[Intel-gfx] [PATCH] drm/i915: encoder/connector internal structure rework

[ykzhao]

On Tue, 2010-03-02 at 11:15 +0800, Zhenyu Wang wrote:
> On 2010.03.02 11:01:41 +0800, ykzhao wrote:
> > > > Why? As from sdvo spec, I think both s-video/composite could exist,
> > > > and
> > > > for TMDS0/1 too. It just forces that when TMDS1 exists, there must be
> > > > TMDS0.
> > > 
> > > Hmm.. I doubt if more than one of CVBS/COMP/SVID could work simultaneously, even though there is a kind of SDVO TV card that have these 3 interfaces. blending them together as one connector called TV may be better. Just like the intgegrated TV, even though sometimes we can see more than one built-in connector ( 80% only have one external connector though ), they can't work at the same time.
> > 
> > I agree. It seems that the SDVO-TV card can report its capability that
> > it can support CVBS/SVIDE/COMP. As they use the same SDVO-TV encoder,
> > maybe they can't work simultaneously.
> > 
> > At most cases the SDVO-DVI card will export only one external connector.
> > Even when it can export more than one connector, the two connectors
> > can't work simultaneously as one SDVO encoder is shared by the two
> > connectors.
> > 
> 
> This actually doesn't conflict with current code, as we already assumed
> only one type of connector can work with SDVO encoder for a single time, 
> the 'attached_output' is set to response value for GetAttachedDisplays request.

Yes. The attached_output can reflect the external request value of
GetAttachedDisplays. And it can work.

But then we will create two connectors with the same connector type for
one SDVO encoder. Maybe one of the two connectors is never used. 

If we can group the same type of capability bit and create one
connector, we can update the correct attached_output according to the
response of GetAttachedDisplay request.

Thanks
    Yakui
>