raspberry pi boots before the AV-amplifier so at the time of boot there is really no display present that is why I have hdmi_force_hotplug=1
In this case you have to run "getedid create" to permanently install the edid - hdmi options in config.txt no longer have any effect. See also the release notes https://libreelec.tv/2022/03/09/lib…nges-since-le92
so long,
Hias
Hmm, I would have expected that the WM8804's reset line is wired to a GPIO and controlled by the driver to avoid such issues - but that doesn't seem to be the case (the hifiberry digi and digi pro overlays don't contain wlf,reset-gpio DT properties).
It's best you contact the manufacturer of your sound card, he should know how to deal with the issue.
so long,
Hias
I believe this should be reworked as well otherwise it's easy to fell into deadlock (when an USB boot is the only possible way for some reason). And in this case both bootloader upgrade and any bootloader configuration changes are blocked.
LE already cleans up bootloader update files in /flash during bootup if bootloader (self-)update didn't pick them up and/or remove them. This also won't deadlock, bootloader self-update had such a potential issue before timestamps were introduced, but this has been resolved long ago.
so long,
Hias
I could reproduce the issue, but it'll be a bit tricky to solve - we probably should rework bootloader update handling in the LE settings addon (which is quite a mess).
The problem is that rpi-eeprom-update embeds a timestamp in the .sig file, and as the systemd service runs prior to time being set via NTP it'll be before the "current" eeprom timestamp and thus the bootloader skips self-update when booting from USB.
Just ssh in and run "rpi-eeprom-update -a" as a workaround for now, this will work fine (at least if system time has been set via NTP).
so long,
Hias
Forage that was the output of rpi-eeprom-update, not rpi-eeprom-config.
But as the output shows that your bootloader eeprom is up to date I'd say everything is fine on your side 
so long,
Hias
I cannot reproduce this here with 10.0.2 running from SD card, Apr 2021 bootloader was successfully updated to Jan 2021 version as expected.
Are you booting from USB? Please also post the output of rpi-eeprom-config, if self updates are disabled this would explain the issue when trying to update with USB boot.
so long,
Hias
Can we download it? (tar.gz)
Yes, you can download the tar.gz, build it and help us testing
so long,
Hias
Running it for the afternoon, unfortunately it has the same issue as 10.0.1 where it keeps powering my AVR/TV back on after shutdown.
Please keep discussion in this thread to HDR, there's already a thread about the AVR issue- post your issue there:
The experimental patch is in my builds since a month ago, I don't use an AVR so can't help much here.
so long,
Hias
It's best you ask this question on the Retropie forum - I have no idea about retropie and what you need to do to build additional kernel modules for it.
so long,
Hias
And yes, disabling script.embuary.info seems to help, can't crash Kodi by navigating in video lists anymore.
Then I've reenabled embuary and got this error in kodi.log
Please report that to the addon author. It could be something in the addon or in the addon depends that triggers this issue.
Kodi switched to Python 3.9.10 about a month ago https://github.com/xbmc/xbmc/pull/20853 and one of the comments in the PR also mentioned kodi crashes on android if script.module.simplecache is used (which is also used by script.embuary.info).
So this doesn't seem to be LE specific and it's best to take it up to the relevant authors so they can investigate it.
so long,
Hias
If you want to get this solved please try the suggested addon change.
Also try with a clean install using stock estuary skin and then try installing addons one by one and carefully test.
I'm not seeing any crashes here on RPi4 - also using the youtube addon.
If you just switch to older builds nothing will change as no one of us devs could reproduce the issue so far.
so long,
Hias
Here's another crash log, produced with a build with DEBUG=Python3,kodi
The last lines in the log show that script.embuary.info ran just before the crash:
2022-02-16 17:22:12.053 T:1305 info <general>: CPythonInvoker(9, /storage/.kodi/addons/script.embuary.info/default.py): script successfully run
2022-02-16 17:22:12.155 T:1305 info <general>: Python interpreter stopped
2022-02-16 17:22:24.776 T:1306 info <general>: initializing python engine.Please try with that addon removed, it seems to cause crashes on other systems, too. See eg this bug report https://github.com/sualfred/script.embuary.info/issues/81
so long,
Hias
Note that you also have to build the sound card driver https://github.com/LibreELEC/slice-drivers
Here's my current WIP branch for kernels 5.10/5.15 https://github.com/HiassofT/slice-drivers/tree/kernel-5.10+
so long,
Hias
Why do you want DWC2? It's a quite limited USB controller and you're far better off using the XHCI controller you get from otg_mode=1.
so long,
Hias
Can you please test if this build boots fine with NVMe?
I've enabled CONFIG_BLK_DEV_NVME, if you can confirm that it works we'll include it in future LE builds.
so long,
Hias
We didn't work around it, latest RPi kernel/firmware and ffmpeg from jc's branch should be all you need.
I haven't seen the flip_done issue for quite a while now (both with 5.10 and 5.15 kernels).
so long,
Hias