All Activity

I actually dont remember the exact order of install, but from what i can recall first i compiled and installed libva-v4l2-request-HACK_HEVC.zip and followed that tutorial then i downloaded h618_hwdec.tar.gz from https://www.elektroda.pl/rtvforum/topic4018092.html#20840047, those are the libs i replaced in /lib/aarch64-linux-gnu/ install latest edge kernel from armbian-config (or enable rolling updates), download kernel hearders and enable gpu in device tree overlays to compile mesa first install deps than creat a python venv in mesa src directory to download latest meson in it sudo apt-get install bison build-essential ccache clang cmake flex git glslang-tools libclang-dev libdrm-dev libelf-dev libexpat1-dev libgl1-mesa-dev libgles2-mesa-dev libgoogle-perftools-dev libgtk-3-dev libllvm-18-dev libomxil-bellagio-dev libpciaccess-dev libsensors-dev libspirv-tools-dev libunwind-dev libva-dev libvdpau-dev libvulkan-dev libwayland-dev libx11-dev libx11-xcb-dev libxatracker-dev libxcb-dri2-0-dev libxcb-dri3-dev libxcb-glx0-dev libxcb-present-dev libxcb-randr0-dev libxcb-shm0-dev libxcb-sync-dev libxcb-xfixes0-dev libxdamage-dev libxext-dev libxfixes-dev libxrandr-dev libxshmfence-dev libxxf86vm-dev libzstd-dev llvm-18-dev meson ninja-build pkg-config python3-mako python3-pil wayland-protocols python3-venv python3 -m venv mesa_venv source mesa_venv/bin/activate pip3 install meson but if i tried to compile from inside the venv i was gettin a mako version error so i exited the venv (just open another terminal window) used this exact command to be able to build from outside the python3 venv PYTHON=~/src/mesa/mesa_venv/bin/python3 ~/src/mesa/mesa_venv/bin/meson setup build -Dbuildtype=release -Dprefix=/usr -Dgallium-drivers=panfrost -Dvulkan-drivers=panfrost -Dplatforms=x11 -Dvideo-codecs=all (replace '~/src/mesa/mesa_venv/bin/python3' and '~/src/mesa/mesa_venv/bin/meson' with the path where you cloned the source...) install any other dependencies it complains about then install with sudo ninja -C build install you can confirm with glxinfo -B gabriel@orangepizero3:~/$ glxinfo -B name of display: :0.0 display: :0 screen: 0 direct rendering: Yes Extended renderer info (GLX_MESA_query_renderer): Vendor: Mesa (0xffffffff) Device: Mali-G31 (Panfrost) (0xffffffff) Version: 25.2.0 Accelerated: yes Video memory: 3922MB ... i don't know if that libva hack is working but vainfo recognize gabriel@orangepizero3:~$ LIBVA_DRIVER_NAME=v4l2_request vainfo libva info: VA-API version 1.20.0 libva info: User environment variable requested driver 'v4l2_request' libva info: Trying to open /usr/lib/aarch64-linux-gnu/dri/v4l2_request_drv_video.so libva info: Found init function __vaDriverInit_1_20 libva info: va_openDriver() returns 0 vainfo: VA-API version: 1.20 (libva 2.12.0) vainfo: Driver version: v4l2-request vainfo: Supported profile and entrypoints VAProfileMPEG2Simple : VAEntrypointVLD VAProfileMPEG2Main : VAEntrypointVLD VAProfileH264Main : VAEntrypointVLD VAProfileH264High : VAEntrypointVLD VAProfileH264ConstrainedBaseline: VAEntrypointVLD VAProfileH264MultiviewHigh : VAEntrypointVLD VAProfileH264StereoHigh : VAEntrypointVLD VAProfileHEVCMain : VAEntrypointVLD and finally mpv (with yt-dlp installed) gabriel@orangepizero3:~/$ mpv --hwdec=drm --profile=fast https://www.youtube.com/watch?v=Dyzf2evNjwc (+) Video --vid=1 (*) (h264 1920x1080 29.970fps) (+) Audio --aid=1 --alang=eng (*) (opus 2ch 48000Hz) File tags: Uploader: Zen Mechanics Channel_URL: https://www.youtube.com/channel/UCC8n8afKsdaYuI7Pr0J4pgA Using hardware decoding (drm). AO: [pulse] 48000Hz stereo 2ch float VO: [gpu] 1920x1080 drm_prime[nv12] it is a bit confusing but i hope it helps...

I've recently upgraded my NanoPi M4 V2 to Armbian_25.2.1_Nanopim4v2_bookworm_current_6.12.13_minimal.img.xz and since that, I have been getting random crashing of it, to the point that I have to manually power it down and up again to get it back running again. The only thing I've noticed is that this normally happens when there is intensive disk usage (like downloading a bunch of docker images, or when downloading large files from the internet). Here's the output from armbianmonitor -u pedro@nanopim4v2:~/docker$ armbianmonitor -u Collecting info and sending to paste.armbian.com, wait... dmesg: read kernel buffer failed: Operation not permitted dmesg: read kernel buffer failed: Operation not permitted https://paste.armbian.com/jefilizufo Please post the URL in the forum where you've been asked for.

In the past we had dedicated image for rpi5 ... It could be added here: https://github.com/armbian/build/blob/main/packages/bsp/common/usr/lib/armbian/armbian-firstrun#L112-L122

The Gnome desktop Power off menu selection and `sudo shutdown now` both reboot rather than powering down.

yeah, frustration is quite clear, but probably it is manufacturer who should be most/only interested in that. that also cost quite a lot, so it is probably better to demand where money were paid to, i guess.

That is the expected behavior. You always need to "install" u-boot after a new version from apt. The apt package just provides the binaries that you can install.

Tried rk3318 debian 12 with linux kernel 6.1.63 or up to 6.12.39 and no hdmi output, I installed on SD card. Though debian 11 with kernel 5.15.35 works Interesting thing that debian 12 with linux kernel 6.16.0-rc3 worked Is there anything I can do to provide useful information to debug?

At the same time as mmcblk1 disappeared someone had changed the ethernet port over to the secondary port as I could not get into the machine and tried the second port which had been configured to the same static IP address as the primary port was on.

You are right. I just downloaded something from youtube and assumed it was H264. When I checked it in VLC in my PC, the codec is: AOMedia's AV1 video (av01) I picked another mp4, which I verified in VLC as H264-MPEG4, 720p, and it played with 10% CPU This test below is with Bookworm XFCE, compositor ON, Linux 6.15.4 with default configuration Then I repeat the test with a 1080p H264 mp4. mpv tries to use hardware acceleration, but the mpv is black and there's continuous DRM_IOCTL errors: Then I tried this inside /etc/mpv/mpv.conf hwdec=drm-prime drm-drmprime-video-plane=primary drm-draw-plane=overlay and the 1080p H264 starts playing fluidly with just 40% CPU (compositor off) With compositor on, I get same 40% CPU Same video glitches, with or without compositor, but I am getting hardware acceleration (although there's the hwdec unsupported message)

>PLEASE STOP! This is both a time and money nightmare for me. I am giving as much information as I can in order to give feedback in order to get support. What I am posting is not all nonsense ! Okay the Gnome downgrade was incorrect and was based on what ifnformation I could get at the time. Why the issue system put up the postgres tag I really don't know I removed it. It is now repeatedly demostratably kernel dependant. On 6.12.32-current-rockchip64 and 6.12.39-current-rockchip64 the display is failing, 6.12.22-current-rockchip64, and 6.12.17-current-rockchip64 the display is working fine. 6.15.1-edge-rockchip64 is also working fine. No sure about 6.15.7 6.16.0-rc3-edge and 6.16.1-edge are not even booting, with BananaPi-M7. I am going to get kernel output using USB UART for all these boots.

Problem solved... My ignorance... For inexplicable reasons, u-boot was updated without an error message, but actually NOT written to the SD card. The ancient version apparently cannot start without uImage. Now I have manually written the SD card - more precisely an image of it - with the new u-boot and then wrote a new SD card from that image, uImage can be deleted and it runs...

And it turns out that nothing can be done? For some reason, it works for everyone, but something happened to me and it doesn't work anymore.

moved to tvboxes

Just out of curiosity, you are trying to play content that is encoded in AV1. Are you sure that your device is equipped with an AV1 decoder IP and that the kernel has the appropriate driver?

The seller can't help me and sends me to technical support. Technical support sends me to this forum. I bought it here. https://www.ozon.ru/product/tsentr-hab-umnogo-doma-home-assistant-4-64-zigbee-ble-hub-1657561166/

I started building mesa 25.2 according to https://docs.mesa3d.org/install.html When I started meson, it complained about meson-1.0.1 is too old (Bookworm)... needs 1.4.0 Trixie will provide meson 1.7... is this what you did? Start with a self built Armbian-Trixie? When you type "replace the libs", do you mean to uninstall debian package, and install the just compiled source? libva2 - Video Acceleration (VA) API for Linux -- runtime Is this independent from ffmpeg (I only see libav in ffmpeg, not libva) Does mpv need any special configuration?

Everything works, but I don't know why (;-) I upgraded to a new kernel 6.12.34-current-meson64 and also to the current u-boot. If I take a look at a current armbian image, every image.. vmlinux.. etc. is available in the boot folder, but no uImage. In my boot folder, however, in addition to the these files in the current armbian image, in my boot folder is still is a uImage from 2022. uname -r -> 6.12.34-current-meson6 So the current kernel is loaded and everything works. But if I now delete the uImage from 2022, the computer no longer boots. How does a current armbian image manage to boot without uImage? Is a uImage created at the first start? Manually I have created a new uImage via mkimage -A arm -O linux -T kernel -C none -a 0x01080000 -e 0x01080000 -n "Linux" -d ./vmlinuz-* ./uImage and the computer boots again, without uImage and only with the vmlinuz files not. Long story - can I leave it like that, or is there a way to boot without uImage.

How did you flash the u-boot ? Were you able to run armbian with that u-boot and install on the NVME drive? Thanks

I have tried again with trixie, ensuring that the kernel patches were taken... no difference. Then I build a Bookworm Armbian image, with linux 6.15.4. It wasn't visible wether the kernel was downloaded from armbian, or the last one was included in my OS image. Just for fun, I tried it with Bookworm's default ffmpeg-5.1.6, with xfce with compositor off... and I got a 1080p MP4 playing with 100% mpv log: Note that mpv still complains about unsupported hwdec: drm I repeated the test with compositor ON: killall xfwm4 && xfwm4 --compositor=on --display :0.0 & Same result: CPU 100%, but nice full FPS video Then I tried with the ffmpeg-x.x.x-v4l2 request Adding the repo, updating, and checking the update: Then I tried playing the mp4, I got no complaint about "unsupported hwdec: drm", but the CPU was still 100% For the first time, I get some video glitches when I resize, and repeated messages: Failed to create scanout resource DRM_IOCTL_MODE_CREATE_DUMB failed: Cannot allocate memory compositor=on or off doesn't make a difference for me. mpv log: Summary: Armbian Trixie (and the correspoinding ffmpeg 7.1.1) will always result in "unsupported hwdec: drm", with default ffmpeg, or ffmpeg-v4l2 request Armbian Bookworm (ffmpeg 5.1.6) will never complain about "unsupported hwdec: drm", with default ffmpeg or ffmpeg-v4l2request. Still 100% CPU

use my works: h616.tar.xz

follow the official instructions, change these dts to user overlay, simplicifed deployment just use `sudo make install` h616.tar.xz

Currently I am running HA supervised on Debian 12 bookworm. Debian boots from USB drive without any issue. So your suggestion should work I believe. Thanks for the suggestion. You are quick as always to reply.

I would. Happened in the past but no clue which hw. Probably also something aw related.

This should have been addressed: https://github.com/armbian/linux-rockchip/pull/373 May take a while until an updated kernel is available via apt. You can always build a up-to-date package by yourself using the build framework.

OK what makes you think you got hacked?