Re: [PATCH v2 14/21] drm/i915/dp: Disable link retraining after the last fallback step

[Imre Deak <imre.deak@intel.com>]

On Thu, May 23, 2024 at 06:28:12PM +0300, Ville Syrjälä wrote:
> On Mon, May 20, 2024 at 09:58:12PM +0300, Imre Deak wrote:
> > After a link training failure if the link parameters can't be further
> > reduced, there is no point in trying to retrain the link in the driver.
> > This avoids excessive retrain attempts after detecting a bad link, for
> > instance while handling MST HPD IRQs, which is likely redundant as the
> > link training failed already twice with the same minimum link
> > parameters. Userspace can still try to retrain the link with these
> > parameters via a modeset.
> > 
> > While at it make the error message more accurate and emit instead a
> > debug message if the link training failure was only forced for testing
> > purposes.
> > 
> > Signed-off-by: Imre Deak <imre.deak@intel.com>
> > ---
> >  .../drm/i915/display/intel_display_types.h    |  1 +
> >  drivers/gpu/drm/i915/display/intel_dp.c       |  4 ++++
> >  .../drm/i915/display/intel_dp_link_training.c | 22 +++++++++++++------
> >  3 files changed, 20 insertions(+), 7 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/display/intel_display_types.h b/drivers/gpu/drm/i915/display/intel_display_types.h
> > index bde518c843468..eb0cac3e27acf 100644
> > --- a/drivers/gpu/drm/i915/display/intel_display_types.h
> > +++ b/drivers/gpu/drm/i915/display/intel_display_types.h
> > @@ -1765,6 +1765,7 @@ struct intel_dp {
> >  		int max_lane_count;
> >  		/* Max rate for the current link */
> >  		int max_rate;
> > +		bool retrain_disabled;
> >  		/* Sequential link training failures after a passing LT */
> >  		int seq_train_failures;
> >  	} link;
> > diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c
> > index b72dbd7becb74..34d64fe3302ef 100644
> > --- a/drivers/gpu/drm/i915/display/intel_dp.c
> > +++ b/drivers/gpu/drm/i915/display/intel_dp.c
> > @@ -2950,6 +2950,7 @@ static void intel_dp_reset_link_params(struct intel_dp *intel_dp)
> >  {
> >  	intel_dp->link.max_lane_count = intel_dp_max_common_lane_count(intel_dp);
> >  	intel_dp->link.max_rate = intel_dp_max_common_rate(intel_dp);
> > +	intel_dp->link.retrain_disabled = false;
> >  	intel_dp->link.seq_train_failures = 0;
> >  }
> >  
> > @@ -5061,6 +5062,9 @@ intel_dp_needs_link_retrain(struct intel_dp *intel_dp)
> >  					intel_dp->lane_count))
> >  		return false;
> >  
> > +	if (intel_dp->link.retrain_disabled)
> > +		return false;
> > +
> >  	if (intel_dp->link.seq_train_failures)
> >  		return true;
> >  
> > diff --git a/drivers/gpu/drm/i915/display/intel_dp_link_training.c b/drivers/gpu/drm/i915/display/intel_dp_link_training.c
> > index 97d499e4b6ef7..375f59afd4dec 100644
> > --- a/drivers/gpu/drm/i915/display/intel_dp_link_training.c
> > +++ b/drivers/gpu/drm/i915/display/intel_dp_link_training.c
> > @@ -1165,10 +1165,8 @@ static int intel_dp_get_link_train_fallback_values(struct intel_dp *intel_dp,
> >  		new_link_rate = intel_dp_max_common_rate(intel_dp);
> >  	}
> >  
> > -	if (new_lane_count < 0) {
> > -		drm_err(&i915->drm, "Link Training Unsuccessful\n");
> > +	if (new_lane_count < 0)
> >  		return -1;
> > -	}
> >  
> >  	if (intel_dp_is_edp(intel_dp) &&
> >  	    !intel_dp_can_link_train_fallback_for_edp(intel_dp, new_link_rate, new_lane_count)) {
> > @@ -1187,7 +1185,7 @@ static int intel_dp_get_link_train_fallback_values(struct intel_dp *intel_dp,
> >  	return 0;
> >  }
> >  
> > -static void intel_dp_schedule_fallback_link_training(struct intel_atomic_state *state,
> > +static bool intel_dp_schedule_fallback_link_training(struct intel_atomic_state *state,
> >  						     struct intel_dp *intel_dp,
> >  						     const struct intel_crtc_state *crtc_state)
> >  {
> > @@ -1195,7 +1193,7 @@ static void intel_dp_schedule_fallback_link_training(struct intel_atomic_state *
> >  
> >  	if (!intel_digital_port_connected(&dp_to_dig_port(intel_dp)->base)) {
> >  		lt_dbg(intel_dp, DP_PHY_DPRX, "Link Training failed on disconnected sink.\n");
> > -		return;
> > +		return true;
> >  	}
> >  
> >  	if (intel_dp->hobl_active) {
> > @@ -1203,11 +1201,13 @@ static void intel_dp_schedule_fallback_link_training(struct intel_atomic_state *
> >  		       "Link Training failed with HOBL active, not enabling it from now on\n");
> >  		intel_dp->hobl_failed = true;
> >  	} else if (intel_dp_get_link_train_fallback_values(intel_dp, crtc_state)) {
> > -		return;
> > +		return false;
> >  	}
> >  
> >  	/* Schedule a Hotplug Uevent to userspace to start modeset */
> >  	intel_dp_queue_modeset_retry_for_link(state, encoder, crtc_state);
> > +
> > +	return true;
> >  }
> >  
> >  /* Perform the link training on all LTTPRs and the DPRX on a link. */
> > @@ -1518,7 +1518,15 @@ void intel_dp_start_link_train(struct intel_atomic_state *state,
> >  		return;
> >  	}
> >  
> > -	intel_dp_schedule_fallback_link_training(state, intel_dp, crtc_state);
> > +	if (intel_dp_schedule_fallback_link_training(state, intel_dp, crtc_state))
> > +		return;
> > +
> > +	intel_dp->link.retrain_disabled = true;
> > +
> > +	if (!passed)
> > +		lt_err(intel_dp, DP_PHY_DPRX, "Can't reduce link training parameters after failure\n");
> > +	else
> > +		lt_dbg(intel_dp, DP_PHY_DPRX, "Can't reduce link training parameters after forced failure\n");
> 
> The forced failure stuff is in later patch, is it not?

Yes, I had a different patch order earlier, will move the error->debug
message change there.

> 
> >  }
> >  
> >  void intel_dp_128b132b_sdp_crc16(struct intel_dp *intel_dp,
> > -- 
> > 2.43.3
> 
> -- 
> Ville Syrjälä
> Intel