Am I missing something?
Not sure if you read first post in this topic in details:
- wifi modules without mainline driver are not supported (mostly Realtek - RTL8189ETV, RTL8189FTV, RTL8822BS, etc.)
we don't want to maintain a ton of out of tree wifi drivers. As I said - you can make your own build with wifi, but driver won't be accepted in official repository.
Is it a matter of dts/dtb? A missing driver?
Mostly driver, but maybe dts too. I didn't check if wifi is already described in DT.
I found nothing related to this question in closed PRs on github.
Few PRs tried to add relevant driver (not OPi Lite specifically).
This is interlaced MPEG2 demo: http://jernej.libreelec.tv/videos/mpeg2/burosch1.mpg I'm sure you'll find more on net.
Any hints for places to look for?
Check closed PRs at LE repo and first post here. Each contain reason why it was rejected (and your PR will be rejected too if you'll include it).
If you take a look just a few posts above, you'll see memory problems on H3 board with 1 GiB of RAM... It might work for you but not in general.
mike2002 you should really disable windowing logging - it produces huge amount of entries which are rarely interesting. Anyway, I have no idea why out of memory happens. There is nothing specific in log...
Sorry, there is no plan to officially support boards with < 1 GiB RAM. There is one in nightly builds, but it won't be released as stable and it's not developed by me.
You can however try to build image yourself and even PR needed changes to official repo. Bare minimum that needs to be done is adding new entry here: LibreELEC.tv/uboot_helper at master · LibreELEC/LibreELEC.tv · GitHub
Note that OPi Lite was supported in my unofficial OpenELEC 7 port and AFAIK there were no issues with it (HW decoding works). LE 7 was made by someone else.
I confirm that wifi doesn't work on Pine H64... For Bluetooth I don't know.
Thanks! Bluetooth is provided by same chip, so it won't work either.
Lewis OTA won't update U-Boot, which includes a couple of other important binaries for system stability, suspend/resume and power on functionality. Because of that I don't recommend it. Next release will have updated version of these binaries...
@all
There are few boards that we are not able to test, so I would like to get status of following boards (playback including deinterlacing, network, USB, suspend/resume, ...):
Libre Computer ALL-H3-CC (H3 version)
Nano Pi M1 (H3)
Orange Pi 2
Orange Pi PC
Orange Pi PC 2
Orange Pi Plus
Pine64 LTS
Additionally, can someone confirm that WIFI and Bluetooth don't work on PineH64 model B (my board may be broken in this regard).
jdww don't double post, I replied in second thread
argh... it's been overlooked.
Both images use software CEC implementation... You can build addons yourself - just make sure that you checkout same sources.
On a side note, LE10 beta1 with Allwinner images was just released. I'm considering closing this topic and move towards one thread one issue style.
norberto_ nothing interesting there either. Which is the last image that works for you (that you know)?
AZukr make sure you have appropriate refresh rate set. If it set to 30 Hz, then yes, microstutters will be noticeable. At 60 Hz it should not be really noticeable. Best way to avoid them altogether is to have refresh rate set to multiple of fps (there is option to set this automatically before playback), but that's not always possible. Note that comparison to legacy RPi2 means nothing to me, whole playback stack has nothing in common with H6.
I'm interested in dmesg output, not kodi log. Anyway, I'm out of ideas.
norberto_ Please test all 3 updates here: Index of /test/cec/ Hopefully at least one of them will work for you. All contain some CEC driver modifications.
Note that addons repo always targets latest nightly images only (they can be few days out of sync when Kodi is updated).
EDIT: Please provide output of dmesg for each update.
FYI, although SoCs on Rock64 and Pine64 are from different companies, they use same family of HDMI controllers which is covered by same driver (DW-HDMI). This means that CEC behaviour should be same.
No, not really. I guess some digging around code and adding some debug output would be needed...
No CEC issue in the log...
