[Intel-gfx] [PATCH 2/2] drm/i915/dp: Validate the compliance test link parameters

[Manasi Navare]

On Fri, Jun 02, 2017 at 01:34:10PM +0300, Ville Syrjälä wrote:
> On Thu, Jun 01, 2017 at 05:51:27PM -0700, Manasi Navare wrote:
> > Validate the compliance test link parameters when the compliance
> > test dpcd registers are read. Also validate them in compute_config
> > before using them since the max values might have been reduced
> > due to link training fallback.
> > 
> > Signed-off-by: Manasi Navare <manasi.d.navare at intel.com>
> > ---
> >  drivers/gpu/drm/i915/intel_dp.c | 29 +++++++++++++----------------
> >  1 file changed, 13 insertions(+), 16 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
> > index 832786d..cda0f0a 100644
> > --- a/drivers/gpu/drm/i915/intel_dp.c
> > +++ b/drivers/gpu/drm/i915/intel_dp.c
> > @@ -1678,12 +1678,15 @@ intel_dp_compute_config(struct intel_encoder *encoder,
> >  	if (intel_dp->compliance.test_type == DP_TEST_LINK_TRAINING) {
> >  		int index;
> >  
> > -		index = intel_dp_rate_index(intel_dp->common_rates,
> > -					    intel_dp->num_common_rates,
> > -					    intel_dp->compliance.test_link_rate);
> > -		if (index >= 0)
> > +		/* Validate the compliance test data */
> > +		if (intel_dp_link_params_valid(intel_dp, intel_dp->compliance.test_link_rate,
> > +					       intel_dp->compliance.test_lane_count)) {
> > +			index = intel_dp_rate_index(intel_dp->common_rates,
> > +						    intel_dp->num_common_rates,
> > +						    intel_dp->compliance.test_link_rate);
> >  			min_clock = max_clock = index;
> > -		min_lane_count = max_lane_count = intel_dp->compliance.test_lane_count;
> > +			min_lane_count = max_lane_count = intel_dp->compliance.test_lane_count;
> 
> This looks a bit questionable. What if eg. just the link_rate is out of
> bounds but the lane count is good?
>

Hi Ville,

Thanks for your feedback/concern here.
But since the test requests both link rate and lane count to be used for the
compliance run, even if one of the parameters are out of bound then we need to bail
from using that combination in the compliance test run and instead go with the
fallback values.

Thought?

Manasi
 
> I think this could be solved in a different way with the patch I posted
> maybe one year ago that allowed rate_index() to return a non-exact match,
> and by clamping the test_lane count.
> 
> > +		}
> >  	}
> >  	DRM_DEBUG_KMS("DP link computation with max lane count %i "
> >  		      "max bw %d pixel clock %iKHz\n",
> > @@ -3961,8 +3964,7 @@ intel_dp_get_sink_irq_esi(struct intel_dp *intel_dp, u8 *sink_irq_vector)
> >  static uint8_t intel_dp_autotest_link_training(struct intel_dp *intel_dp)
> >  {
> >  	int status = 0;
> > -	int min_lane_count = 1;
> > -	int link_rate_index, test_link_rate;
> > +	int test_link_rate;
> >  	uint8_t test_lane_count, test_link_bw;
> >  	/* (DP CTS 1.2)
> >  	 * 4.3.1.11
> > @@ -3976,10 +3978,6 @@ static uint8_t intel_dp_autotest_link_training(struct intel_dp *intel_dp)
> >  		return DP_TEST_NAK;
> >  	}
> >  	test_lane_count &= DP_MAX_LANE_COUNT_MASK;
> > -	/* Validate the requested lane count */
> > -	if (test_lane_count < min_lane_count ||
> > -	    test_lane_count > intel_dp->max_link_lane_count)
> > -		return DP_TEST_NAK;
> >  
> >  	status = drm_dp_dpcd_readb(&intel_dp->aux, DP_TEST_LINK_RATE,
> >  				   &test_link_bw);
> > @@ -3987,12 +3985,11 @@ static uint8_t intel_dp_autotest_link_training(struct intel_dp *intel_dp)
> >  		DRM_DEBUG_KMS("Link Rate read failed\n");
> >  		return DP_TEST_NAK;
> >  	}
> > -	/* Validate the requested link rate */
> >  	test_link_rate = drm_dp_bw_code_to_link_rate(test_link_bw);
> > -	link_rate_index = intel_dp_rate_index(intel_dp->common_rates,
> > -					      intel_dp->num_common_rates,
> > -					      test_link_rate);
> > -	if (link_rate_index < 0)
> > +
> > +	/* Validate the requested link rate and lane count */
> > +	if (!intel_dp_link_params_valid(intel_dp, test_link_rate,
> > +					test_lane_count))
> >  		return DP_TEST_NAK;
> >  
> >  	intel_dp->compliance.test_lane_count = test_lane_count;
> > -- 
> > 2.1.4
> > 
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx at lists.freedesktop.org
> > https://lists.freedesktop.org/mailman/listinfo/intel-gfx
> 
> -- 
> Ville Syrjälä
> Intel OTC