Re: [RFC 2/3] drm/mediatek: add support for Mediatek SoC MT2701

[YT Shen]

Hi CK,
On Fri, 2016-05-13 at 11:59 +0800, CK Hu wrote:
> Hi, YT:
> 
> On Thu, 2016-05-12 at 19:49 +0800, yt.shen@xxxxxxxxxxxx wrote:
> > From: YT Shen <yt.shen@xxxxxxxxxxxx>
> > 
> > This patch add support for the Mediatek MT2701 DISP subsystem.
> > There is only one OVL engine in MT2701, and we have shadow
> > register support here.
> > 
> > Signed-off-by: YT Shen <yt.shen@xxxxxxxxxxxx>
> > ---
> 
> > @@ -385,12 +422,16 @@ static void mtk_drm_crtc_atomic_begin(struct drm_crtc *crtc,
> >  		mtk_crtc->event = state->base.event;
> >  		state->base.event = NULL;
> >  	}
> > +
> > +	if (priv->data->shadow_register)
> > +		mtk_disp_mutex_acquire(mtk_crtc->mutex);
> >  }
> >  
> > @@ -409,6 +450,11 @@ static void mtk_drm_crtc_atomic_flush(struct drm_crtc *crtc,
> >  	}
> >  	if (pending_planes)
> >  		mtk_crtc->pending_planes = true;
> > +
> > +	if (priv->data->shadow_register) {
> > +		mtk_crtc_ddp_config(crtc);
> > +		mtk_disp_mutex_release(mtk_crtc->mutex);
> > +	}
> >  }
> >  
> 
> I think it's better to call mtk_disp_mutex_acquire() and
> mtk_disp_mutex_release() as near as possible to prevent mutex locked for
> a long time. All HW register access of this atomic setting is in
> mtk_crtc_ddp_config(), so it's better to move mtk_disp_mutex_acquire()
> just before mtk_crtc_ddp_config().
> 
> Regards,
> CK
I'll do this in the next version.  Thanks for the review.

Regards,
yt.shen