Re: Display not working if omapdrm is builtin with 3.15rc

[Tomi Valkeinen <tomi.valkeinen@ti.com>]

[-- Attachment #1: Type: text/plain, Size: 1647 bytes --]

On 02/05/14 13:47, Koen Kooi wrote:
> Hi,
> 
> I've been trying to get DVI on my beagleboard XM to output a picture with mainline and with 3.15rc it still doesn't work. If I statically build in everything the drm encoders and connectors all go into deferred probe mode and fail, falling back to 1024x768. I with omapdrm and encoders/outputs/etc as modules deferred probe is avoided, far fewer error messages and a 1280x1024 resolution is being used. But still the monitor still says 'no signal' detected.
> 
> So what's the magic to get an actual signal on the hdmi connector?

Sigh, looks like my beagle xM died on me while testing this...

The reason for this is that the twl_gpio (which TFP410 uses) gets
initialized really late, so display things get deferred. Then, before
TFP410 has initialized, omapdrm starts. It finds VENC (tv-out) display
available, so it starts. After that TFP410 gets probed, but as DRM
doesn't support adding displays later, it is never actually added to
omapdrm.

Also, the VENC output doesn't even work by default if
CONFIG_DRM_OMAP_NUM_CRTCS is 1. It needs to be set to 2 to support the
second output (tv-out in this case). But my board died before I managed
to test this.

So, the probing is quite broken. And I have no good ideas how to fix it,
only bad ones...

A simple thing to do for testing is to either disable
CONFIG_OMAP2_DSS_VENC from the kernel, or remove the svideo-connector
from the board's .dts file. This results in tv-out not being available,
so when omapdrm starts and finds no displays, it'll defer its probe,
thus getting the DVI next time it probes.

 Tomi



[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]