Re: [PATCH v1, 1/2] drm/mediatek: Fixup external display black screen issue

[CK Hu]

Hi, Yongqiang:

On Tue, 2019-12-03 at 15:43 +0800, Yongqiang Niu wrote:
> On Tue, 2019-12-03 at 13:48 +0800, CK Hu wrote:
> > Hi, Yongqiang:
> > 
> > On Wed, 2019-11-27 at 09:17 +0800, yongqiang.niu@xxxxxxxxxxxx wrote:
> > > From: Yongqiang Niu <yongqiang.niu@xxxxxxxxxxxx>
> > > 
> > > Problem:
> > > overlay hangup when external display hotplut test
> > > 
> > > Fix:
> > > disable overlay when crtc disable
> > 
> > I think you do two things in this patch. The first is to config layer
> > before component start, and the second is disable layer when crtc
> > disable. So separate to two patches.
> > 
> > > 
> > > Signed-off-by: Yongqiang Niu <yongqiang.niu@xxxxxxxxxxxx>
> > > ---
> > >  drivers/gpu/drm/mediatek/mtk_drm_crtc.c | 39 +++++++++++++++++++++------------
> > >  1 file changed, 25 insertions(+), 14 deletions(-)
> > > 
> > > diff --git a/drivers/gpu/drm/mediatek/mtk_drm_crtc.c b/drivers/gpu/drm/mediatek/mtk_drm_crtc.c
> > > index 4fb346c..7eca02f 100644
> > > --- a/drivers/gpu/drm/mediatek/mtk_drm_crtc.c
> > > +++ b/drivers/gpu/drm/mediatek/mtk_drm_crtc.c
> > > @@ -369,6 +369,20 @@ static int mtk_crtc_ddp_hw_init(struct mtk_drm_crtc *mtk_crtc)
> > >  	mtk_disp_mutex_add_comp(mtk_crtc->mutex, mtk_crtc->ddp_comp[i]->id);
> > >  	mtk_disp_mutex_enable(mtk_crtc->mutex);
> > >  
> > > +	/* Initially configure all planes */
> > > +	for (i = 0; i < mtk_crtc->layer_nr; i++) {
> > > +		struct drm_plane *plane = &mtk_crtc->planes[i];
> > > +		struct mtk_plane_state *plane_state;
> > > +		struct mtk_ddp_comp *comp;
> > > +		unsigned int local_layer;
> > > +
> > > +		plane_state = to_mtk_plane_state(plane->state);
> > > +		comp = mtk_drm_ddp_comp_for_plane(crtc, plane, &local_layer);
> > > +		if (comp)
> > > +			mtk_ddp_comp_layer_config(comp, local_layer,
> > > +						  plane_state, NULL);
> > > +	}
> > > +
> > >  	for (i = 0; i < mtk_crtc->ddp_comp_nr; i++) {
> > >  		struct mtk_ddp_comp *comp = mtk_crtc->ddp_comp[i];
> > >  		enum mtk_ddp_comp_id prev;
> > > @@ -385,20 +399,6 @@ static int mtk_crtc_ddp_hw_init(struct mtk_drm_crtc *mtk_crtc)
> > >  		mtk_ddp_comp_start(comp);
> > >  	}
> > >  
> > > -	/* Initially configure all planes */
> > > -	for (i = 0; i < mtk_crtc->layer_nr; i++) {
> > > -		struct drm_plane *plane = &mtk_crtc->planes[i];
> > > -		struct mtk_plane_state *plane_state;
> > > -		struct mtk_ddp_comp *comp;
> > > -		unsigned int local_layer;
> > > -
> > > -		plane_state = to_mtk_plane_state(plane->state);
> > > -		comp = mtk_drm_ddp_comp_for_plane(crtc, plane, &local_layer);
> > > -		if (comp)
> > > -			mtk_ddp_comp_layer_config(comp, local_layer,
> > > -						  plane_state, NULL);
> > > -	}
> > > -
> > >  	return 0;
> > >  
> > >  err_mutex_unprepare:
> > > @@ -607,10 +607,21 @@ static void mtk_drm_crtc_atomic_disable(struct drm_crtc *crtc,
> > >  	for (i = 0; i < mtk_crtc->layer_nr; i++) {
> > >  		struct drm_plane *plane = &mtk_crtc->planes[i];
> > >  		struct mtk_plane_state *plane_state;
> > > +		struct mtk_ddp_comp *comp = mtk_crtc->ddp_comp[0];
> > > +		unsigned int comp_layer_nr = mtk_ddp_comp_layer_nr(comp);
> > > +		unsigned int local_layer;
> > >  
> > >  		plane_state = to_mtk_plane_state(plane->state);
> > >  		plane_state->pending.enable = false;
> > >  		plane_state->pending.config = true;
> > > +
> > > +		if (i >= comp_layer_nr) {
> > > +			comp = mtk_crtc->ddp_comp[1];
> > > +			local_layer = i - comp_layer_nr;
> > > +		} else
> > > +			local_layer = i;
> > > +		mtk_ddp_comp_layer_config(comp, local_layer,
> > > +					  plane_state, NULL);
> > 
> > I'm confused with this part. The design of this loop is to set
> > plane_state->pending.enable = false and wait for irq handler to write
> > register. Why do you directly write register?
> > 
> > Regards,
> > CK
> 
> when cmdq enable, mtk_crtc->cmdq_client will be always true when crtc
> create, there is no chance for mtk_crtc_ddp_config process in ddp irq 
> callback function

I think this is a bug of patch [1] which has not been upstream yet. So
this part should be moved to that patch.

[1] https://patchwork.kernel.org/patch/11270637/

> > 
> > >  	}
> > >  	mtk_crtc->pending_planes = true;
> > >  
> > 
> > 
> 
>