[Openchrome-users] DFP problems on Samsung NC20

[Xavier Bachelot]

Matthieu Herrb wrote:
> Xavier Bachelot wrote:
>> Matthieu Herrb wrote:
>>> Hi
>>>
>>> since revision 742 of the driver (VX800 integrated TMDS support), I get
>>> a black screen on the NC20 if running without and xorg.conf file.
>>> Reverting that revision, or using an xorg.conf file that forces 'Option
>>> "ActiveDevice" "LCD"' fixes the problem.
>>>
>>> after looking a bit closer to the code, this option as the effect of
>>> ignoring the BIOS that pretends to support VIA_DEVICE_DFP.
>>>
>>> I don't know what the right fix could be, may be a new kind of model
>>> depend quirk table is needed to disable the TDMS support on specific models?
>>>
>>> The patch attached, only enables it if the ViaCardId entry for a card
>>> explicitely lists VIA_DEVICE_DFP. I don't know if it's the right
>>> approach or not.
>>>
>> Using a table doesn't really scale, we need better outputs detection imho.
>> I'd rather force the use of the ActiveDevice option to enable DFP for
>> now. There are only a handful of boards with DVI output anyway (Epia
>> M700, are there others ?).
>> The attached patch shouldn't enable DFP output unless ActiveDevice is
>> explicitly set to DFP. Does it do the job for you ?
> 
> Yes, that works for me too.
> 
Hi Matthieu,

Could you please provide xorg logs from runs both with and without the
patch applied and no xorg conf ? Something is wrong with the DFP
detection code and maybe it'll help understand why it detects a DFP on a
machine that don't have one and doesn't on a machine that does.

Thanks in advance,
Xavier