EDID failed with Connect monitors

[Matt Waddel]

The 47" Connect monitors are not detected during boot so there is no display. These monitors work with other ARM low cost systems like Panda and Quickstart.

The entire boot message sequence is posted here.
https://pastebin.linaro.org/291/

[Lee Jones (lag)]

I see the same thing with: Philips 244E monitor using HDMI.

[Patrik Klinger (patrik-klinger)]

Is there any indication on what the effort would be to solve this bug?

[Anmar Oueja (anmar)]

This bug is critical to resolve for the Connect in Hong Kong since it affects the TV displays we use.

[Mathieu Poirier (mathieu.poirier-deactivatedaccount)]

A couple of things need to happen in order to fix this problem.

1. Fix failing i2c tranfers when reading the EDID - could be that the timeout is set too aggressively.
2. Fix MCDE driver to handle the resolution handed out by HDMIServices, something that is not happening for the specific TVs that are used in this case.

[Sunil Kamath (sunil-kamath)]

Mathieu and Vathsala to work together for this issue as we do not have 47 inch monitor in Bangalore. Mathieu to update on workaround details.

[Anmar Oueja (anmar)]

Actually that is Matt Wadell (not Mathieu)

[Mathieu Poirier (mathieu.poirier-deactivatedaccount)]

On 12-05-11 04:32 AM, Sunil Kamath wrote:
> Mathieu and Vathsala to work together for this issue as we do not have
> 47 inch monitor in Bangalore. Mathieu to update on workaround details.
>
I'm not aware of any work around.

[Matt Waddel (mwaddel)]

This doesn't technically fix the EDID problem, but there is a work-around to display 1920x1080 using the fbset command. I was planning on using that at the HK Connect if there isn't any progress on EDID.

[Anmar Oueja (anmar)]

Matt: Does that mean the workaround is yielding good display. I
thought you were having some kid of wavy lines at the top of the
screen. If the workaround is enough, we can switch to other bugs.

anmar

[Matt Waddel (mwaddel)]

The workaround is yielding wavy lines on my small (22") HDMI screen and since that is not the screen we will be using at the HK Connect I don't want to spend anymore time trying to fix it here.

Ricardo did not see that problem on the screens they were using at UDS (which would be similar to the big screens we would be using in HK). So the fbset parameters will probably require some tweaking once we try the actual screens, but I think it is enough for now.

[vathsala (vathsala-nagaraju)]

Mathieu :

https://pastebin.linaro.org/291/ : i am not able to access these logs. Could you forward it to <email address hidden>

1. Fix failing i2c tranfers when reading the EDID - could be that the timeout is set too aggressively.
2. Fix MCDE driver to handle the resolution handed out by HDMIServices, something that is not happening for the specific TVs that are used in this case.
Could you enable hdmi logs at kernel level in config

    Can you enable the HDMI logs at kernel level :
 Devicedriver-->graphicsupport-->Av8100 driver support--> HDMI and Av8100 debug messages
 Please send us the with logcat and kernel messages

[Anmar Oueja (anmar)]

I put in the pastebin Matt did in an open one here: http://paste.ubuntu.com/992449/

[Matt Waddel (mwaddel)]

Here is the bootlog for the most recent snapshot (today):

http://paste.ubuntu.com/993016/

Here is the bootlog for the "same" kernel with "HDMI and Av8100 debug messages" enabled:

http://paste.ubuntu.com/993017/

I didn't see much difference, but maybe I'm missing something.

Here's the kernel info I used:

https://github.com/jcrigby/packaged-linux-linaro-3.3-ci/commit/56e4eb39c08ee2f221140f6742110ad5f44eb200

set to commit: 5f27d8d2890dc406839b4aba4c2f262dcb118a52