Openelec Build for OPI PC and 2 now with HW decoding

Dima · Published in 2016-5-2 15:16:15

jernej replied at 2016-5-2 14:56
When kodi is stopped, you must use "systemctl suspend" command. Can you please try this. Kodi alwa ...

So, with autostart.sh lircd (the buggy combination):
The systemctl suspend command worked, however, power button didn't work to put it to sleep.

slaine · Published in 2016-5-2 16:08:45

I have a problem with deinterleacing on pvr client (PVR VNSI). When the deinterleacing is on the horizontal lines appears on screen and I have problem with channel switching. When the deinterleacing is off the problems disappear.

jernej · Published in 2016-5-2 16:48:51

Could you post some kind of image for horizontal lines? HW deinterlacing is known to be buggy. I have a long term plan to fix this. However, few things must be done before.

jernej · Published in 2016-5-2 16:53:20

Red-Z replied at 2016-5-1 17:21
G'day,

I have some trouble with the latest build on OPi PC. After booting a fresh installation, the ...

It seems that Kodi wants to switch resolution on the fly due to some error in resolution detection. Maybe your TV doesn't support EDID (I'm not sure if that is possible). However, resolution switching on the fly isn't yet supported due to bug in Mali driver and workaround is needed. I have a long to do list but unfortunately this is not very high.

Dima · Published in 2016-5-2 17:23:17

Quick update after even more testing: systemctl suspend also works without killing kodi. And KEY_POWER also puts it into sleep. So there is something that breaks the power button to suspend when lircd is run from autostart.sh

Last one thing I tested: I put sleep 20 in autostart.sh code:

#!/bin/sh
(
sleep 20; \
lircd /storage/.config/lircd.conf;
) &

Copy code

This breaks initial function of the remote (don't know why), but the power button works. After waking up from a power button suspend, the remote works again!

So my suspicion is that Kodi does something wrong when lircd is running when it starts up.

I there's anything else that I could help to debug, let me know. I've tested every possible scenario I could imagine.

jernej · Published in 2016-5-2 17:44:58

Dima replied at 2016-5-2 10:23
Quick update after even more testing: systemctl suspend also works without killing kodi. And KEY_PO ...

I think there might be more than one issue, so we need to sort some things out. First of, which board do you have? Plus? When you said that suspend works, do you mean that also resume works? Do you have any kind of USB device like, but not limited to, mouse or keyboard connected during suspend/resume?

Dima · Published in 2016-5-2 18:00:35

jernej replied at 2016-5-2 17:44
I think there might be more than one issue, so we need to sort some things out. First of, which bo ...

I have the Orangepi PC.

Resume always works. It's the suspend that doesn't work for me using the power button.
At this moment, I do have the mouse plugged in. However, without the mouse it's the same result.

Actually, for me: I can live with this. I don't actually need the power button suspend, as long as resume works. I just try to help out as much as possible to find the "root cause" of this problem.

jernej · Published in 2016-5-2 18:13:02

Dima replied at 2016-5-2 11:00
I have the Orangepi PC.

Resume always works. It's the suspend that doesn't work for me using the ...

Ah, ok. I always though that you have plus board. USB is the issue on OPi2 and Plus boards.

Maybe crashing deinterlacing driver has something to do with it. After second resume, there is a lot of errors in dmesg also for other programs like systemd-sleep, but if this driver is disabled, there is no problems at all. Could you please build image without it (CONFIG_SUNXI_DI) and test? Kodi will still work, but there would be no HW deinterlacing.

Dima · Published in 2016-5-2 18:40:12

jernej replied at 2016-5-2 18:13
Ah, ok. I always though that you have plus board. USB is the issue on OPi2 and Plus boards.

Maybe ...

Oh, OK. Sorry for the confusion.

At this moment I cannot build a test image because for the next week I'll only have my ultrabook at my disposal (which is painfully slow to build images).
It would be strange if it's caused by the DI driver as all the problems are in fact on inital boot... After a suspend and wakeup, everything works just like it should (except DI). But would be a cool thing to test

Dima · Published in 2016-5-2 19:39:05

BTW: I saw this commit: 5ad82638fb5bb9f150f8c680e455128d5cd0ceb8 (Unify all images and enable full system updates without losing changes)
Does it mean we will soon be able to update without rewriting the microSD card? And maybe OTA?