Use the AMLGX "box" image and set the dtb name to boot in uEnv.ini, then trigger recovery mode so it searches for the boot scripts included in that image (which are similar but slightly different to legacy ones). The "lepotato" image is designed for booting an actual LePotato board using mainline u-boot which uses a completely different extlinux.conf boot layout.
Posts by chewitt
-
-
The short answer is 1) No, 2) No.
-
The Python 3.9 bump is causing issues in a few places - notably one of the caching add-ons that's consumed by some streaming and skin add-ons. It's not anything specific to my images (or LE) and will probably get resolved upstream soon enough.
-
Code
2022-02-21 09:48:57.848 T:1008 INFO <general>: VideoPlayer::OpenFile: /var/media/sda1-usb-General_USB_Flas/test/kodi/iphone6s_4k(h.264).movRPi hardware (all generations including RPi4) does not support 4K H264 hardware decode, so the file is being software decoded and that's going to suck and drop frames like crazy.
-
If you want to disable MPEG2 in the driver that patch will still work (nothing changed in the driver). One of the HardKernel forum regulars has started to look into it .. but he's still in the early stages of tracing what the code actually does.
-
If anyone is using a Tanix TX3-mini box (S905W) the VFD display is now supported in my test images with upstream code! .. as this series was posted to the mailing list https://patchwork.kernel.org/project/linux-…[email protected]/ and I've added the super-simple clock script here https://github.com/chewitt/LibreE…401de6c2979fdd7 to put some numbers on it. In time (and assuming the driver patches are merged) someone with a bigger clue about Python than me will be able to resurrect the fd628.service add-on we've had in the past as that was fancier, but that involves Python, which I suck at. NB: There's also no guarantee the above driver will work on other boxes with VFD bits, but it's a start in the right direction
-
same here to be honest - rebooting after every episode watched so I switched to mainline nightlies instead
Provide a proper debug log.
-
It's not a complete debug log. There is no explanation of what hardware the problems is seen on. The filename shows it's something you've probably downloaded from a torrent site. Oh dear..
-
chewitt maybe these files stay somewhere on the server and you could share?
I use my own images for testing, not nightlies, so I don't have old images anywhere and the current test server holds one or two days files and then deletes them. I also have no stability issues (within the normal range of what works/doesn't-work with the vdec) so your problems are probably a skin or some add-on not being happy about the recent Python 3.9 bump in master.
-
48Hz movies are certainly bad enough.
No TV panels have native modes at that refresh though, so you won't see it advertised via EDID or used, unless you forced it somehow and then the bad results are entirely your own problem.
-
Look at the overlays configuration in cmdline.txt for the LE 9.2.8 image and replicate that in boot params with the other distro. If the kernel versions match you can take our overlay files too. If they don't .. the sources for the overlays can be found in our repo and you'll have to compile them to match the kernel the other distro is using.
-
Amlogic G12B hardware (inc. VIM3) has working H264, working 8-bit VP9 and working but imperfect HEVC decoding in the LE codebase and no support for 10-bit media or encoding. I run zero testing with conventional desktop OS like Ubuntu so cannot speak for how well our patchset might work in browsers or with WebRTC applications.
https://github.com/LibreELEC/LibreELEC.tv/ <= all sources (URLs or patches) are clearly listed here. If you want proper support and assistance with the upstream codebase I'd suggest you reach out to the Amlogic kernel maintainers at Baylibre who authored most of the drivers - who would love someone to fund more man-hours of development time so they can add more functionality and finish/polish features.
-
The Slice DAC requires a device-tree overlay, but this is pre-configured in the LE 9.2.8 'Slice3' image, so what image are you trying to use?
-
The AMLMX image is more POC than usable. Most of the boot workflow is still rough/broken or being discovered (more u-boot archaeology needed) but if you know enough about LE boot processes and have UART cables attached then it's not too hard to bodge something and get something booted. S802/S805 should get Kodi up. S812 will boot but 8/10 boots will result in CPU stalls (disabling CPU cores to make it a single core device solves that so but there's been no Eureka moment yet). There's also no vdec at the current time. Otherwise Linux 5.17-rc4 boots and generally things sort of work. You're welcome to play around with the image, but don't expect too much.
-
Put me on the list for wanting tftp support.
Mainline u-boot already supports it (all the Amlogic boards in kernel-ci boot that way) so you should be able to write u-boot to SPI and boot the device (replacing petitboot). I have no personal experience with it so can't guide you, but it should be do-able.
-
-
emveepee I did some test boots from an SD card yesterday using the image in my web-share (5.17-rc4 and u-boot 2022.01) and it worked fine from a boot perspective. I'd recommend getting an emmc module; they're faster and more reliable than SD cards.
-
Τα ελληνικά σου είναι καλύτερα από τα δικά μου ελληνικά
