Posts by chewitt

NB: I rebased the AMLMX image (for Meson 8 devices) on LE master branch using Martin B's latest Linux 6.2-rc kernel yesterday. It's untested as I don't have a spare monitor at the moment, and prob. needs some manual fiddling to boot on some devices due to the prehistoric u-boot versions being used, but people are welcome to play: https://chewitt.libreelec.tv/testing/LibreE…88.0-box.img.gz

Quote from jim_p

Thank you for all the clarifications. I want to try a le11 nightly image on s905 sometime (= when I will feel better and done with all the medical tests), so does it contain the device tree for my box (mxqpro4k 1gb)? Also, can it boot from a usb stick, because I have no sdcards handy?

There's no specific device tree but there are a number of other GXBB (S905) box devices present and one will probably be close enough to work. The upstream kernel doesn't require RAM specs to be coded in the device-tree which makes experiments a little easier. The box will always boot from emmc (hardcoded boot priorities) but you should be able to use the 'toothpick' method to trigger a recovery mode search for boot.scr and loading the upstream kernel from SD or USB.

If you have further Q's let's take them to a separate thread (or the AMLGX general thread) to keep this one on-topic.

Our settings add-on submits pastes anonymously, so there's no method to delete pastes. The site does purge pastes at some point, but there appears to be no formal policy on when (that we ever found). To ensure pastes are available for as long as possible (sounds like a reasonable goal) I'd expect purge to be FIFO and triggered on disk consumption not paste age.

_emanuel_ We inherit "keep-power-in-suspend;" from the w400 dtsi SDIO node so I'm not sure what's wrong or missing there. I do note that you've told me the sdio-pwrseq node is "GPIO_ACTIVE_LOW" whereas the vendor dts shows "GPIO_ACTIVE_HIGH" .. not sure if that's relevant.

Please also see if https://github.com/chewitt/linux/…e0c5cd69ba1c6f0 gives control over the red power LED, power button, and the Volume up/down/enter keys now?

Have a look at /storage/.kodi/temp/kodi.log for errors/issues. My thoughts would be a corrupted DB file or something add-on related. Worst case stop Kodi and rename /storage/.kodi to /storage/.kodi-old and restart; then start copying back config/content items from the old install until you either find the problem or end up with a working install again.

_emanuel_ I've rebased the branch with the SDIO change and this Fixup! patch https://github.com/chewitt/linux/…6757ed536b40b9d .. please see if that resolves the boot hang for you?

Quote from Omina

I have two 3.5" 4TB hard disks, plugged into an external drive dock ... and used the Storage Spaces utility to combine the two drives into an 8TB volume (setting up RAID 1)

To be pedantic: If you combine 2x 4TB drives into an 8TB volume you have RAID0 (striping) not RAID1 (mirroring) so data is "at risk" under any OS because RAID0 provides zero redundancy.

Quote from mdofperth

Any ideas on how I fix this and get my IVACY subscription running on LibreELEC?

LE10/LE11 images use Python 3.x, the last release to use Python 2.x was LE9.2. So options are:

a) Persuade <insert name of whatever VPN service> to provide a Python3 compatible version of their add-on

b) Find and extract the connection details and credentials/certs used from the add-on and use them with e.g. Zomboided's add-on.

Option A has no further action from us (the add-on, the service, and everything related to it is someone else's problem). If you go for Option B redirect future Q's to Zomboided's add-on support thread.

You probably need to test the specific (not Open) SUSE images provided for these devices as those are more likely to have some tweaked power management code. If not, you're needing to read up on ACPI debugging and/or sending emails to the Linux kernel mailing list asking for help on how to debug it. Understanding power states is one of those less-easily understood bit of the kernel .. I don't have any docs I can point you to.

The Kodi shows that there is no HDMI source detected. You can try appending "HDMI-A-1:1920x1080M@60" to boot params (same as the 'ssh') but this will only give you limited resolutions. You an also try an LE11 nightly: https://test.libreelec.tv/11.0/RPi/RPi4/

jim_p As per the wiki, AMLGX images are for 64-bit hardware (S905 and up) and AMLMX is for Meson 8 (S802/S805/S182) although S812 is not functioning. There hasn't been much upstream work done on the Meson 8 codebase in the last year as the sole developer (Martin B) has been keeping himself busy with other things.

The last AMLMX pre-pre-Alpha test image I created was in Feb '22: https://chewitt.libreelec.tv/testing/ - It did boot on an Odroid C1 but I can't speak for anything else.

LE provides all core downloads via 3rd party mirror servers: https://releases.libreelec.tv/mirrorstats

It is possible for some users to download direct of our infrastructure, but those downloads are a minority and do not use enough bandwidth to concern or cause expense to the project. So we don't really care if you set rate limits or run at full speed. The other 99.99% of users will be downloading from a mirror; so all the bandwidth consumed is someone else's - donated for free for that purpose.

Almost all open-source projects do some variant of the same thing (hosting on GitHub, via mirrors, via some major repo, etc.) so while it's nice to be a good netizen, you're probably over-thinking this issue.

Quote from chewitt

NB: Userspace shutdown scripts are not the problem or solution here. The issue (and any resolution) will be in lower-level code.

^ I'll quote myself again.

Quote from lafleur

I'd like to run the latest LibreELEC image that used the vendor GPU drivers. Or is anyone aware of LE9.x images usable for the Radxa Zero ? They seem to have all disappeared in the end.

LE (or me) has never created an image using vendor GPU drivers for Radxa Zero. We switched to mesa/panfrost before support for the Zero was added upstream.

frakkin64 I'm not sure that bug would be found in LE images since we aren't using EFI boot code paths in u-boot? - I can see how it would show up with dekstop distros though.

Quote from knedlyk

I'm curious how to use widevine with pure aarch64 kernel and system.

Google ships 64-bit versions of Widevine for x86_64, and 32-bit versions for ARM. There are no known sources for 64-bit ARM libs.

Comparisons with another Linux OS (Ubuntu, Fedora, etc.) are relevant - if it works fine there, there's some level of proof that the hardware and firmware combo can play nice with Linux. Proof that it runs great under Windows isn't particularly insightful since a) You'd expect a product built to run Windows to play nice with Windows, b) The Windows power architecture is completely different to Linux.

Issues with Shutdown/Restart are normally related to BIOS/firwmare versions (supporting ACPI states correctly) and power/config settings in the BIOS/firmware (what's needed for Windows isn't necessarily what's needed for Linux). And some hardware is just broken and there's not much we can do about it. You might want to start by telling us what hardware you have?