[PATCH] drm/mediatek: hdmi: Filter interlaced resolutions

Daniel Vetter daniel at ffwll.ch
Wed May 24 12:24:51 UTC 2017


On Wed, May 24, 2017 at 05:20:45PM +0800, CK Hu wrote:
> On Tue, 2017-05-23 at 15:12 +0200, Daniel Vetter wrote:
> > On Tue, May 23, 2017 at 05:28:15PM +0800, CK Hu wrote:
> > > Hi, Bibby:
> > > 
> > > I've applied this patch to my branch mediatek-drm-fixes-4.12-rc1,
> > > thanks.
> > > 
> > > Regards,
> > > CK
> > > 
> > > On Tue, 2017-01-24 at 13:10 +0800, Bibby Hsieh wrote:
> > > > Current Mediatek DRM driver does not support interlaced mode, and
> > > > will hang if such resolution is used: Filter those to prevent
> > > > kernel hangs, until the DRM driver is fixed properly.
> > > > 
> > > > Signed-off-by: Bibby Hsieh <bibby.hsieh at mediatek.com>
> > > > ---
> > > >  drivers/gpu/drm/mediatek/mtk_hdmi.c | 2 ++
> > > >  1 file changed, 2 insertions(+)
> > > > 
> > > > diff --git a/drivers/gpu/drm/mediatek/mtk_hdmi.c b/drivers/gpu/drm/mediatek/mtk_hdmi.c
> > > > index 0e8c4d9..e33678d 100644
> > > > --- a/drivers/gpu/drm/mediatek/mtk_hdmi.c
> > > > +++ b/drivers/gpu/drm/mediatek/mtk_hdmi.c
> > > > @@ -1244,6 +1244,8 @@ static int mtk_hdmi_conn_mode_valid(struct drm_connector *conn,
> > > >  			return MODE_BAD;
> > > >  	}
> > > >  
> > > > +	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
> > > > +		return MODE_NO_INTERLACE;
> > > >  	if (mode->clock < 27000)
> > > >  		return MODE_CLOCK_LOW;
> > > >  	if (mode->clock > 297000)
> > 
> > You probably want to check out Jose's mode_valid work, since only
> > filtering in the connector's ->mode_valid callback isn't enough. You also
> > need to filter in ->mode_fixup (or some other place called at atomic_check
> > time).
> > -Daniel
> 
> Hi, Daniel:
> 
> Yes, I still need to filter in ->mode_fixup.
> 
> Hi, Bibby:
> 
> Because Jose's work plan to use mode_valid to replace mode_fixup, so
> it's better to refine this patch after Jose's work is done. And I would
> now remove this patch from my branch.

Btw if you have a revised version of your patch on top of Jose's series, a
tested-by on that would be awesome. The more people find it useful, the
faster it will land.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch


More information about the dri-devel mailing list