I saw a mention about autostart.sh in /storage/.config and put my alias there, but it didn't work (no error, but no alias command either)
It did work, but only in the context/lifetime of the autostart.sh shell session. Once the script exited, the config died too.
The diagram shows you are using CE. Problems should be posted in the CE forum; because we have nothing to do with their images and have no knowledge of how audio routing has been wrangled/mangled for the Amlogic BSP codebase.
Everything is in our main GitHub repo.
I don't see how this is relevant to this particular case.
It's relevant because explicit configuration of things is a general design objective for the buildsystem. Whether that was/wasn't an option in the past is irrelevant because what's possible today takes prescedence.
I plan to leave the PR open for a while for people to show up with comments on the state of MPEG2 decoding.
Please consider updating service.bluetooth-audio so that it stores and applies the full Kodi PulseAudio device identifier:
If you've investigated the problem and already figured out the fix/solution, perhaps send us a pull request?
Our general preference with packages is to always be explicit in what we enable. That way package featuresets never magically change because some upstream maintainer changed the scope of "all" without us noticing.
Looks like it should changed to -Dvideo-codecs=all to enable all codecs rather than list them individually.
Our preference has always been to be explicit about what we build and support rather than use catch-all configs.
ConnMan owns resolv.conf and it exists in /run which is not persistent over reboots. You can configure things using the settings add-on which has a dbus-agent that speaks to connmanctl, or by using connmanctl from the CLI, or by cloning /etc/connman/main.conf to /storage/.config/connman_main.conf and editing defaults there (reboot to effect changes).
It might be better to enable the capability in code and allow people to disable through Kodi VAAPI settings if there are issues rather than the current block in code?
Is there a reason why it's disabled/missing in the Mesa package.mk?
There's no reason that I'm aware of, although for context I haven't paid much attention to x86_64 matters for about a decade. You are welcome to send a pull request. We'll leave it open for a bit and see what comments it attracts.
Clear SPI and boot LE from SD card. Establish that you can see the NVME drive, i.e. the kernel modules and device-tree stuff is there to support that function. Format the NVME drive with a single EXT4 partition 100% sized with a unique disklabel. (re)Mount /flash as writeable and reconfigure /flash/extlinux/extlinux.conf so disk=label=STORAGE points to the NVME disklabel instead. It will boot from SD card, but Kodi effectively runs form NVME so you see the benefit of faster I/O access.
2026-06-17 06:29:45.745 T:807 info <general>: Running on LibreELEC (community): nightly-20260612-5c96e68 13.0, kernel: Linux x86 64-bit version 7.0.8Please bump to a current nightly as the RTL8157 changes I can see were merged for Linux 7.1 and are not present in Linux 7.0
run lsusb -tv | paste and pastekodi and share the URLs please
I wonder if we can get the LibreElec team to add support for the RTL8157 chipset ?
The r8152 kernel driver shows signs of RTL8157 support so have a look at the system log for missing firmware messages. If that's all that's needed (or perhaps adding new USB id's) then it's trivial to add support.
NB: on the flip-side if it needs a downstream/vendor driver we'll politely refuse; it needs to be supported upstream.
Having "Adjust display HDR mode" disabled results in HDR media being tonemapped to SDR whereas enabled implies the hardware is HDR capable (but while the TV might be, the kernel DRM layer says otherwise) so Kodi doesn't tonemap and you see the normal washed out colours when HDR media is used with an SDR colourspace.
K22 is a big improvement on the previous HDR experience in Kodi but a lot of the underlying logic and config options are still based on K21 capabilities so there's some cleanup/reworking needed. Hopefully by the time K23 ships that's been done; along with more of the kernel level support for colourspace/bit-depth being implemented and matured.
The DP-1 connector shown in modetest output does list HDR_OUTPUT_METADATA support, but even if the silicon out lists DP-1 as an option there's no guarantee that it physically exists on the board and/or is wired up properly.
Have you checked the state of the pins inside the USB connector? .. wondering if it's a mechanical issue
Run "modetest | paste" and share the URL so we can see the DRM connector properties.
