[Intel-gfx] [PATCH 10/10] drm/i915: Leave FDI running after failed link training on LPT-H

[Ville Syrjälä]

On Fri, Dec 04, 2015 at 11:09:11AM +0100, Daniel Vetter wrote:
> On Tue, Dec 01, 2015 at 03:08:41PM +0200, ville.syrjala at linux.intel.com wrote:
> > From: Ville Syrjälä <ville.syrjala at linux.intel.com>
> > 
> > Currently we disable some parts of FDI setup after a failed link
> > training. But despite that we continue with the modeset as if everything
> > is fine. This results in tons of noise from the state checker, and
> > it means we're not following the proper modeset sequence for the rest of
> > crtc enabling, nor for crtc disabling.
> > 
> > Ideally we should abort the modeset and follow the proper disable
> > suquenced to shut off everything we enabled so far, but that would
> > require a big rework of the modeset code. So instead just leave FDI
> > up and running in its untrained state, and log an error. This is
> > what we do on older platforms too.
> 
> No, ideally we should light the pipe up and tell userspace through some
> uevent that something bad has happened.

That's not ideal from the hardware POV, and it's rather against the spec
even. But this patch does exactly that.

> Especially with async commits
> failing to light up the pipe looks to userspace like the kernel just
> randomly disabled it. And that usually results in random crashes when the
> next flip or vblank wait fails and everything comes crashing down. And
> even if we wouldn't -EINVAL these userspace would still have a problem
> since the flip or vblank it wants to wait for simply never happens and
> then it's just stalled.
> 
> We've had tons of problems with this in the past, and therefore removed
> all the in-kernel sources for shutting down a pipe. E.g. DP re-training
> also forces the pipe to on again, even if it fails.
> 
> Yup dp mst is an exception and I need to fix it eventually.
> 
> Cheers, Daniel
> > 
> > Signed-off-by: Ville Syrjälä <ville.syrjala at linux.intel.com>
> > ---
> >  drivers/gpu/drm/i915/intel_ddi.c | 24 +++++++++++++++---------
> >  1 file changed, 15 insertions(+), 9 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c
> > index 76ce7c2960b6..b312a47a0654 100644
> > --- a/drivers/gpu/drm/i915/intel_ddi.c
> > +++ b/drivers/gpu/drm/i915/intel_ddi.c
> > @@ -675,15 +675,16 @@ void hsw_fdi_link_train(struct drm_crtc *crtc)
> >  		temp = I915_READ(DP_TP_STATUS(PORT_E));
> >  		if (temp & DP_TP_STATUS_AUTOTRAIN_DONE) {
> >  			DRM_DEBUG_KMS("FDI link training done on step %d\n", i);
> > +			break;
> > +		}
> >  
> > -			/* Enable normal pixel sending for FDI */
> > -			I915_WRITE(DP_TP_CTL(PORT_E),
> > -				   DP_TP_CTL_FDI_AUTOTRAIN |
> > -				   DP_TP_CTL_LINK_TRAIN_NORMAL |
> > -				   DP_TP_CTL_ENHANCED_FRAME_ENABLE |
> > -				   DP_TP_CTL_ENABLE);
> > -
> > -			return;
> > +		/*
> > +		 * Leave things enabled even if we failed to train FDI.
> > +		 * Results in less fireworks from the state checker.
> > +		 */
> > +		if (i == ARRAY_SIZE(hsw_ddi_translations_fdi) * 2 - 1) {
> > +			DRM_ERROR("FDI link training failed!\n");
> > +			break;
> >  		}
> >  
> >  		temp = I915_READ(DDI_BUF_CTL(PORT_E));
> > @@ -712,7 +713,12 @@ void hsw_fdi_link_train(struct drm_crtc *crtc)
> >  		POSTING_READ(FDI_RX_MISC(PIPE_A));
> >  	}
> >  
> > -	DRM_ERROR("FDI link training failed!\n");
> > +	/* Enable normal pixel sending for FDI */
> > +	I915_WRITE(DP_TP_CTL(PORT_E),
> > +		   DP_TP_CTL_FDI_AUTOTRAIN |
> > +		   DP_TP_CTL_LINK_TRAIN_NORMAL |
> > +		   DP_TP_CTL_ENHANCED_FRAME_ENABLE |
> > +		   DP_TP_CTL_ENABLE);
> >  }
> >  
> >  void intel_ddi_init_dp_buf_reg(struct intel_encoder *encoder)
> > -- 
> > 2.4.10
> > 
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx at lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/intel-gfx
> 
> -- 
> Daniel Vetter
> Software Engineer, Intel Corporation
> http://blog.ffwll.ch

-- 
Ville Syrjälä
Intel OTC