[Intel-gfx] [PATCH] drm/i915: Ignore DRM_MODE_FLAG_DBLCLK flag with DP

[Ville Syrjälä]

On Mon, Feb 02, 2015 at 06:21:31PM +0100, Daniel Vetter wrote:
> On Mon, Feb 02, 2015 at 07:16:33PM +0200, ville.syrjala at linux.intel.com wrote:
> > From: Ville Syrjälä <ville.syrjala at linux.intel.com>
> > 
> > The DP spec has nothing at all to say about double clocked modes. One
> > might assume they don't exist, and if you think about the concept
> > doesn't make much sense since the link already runs at higher fixed
> > frequency. So let's drop the DRM_MODE_FLAG_DBLCLK checks and simply
> > use the mode as if it was not double clocked.
> > 
> > I've tested this on a Dell UP2414Q which claims to support 720x576i
> > and 720x480i double clocked CEA modes, and it seems perfectly happy
> > with both modes using the 1x clock.
> > 
> > Signed-off-by: Ville Syrjälä <ville.syrjala at linux.intel.com>
> 
> What happens when this is for a dp->vga sst dongle? Or an active dvi/hdmi
> dongle (not sure those exist ...), are we sure this wont upset an existing
> screen somewhere?

If it's a native DP sink, I'd expect it to not advertize the mode if it
can't do it. Unless of course we're actually supposed to double clock
with DP. As stated I can't find anything in the spec about such things.
Also the specs are rather vague on what kind of infoframes you're
supposed to send DP sinks. If we're supposed to send the AVI infoframe
then we could send the VIC and the sink could know it needs to deal
with the doubled pixels, but othewise there's no way to tell the sink
about this using any native DP mechanism. I wish the DP spec would 
be more clear on how it interacts with CEA-861, rather than just have
some vague references to it. I suppose I could try to enable the AVI
infoframe and see what happens...

As for dongles, VGA doesn't have a clock so I expect that should just
work (tm). For active DP->HDMI I suppose the dongle should realize it
needs to double up (assuming we're not supposed to double clock DP),
but I've not tried it. We do have one active dongle here so I should
be able to try it.


> -Daniel
> 
> > ---
> >  drivers/gpu/drm/i915/intel_dp.c | 6 ------
> >  1 file changed, 6 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
> > index eea9e36..4b0dadb 100644
> > --- a/drivers/gpu/drm/i915/intel_dp.c
> > +++ b/drivers/gpu/drm/i915/intel_dp.c
> > @@ -222,9 +222,6 @@ intel_dp_mode_valid(struct drm_connector *connector,
> >  	if (mode->clock < 10000)
> >  		return MODE_CLOCK_LOW;
> >  
> > -	if (mode->flags & DRM_MODE_FLAG_DBLCLK)
> > -		return MODE_H_ILLEGAL;
> > -
> >  	return MODE_OK;
> >  }
> >  
> > @@ -1189,9 +1186,6 @@ intel_dp_compute_config(struct intel_encoder *encoder,
> >  						intel_connector->panel.fitting_mode);
> >  	}
> >  
> > -	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLCLK)
> > -		return false;
> > -
> >  	DRM_DEBUG_KMS("DP link computation with max lane count %i "
> >  		      "max bw %02x pixel clock %iKHz\n",
> >  		      max_lane_count, bws[max_clock],
> > -- 
> > 2.0.5
> > 
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx at lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/intel-gfx
> 
> -- 
> Daniel Vetter
> Software Engineer, Intel Corporation
> +41 (0) 79 365 57 48 - http://blog.ffwll.ch

-- 
Ville Syrjälä
Intel OTC