Do you have a 3D-capable TV? tab/sbs are the two modes to output 3D (I don't think it'll work on RPi/kodi ATM though).
So just whitelist the 1920x1080 modes without tab/sbs.
so long,
Hias
Do you have a 3D-capable TV? tab/sbs are the two modes to output 3D (I don't think it'll work on RPi/kodi ATM though).
So just whitelist the 1920x1080 modes without tab/sbs.
so long,
Hias
May I ask what this means?
Drivers for RPi2 and 3 are currently not yet feature complete and stable.
LE10 uses a completely new and different graphics and video decoding stack. For quite some time RPi0-3 had some severe issues (which are mostly solved now) but so far we could do only a very limited amount of testing on RPi2/3 so there might be more bugs lurking.
One of the major missing features on RPi2/3 is GPU accelerated HEVC playback - which means HEVC is practically unusable. Not sure if or when this will be added, it's very tricky and other missing features like deinterlacing (which is also important for RPi4) have higher priority.
We'll look into providing LE10 testbuilds for RPi2/3, if you don't need HEVC these should already be quite usable - and feedback from testers will also help us find bugs.
so long,
Hias
Most likely you are hitting this issue: HDMI screen turns off when sustained high CPU usage on all cores · Issue #1598 · raspberrypi/firmware · GitHub
Due to internal clocking changes the RPi4 started clocking down the HDMI block too much when it throttles (eg when it got too hot or in undervoltage conditions) which led to "no signal".
LE 10.0.1 will contain a fix for that, but you still need to resolve the root cause (too weak power supply).
LE 9.2 was able to visually notify you about undervoltage conditions (by showing a lightning bolt on the screen) but unfortunately that's no longer that case with the new graphics driver in LE10 so you/we need to check the logs (we added the throttled state to pastekodi logs right at the beginning though so we can easily see if there was a power supply issue - these are still quite common).
so long,
Hias
Always be careful with overclocking, it can cause subtle issues and settings that work stable today can result in an unstable system after the next (firmware) update.
Of course you are free to overclock your RPis as much as you like, just keep in mind that if you experience some issue the first thing to check is if removing the overclock fixes it.
so long,
Hias
max_usb_current=1 has no effect on RPi4, you can drop that.
Overclocking capabilities and settings depend on your specific board/chip, you'll need to find out yourself if it works or not.
4kp60 output has some known issues (see release post), these will be fixed in 10.0.1
so long,
Hias
Your power-supply seems to be underrated. The logs show undervoltage reports which then cause the HDMI signal to drop out - and will likely cause all sorts of other issues.
I'd recommend using the official RPi power supply, this is known to be working fine.
so long,
Hias
Which platform and LE version are you using?
LE9.2 on RPi only supports the official RPi 7" touchscreen, on LE10 DSI support is still rather experimental. You might find some more info on the RPi forums, eg here Interfacing (DSI, CSI, I2C, etc.) - Raspberry Pi Forums
so long,
Hias
No, only the kernel is 64bit (to solve the dreaded OOM-killer issue), userspace is still running 32bit.
so long,
Hias
Thanks for the info!
I could reproduce the pixelated output issue here and pinged the RPi devs - it looks though as 480i/576i output via HDMI always had this issue on RPi4.
so long,
Hias
Normal video playback is fine, only files with severe corruption (like the DVB recording with tons of errors) can make kodi/ffmpeg crash.
so long,
Hias
LE 10.0.0 is around the corner, the fix will be included in 10.0.1
so long,
Hias
Please post the outputs of pastekodi and modetest | pastebinit
so long,
Hias
Thanks for the sample, fix should be in nightlies soon [le11] ffmpeg: update rpi patch by HiassofT · Pull Request #5569 · LibreELEC/LibreELEC.tv · GitHub
so long,
Hias
Try adding video=HDMI-A-1:720x576@50iD to the end of cmdline.txt (after "quiet", in the same line).
so long,
Hias
Use unix line endings (LF), not windows ones (CR+LF)
so long,
Hias
there is fixes in RPi 4B, USB attached HDD fails with the message "xhci_hcd 0000:01:00.0: WARN Cannot submit Set TR Deq Ptr" in dmesg · Issue #3981 · raspberrypi/linux · GitHub , can someone check if this is going to fix it as well for this issue ?
the xhci warning and OOM killer kicking in are two different issues - the logs in this thread don't contain any xhci warnings.
The common factor between those two issues is (disk) I/O - if you are doing lots of I/O and have a problematic drive you may run into both issues (but it looks like smartd or udisks doing periodic smart checks seems to be an important point in the xhci issue, so LE should probably be OK as it doesn't run those).
so long,
Hias
We have no plans to add any more changes to 9.2, even more so as the new rev 1.4 RPi4 seems to contain incompatible hardware changes and 9.2 won't boot on it.
If you really feel the need to run 9.2 you can simply change the kernel configs and build LE on your own - see the instructions in our wiki https://wiki.libreelec.tv/development-1/build-basics
so long,
Hias
Run "rpi-eepron-config -e" and change BOOT_ORDER to eg 0xf14 (try USB first, then SD) - see also here Raspberry Pi Documentation - Raspberry Pi Hardware
Using the RPi imager to create a bootable bootloader-update-sd-card and choosing the "USB Boot" variant should also work (haven't tried, but the description says "Boot from USB if available, otherwise boot from SD card"
so long,
Hias