Thanks a lot for the logs!
As suspected the card isn't supported by the official linux kernel drivers but uses the out-of-tree CrazyCat drivers.This is good to know since we now know whom to bother with this issue.
I haven't used Live/Run mode myself yet but I thought it would store user/configuration data on the USB stick. So selecting a different DVB driver set might work. Could you give that a try?
so long,
Hias
Mit ein wenig Overclocking und entsprechender Kühlung (ein etwas grösserer Kühlkörper langt) läuft DVB-T2 HD (1080p50 HEVC) am RPi3B und RPi3B+ flüssig.
Für DVB-T/T2 am RPi ist der offizielle Raspberry Pi TV Hat sehr gut geeignet - der funktioniert einwandfrei.
Mit einem GPIO Extender, um den TV Hat etwas höher zu setzen, ist auch genug Platz für einen Kühlkörper - ich verwende hier die BGA-STD-065 von ABL (ca 28x28mm Grundfläche und ca 11mm hoch).
so long,
Hias
Kodi recevies the remote button presses as LIRC events, not keyboard events, and Kodi doesn't support longpress with those.
I'm not exactly sure how it was handled in LE 8.2 but if you got keyboard events from the remote it was most likely a bug. The LIRC translation by eventlircd is needed because Kodi doesn't support a lot of the buttons on remotes in keyboard mode - see the wiki for details Infrared Remotes [LibreELEC.wiki]
You can disable the LIRC translation via a custom 98-eventlircd.rules udev rule file but keep in mind that this is not officially supported and you might have to change the remote keymap as well so it contains only keycodes that kodi can handle. If you are prepared to go this route you can completely disable LIRC translation by eventlircd by creating an empty udev rule file:
: > /storage/.config/udev.rules.d/98-eventlircd.rulesso long,
Hias
udevil seems to have it's config file hardcoded to /etc/udevil/udevil.conf, which is located in readonly system and from a quick glance there doesn't seem to be an option to override/change that.
But I've already opened a PR to drop mount_point_mode udevil: don't change permissions of mounted filesystems by HiassofT · Pull Request #3165 · LibreELEC/LibreELEC.tv · GitHub
Mounting some filesystem should never result in any modifications of it and doing a chmod() is a really devilish action of udevil.
so long,
Hias
This is caused by udevil, more specifically because we set mount_point_mode = 0755 in /etc/udevil/udevil.conf LibreELEC.tv/udevil.conf at master · LibreELEC/LibreELEC.tv · GitHub
I'm, not exactly sure why this is done but it looks like an oversight to me - it's been this way since the very beginning when udevil was added.
so long,
Hias
Your system is fine. The important thing is the revision number, this allows you to distinguish between 2B, 3B, 3B+ etc.
Revision a020d3 means it's an RPi 3B+ - see here readme.md
so long,
Hias
The log shows that you are playing a movie with a DTS audio track and have DTS passtrhough enabled. So Kodi is transmitting the DTS stream as it is and the volume setting in kodi doesn't have any effect - this is what passthrough is meant to do so everything seems to be working as expected.
If you want multichannel audio output with volume control you have to use HDMI audio output and disable passthrough. Then kodi will decode the AC3/DTS streams, apply the volume setting and transmit multichannel PCM audio over HDMI.
so long,
Hias
I think you should be able to test by booting LE 8.90.009 from an USB stick and using live / run mode instead of the installer.
so long,
Hias
The fix mentioned above is included in today's Milhouse build: LibreELEC Testbuilds for RaspberryPi (Kodi 18.0)
Could you test that on your RPi3 and report back if it works?
so long,
Hias
Thanks for the report, we could reproduce the issue and found a solution for it (see libinput: add support for absolute pointer motion by lrusak · Pull Request #15032 · xbmc/xbmc · GitHub ) which should be included in the next LibreELEC release.
so long,
Hias
Could you test with the latest LE alpha release (currently 8.90.009)?
By default it uses the official DVB drivers from the Linux kernel. With the driver addon chooser you can switch to 2 alternative sets of DVB drivers: latest drivers from linux kernel is a backport of (currently only a bit) newer official kernel drivers and TBS / CrazyCat add some out-of-tree drivers and patches that are not included in the official Linux kernel. LE 8.2 always used the CrazyCat drivers.
It would be interesting to know which set of drivers is required for your card and if the IR problem shows up with all of them or only on some specific ones.
Please also grab dmesg output when testing with the various DVB drivers and post them here (just run "dmesg | paste" and post the URL).
so long,
Hias
Do you get the volume change notification in Kodi when you press the volume buttons and does that change the level of the GUI sounds?
I just tried it here, on LE 8.90.009 and RPi3 plus a Hifiberry Digi and it worked fine. I also checked that it worked as expected while playing audio or a movie.
Note that volume changes won't have any effect when passtrhough is enabled, in this case the unaltered signal will be output.
so long,
Hias
Short answer: yes and yes.
The "wake from halt" function is a feature of the RPi firmware / bootloader. If you short GPIO3 (pin 5 on the header) to ground the RPi will start up again. This works always (AFAIK).
The gpio-shutdown overlay configures a GPIO to send a KEY_POWER keypress. Kodi handles this like the power button on a PC, which by default brings up the shutdown menu. You can change that to do an instant shutdown via keyboard.xml.
so long,
Hias
I've changed the title for you to make it more "interesting" for other users.
1) We know it's LE - that's why you're here,
2) Cinergy DVB-S2 Dual Tuner shouldn't be relevant.
I beg to differ, Cinergy DVB-S2 Dual Tuner could be very relevant - could you change the title back?
This is nothing I've seen before, as 2 IR devices are created and as they are identical the same configuration will be applied to both devices - and both will decode the signals thus giving double button presses.
Unless the device really should have 2 IR receivers (which would be very strange) I suspect this could be a driver bug and should be reported upstream.
finsher which LE version are you using, and which platform? Are you using the CrazyCat / TBS drivers or the standard kernel drivers?
The easiest workaround is to add "ir-keytable -s rc1 -c" to .config/autostart.sh for now.
so long,
Hias
Thanks a lot for the feedback and for your patience testing all those builds!
I think we should be settled now, tomorrow's Milhouse build will include the necessary changes to make bluetooth working on your box and if I didn't mess up and break other stuff they will be included in the next LibreELEC alpha/beta release as well.
so long,
Hias
Ah, no, I'm quite sure you didn't do anything wrong and missing firmware files are also (very certainly) not an issue - brcm/BCM4345C0.hcd was also missing in Mint but BT worked without it.
The issue is very certainly that the serial driver needed by your Atom CPU/board was missing and that the number of serial ports was limited to 4.
Can you check if this build works? libreelec-generic.x86_64-9.0-devel-20181205221534-01837ae.tar
I've changed the serial kernel options to (about) match with current Debian/Ubuntu configs which adds hopefully all options needed for your box.
so long,
Hias
sandin could you test this build? libreelec-generic.x86_64-9.0-milhouse-20181204160711-#1203b-g1f14e69.tar
It includes additional bluetooth drivers, hopefully your broadcomm BT adapter works with that. Either way dmesg would be interesting.
so long,
Hias
sandin I noticed in your log that you still have /storage/.cache/kernel-overlays/90-my-firmware.conf present - you can drop that now.
dmesg isn't showing any bluetooth adapters, so probably yet another kernel config option is missing.
Can you check if you get Bluetooth with Linux Mint? If it works there please post dmesg, journalctl -a, lsmod and lspci output from the working setup. This should provide some more hints.
so long,
Hias
