DPCD: %*ph\n", (int) sizeof(intel_dp->dpcd), intel_dp->dpcd);
> 
> One other alternative is to have
> intel_dp_extended_receiver_capabilities() return true if the cap exists
> and is different from current DPCD, and *all* DPCD logging would be done
> here. Both the old and the new. Like so:
> 
> 	DRM_DEBUG_KMS("DPCD: %*ph\n", (int) sizeof(intel_dp->dpcd), intel_dp->dpcd);
> 	if (intel_dp_extended_receiver_capabilities(intel_dp))
> 		DRM_DEBUG_KMS("DPCD: %*ph\n (ext)", (int) sizeof(intel_dp->dpcd), intel_dp->dpcd);
>

I like the firts idea better where we print the old DPCD before memcpy and caller prints the DPCD (new/old)
anyways after return so will stick to that.

Manasi
 
> BR,
> Jani.
> 
> >  
> >  	return intel_dp->dpcd[DP_DPCD_REV] != 0;
> 
> -- 
> Jani Nikula, Intel Open Source Graphics Center