Re: [Intel-gfx] [PATCH v2 02/12] drm/i915: Make the CRTC state consistent during sanitize-disabling

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

On Thu, May 04, 2023 at 05:38:12PM +0300, Ville Syrjälä wrote:
> On Thu, May 04, 2023 at 02:10:38AM +0300, Imre Deak wrote:
> > Make sure that the CRTC state is reset correctly, as expected after
> > disabling the CRTC.
> > 
> > In particular this change will:
> > - Zero all the CSC blob pointers after intel_crtc_free_hw_state()
> >   has freed them.
> > - Zero the shared DPLL and port PLL pointers and clear the
> >   corresponding active flags in the PLL state.
> > - Reset all the transcoder and pipe fields.
> > 
> > v2:
> > - Reset fully the CRTC state. (Ville)
> > - Clear pipe active flags in the DPLL state.
> > 
> > Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > Signed-off-by: Imre Deak <imre.deak@intel.com>
> > ---
> >  drivers/gpu/drm/i915/display/intel_modeset_setup.c | 13 +++++++------
> >  1 file changed, 7 insertions(+), 6 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/display/intel_modeset_setup.c b/drivers/gpu/drm/i915/display/intel_modeset_setup.c
> > index eefa4018dc0c2..2ca66e49d8863 100644
> > --- a/drivers/gpu/drm/i915/display/intel_modeset_setup.c
> > +++ b/drivers/gpu/drm/i915/display/intel_modeset_setup.c
> > @@ -88,13 +88,14 @@ static void intel_crtc_disable_noatomic(struct intel_crtc *crtc,
> >  	crtc->active = false;
> >  	crtc->base.enabled = false;
> >  
> > -	drm_WARN_ON(&i915->drm,
> > -		    drm_atomic_set_mode_for_crtc(&crtc_state->uapi, NULL) < 0);
> > -	crtc_state->uapi.active = false;
> > -	crtc_state->uapi.connector_mask = 0;
> > -	crtc_state->uapi.encoder_mask = 0;
> > +	if (crtc_state->shared_dpll) {
> > +		crtc_state->shared_dpll->state.pipe_mask &= ~BIT(pipe);
> 
> So that's basiclly unreference_shared_dpll(). I wonder if we can
> refactor that a bit to by usable here?

Yes, makes sense I can add a
intel_unreference_shared_dpll_crtc(const struct intel_crtc *,
				   const struct intel_shared_dpll *,
				   struct intel_shared_dpll_state *);
for this.

> 
> > +		crtc_state->shared_dpll->active_mask &= ~BIT(pipe);
> > +	}
> 
> Isn't this part of disable_shared_dpll()?

Yes, it is.

> Are we not calling that?

Yes, missed that the active flag gets cleared there. Only the above
unreference is needed then, will fix this.

> 
> > +
> > +	__drm_atomic_helper_crtc_destroy_state(&crtc_state->uapi);
> >  	intel_crtc_free_hw_state(crtc_state);
> > -	memset(&crtc_state->hw, 0, sizeof(crtc_state->hw));
> > +	intel_crtc_state_reset(crtc_state, crtc);
> >  
> >  	for_each_encoder_on_crtc(&i915->drm, &crtc->base, encoder)
> >  		encoder->base.crtc = NULL;
> > -- 
> > 2.37.2
> 
> -- 
> Ville Syrjälä
> Intel