Re: [PATCH 1/4] arm: omap: display: Create omapdrm inside omap_display_init

[Tomi Valkeinen <tomi.valkeinen-l0cyMroinI0@public.gmane.org>]

[-- Attachment #1: Type: text/plain, Size: 1471 bytes --]

On 13/09/13 12:39, Archit Taneja wrote:
> On Friday 13 September 2013 03:08 PM, Archit Taneja wrote:
>> On Friday 13 September 2013 02:54 PM, Tomi Valkeinen wrote:
>>> On 13/09/13 12:14, Archit Taneja wrote:
>>>> Move omapdrm device creation inside the omap_display_init so that we
>>>> can
>>>> correctly create the device based on the presence of omapdss.
>>>>
>>>> Originally worked on by Andy Gross.
>>>
>>> If the dmm device is present in the DT data, there is no need to create
>>> the dmm device. It's created automatically.
>>
>> Yes, that is done in a later patch.
>>
>>>
>>> Also, omapfb device is currently created the same way as omapdrm,
>>> independently of omapdss. Why is it a problem to have them like that?
>>
>> In a multiplatform config, we might have CONFIG_DRM_OMAP and
>> CONFIG_DRM_OMAP_MODULE selected.
> 
> I meant these configs might be selected even if the image is booted on
> am3xx platform.

Ah, I see. And the same omap_arch_initcall() is used for am3xxx also,
even if the DSS (and thus DRM) doesn't exist in the HW.

We have the same problem with omapfb, so it'd be good to include that in
the same series.

Hmm. If omap_generic_init() is called on am3xxx, it means we try to
create the dss stuff there also. It should fail to the DSS version
check, printing an error (at least I hope), but we really shouldn't even
call the dss init code on am3xxx.

I wonder how this should be fixed...

 Tomi



[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 901 bytes --]