Active threads

We already did that. I think the problem is that our keys are not packed but added at build time. When a person upgrades from Bookworm to Trixie, this will show up. https://github.com/armbian/configng/commit/5d866b9b105bfd46cb341c21c70ba76e32e1fea2 We are signing with old and new key (which is aligned with current standards) since then.

Hi everyone, I recently got an ESP-WROOM-32 Development Board to use as a companion board for my Armbian SBC setup. My goal is to have the ESP32 collect sensor data and send it to the Armbian host over serial or WiFi. It works okay at first, but after a while running a Python script, the ESP32 crashes or stops sending data. Setup Details: Armbian running on a Banana Pi (or similar SBC) ESP-WROOM-32 connected over serial (USB-to-TTL) and also configured via WiFi Python script (on the Armbian side) reads from the serial port / receives via a simple socket over WiFi Power supply is stable; using 3.3 V TTL logic on ESP and stable USB power Now here are my issues: Script runs fine for some minutes (5–10), then suddenly no data from ESP3. ESP32 sometimes reboots (observed via onboard LED), sometimes appears alive but unresponsive. Logs show no obvious error messages; serial just quiet or times out I've tried lowering the serial baud rate. Added delays / sleeps in code to reduce data flood. Ensured the USB-power to ESP32 is stable / using good cable. Tried Py serial and different socket listener scripts Could USB-TTL or serial USB drivers under Armbian be causing issues (overheating, buffer overflows)? Would switching fully to WiFi (TCP/UDP) instead of serial help stability? Any tools or logging tips on Armbian to catch the crash (kernel logs, serial trace, etc.)? Thanks in advance for any advice. Jordan

I don't see it as the very latest bleeding edge kernel version. I see it as a newish Linux, which already has the armbian patches updated for. Some CPUs end up with newer or older Linux versions, due to varying levels of development success. For example, sunxi's armbian edge is 6.15.4 today, but the linux trunk is up to 6.17-rc6 ... (not available in Armbian)

I tried it on 6.16.7, and the log started working: [ 4.838111] rockchip-drm display-subsystem: [drm] fb0: rockchipdrmfb frame buffer device Error again after initialization: [ 5.118148] rockchip-drm display-subsystem: [drm] Cannot find any crtc or sizes [ 5.637839] rockchip-drm display-subsystem: [drm] Cannot find any crtc or sizes

Has anyone ever gotten the Radxa Zero 3w working in gadget mode? I bought a Zero 3w about a week ago and have been beating this horse most of the time. I can't get the device to connect via USB. A or C. When I try to connect (radxa.local) it will always go through wifi. If I have wlan0 disabled or even my internet off it just will NOT connect. It clearly is using wifi to connect even if I use the hostname. Maybe this is something simple but I and another engineer I work with have been scratching our heads about it. The obligatory AI search was an endless loop of everything we had already tried. Anybody have any advice?

@hexdump hey mate , do you know if new armbian image still uses an offset of 8192 sector ? i'm trying the workaround you proveded in first page for linux hanging excatly after 60 seconds , but it only worked with multitool.img . with new armbian images it doesn't even boot after applying the trick 🙁 thanks

What type of monitor do you have connected? If you have another monitor, I'd try that. What I find works best is a monitor that is native 1920x1080. I sometimes have issues with monitors that are higher resolution or lower resolution than a standard 1080P display. Otherwise finding the uart pads is the only other advice I have, so that you can see what is going on during the boot.

I was considering rebuilding the initrd image for orange pi 5 as I think there was an issue in what was put on the image file. But now I have an a system settup with 6.1.115 - I should be good.

Awesome, @guenter. I'm debugging a board support layer for the Home Assistant OS, which targets the H96 Max. The fabulous work found in this thread, has been both inspiration and foundation for the port to Buildroot. With your very welcome direction, guenter, I'll have a far more polished result, if I ever get my images to boot. Thanks, so much, one and all, M.

Hello, I am selling a working unit of Kobol Helios64 with the fix done, without drives. Based in Czech Republic, able to send within EU. Please DM me if interested.

Tried for several days to get this working with the latest kernel and headers (never got it working and always got the same linkonce error). So I just abandoned the idea of Armbian for x86 and installed Debian which did work, perhaps when the Motorcomm YT6801 driver hits mainline I'll give it another try.

I've tried a reboot by disabling the overlay line in armbianEnv.txt and I've got one of my 2 LEDs back. It seems that ledconf-3 is no longer working for me. @jock If you have some time to check it out, that would be great... Thanks again.

Visit us at the STMicroelectronics booth, where Collabora will highlight how the STM32MP2 chip empowers edge AI solutions for industrial applications. View the full article

current isn't ideal for any rk3588 hw anyway due to lack of hw features mainlined for now. edge or vendor working best.

Let's say the world has been extra certain to bring me IRL problems 😄, as far as a bootloader being present, the Libre Computer Tritium boards are as barebones as can be possible, so any issues should be shared across any/all other H5 boards and really not board specific.

@Dominik Wójt Hey, I've picked up my mxq box again looking to put a newer OS on it. It seems the image you provided works with HDMI. Thanks! However, if I build it from the normal armbian repo it does not work. Any chance you can share what changes you made?

after struggling to make keyboard in french as locale is not taken account sudo dpkg-reconfigure locales sudo dpkg-reconfigure keyboard-configuration sudo systemmctl restart console-setup.service setupcon <-- winner Screenshot taken with screengrab , How am I supposed to live without you screengrab the music..

TV Box Name: H96 MAX V56 CPU: RockChip RK3566 Armbian build file name: Armbian_community_24.8.0-trunk.6_H96-tvbox-3566_jammy_current_6.6.32_gnome_desktop.img.xz Kernel Version: 6.6.32 Kernel Version: 6.12.6 Distribution Installed (focal, buster, etc): Jammy Working Ethernet (Yes/No): yes Max Ethernet Speed that works (100/1000): 1000 Gigabit Does wifi work (Yes/No): Yes Does bluetooth work (Yes/No): depends on User WorkAround Does HDMI audio work (Yes/No): Yes Additional Comments (provide any additional information you feel is important): No TF-Card Boot and No USB boot all done by Rockchip Hardware flash Topic 4/32gb Topic 8/64gb

Hi everyone, I was experience some strange things with two similar devices, the first I installed in february this year, mesa v.24.x and EGL 1.4 and it was working fine The other one I installed this week and it turns out that it uses mesa v.25.x and EGL 1.5, although performance was great with most 3D apps (like glmark2) I had poor performance under sway/kitty or the retroarch menu for instance (something like 15 fps). It took me a few days to find a solution so if you have similar issues please try installing mesa with the kisak repository as it fixed my problems sudo add-apt-repository ppa:kisak/kisak-mesa -y sudo apt update sudo apt upgrade the last command should upgrade mesa to a more recent version than what's in the ubuntu ppa.

Hi, you need the radxa dedicated tool to do it : https://opensource.rock-chips.com/wiki_Rkdeveloptool Once it is compiled and installed, you need to start the device in "maskrom mode", to do that, push the maskrom button located under the board near the hdmi output then power the device while holding the button. The led should turn and stays on. To check if the device is ready, run the following command : rkdeveloptool ld it should print something like that : DevNo=1 Vid=0x2207,Pid=0x350a,LocationID=302 Maskrom it means the device is ready to receive data then you need to upload the bootloader : sudo rkdeveloptool db rk356x_spl_loader_ddr1056_v1.12.109_no_check_todly.bin and finally you can upload the armbian img : sudo rkdeveloptool wl 0 Armbian-XXX.img make sure your device is properly cooled before doing those operations because heat can freeze the process. you can find the loader and more info here : https://docs.radxa.com/en/zero/zero3/low-level-dev/rkdeveloptool there's a windows version of rkdeveloptool too you can find on their website.

That should work for me. Thanks to everyone, who have participated.

before: Orange Pi Zero (H+), CPU: Allwinner H3 (SUN8I 1680) Model: Xunlong Orange Pi Zero DRAM: 512 MiB TF card is Samsung Evo Plus 64GB; After writing the TF card with Win32DiskImager, the same issue occurred: normal boot-up and shutdown, but when executing a reboot, it gets stuck at the final stage of service termination before restarting. ...... [ OK ] Reached target shutdown.target - System Shutdown. [ OK ] Deactivated swap dev-zram0.swap - /dev/zram0. [ OK ] Reached target umount.target - Unmount All Filesystems. [ OK ] Reached target final.target - Late Shutdown Services. [ OK ] Finished systemd-reboot.service - System Reboot. [ OK ] Reached target reboot.target - System Reboot. [44967.125952] systemd-shutdown[1]: Failed to set timeout to 10min: Invalid argument systemd-shutdown[1]: Failed to set timeout to 10min: Invalid argument [44967.761002] reboot: Restarting system This is where it starts to lose responsiveness! after: Following someone's guidance, I was told to install or update u-boot: run armbian-install, then selected “Install/Update the bootloader on SD card (/dev/mmcblk0)”. After the update completed, needed to shut down the system and disconnect the power (the first time I performed a reboot here, but it got stuck again at the final stage of stopping services before restarting, just like before). so I had to disconnect the power, then plug it back in for an automatic boot-up. After that, performing a reboot worked normally without getting stuck). ...... [ OK ] Reached target shutdown.target - System Shutdown. [ OK ] Deactivated swap dev-zram0.swap - /dev/zram0. [ OK ] Reached target umount.target - Unmount All Filesystems. [ OK ] Reached target final.target - Late Shutdown Services. [ OK ] Finished systemd-reboot.service - System Reboot. [ OK ] Reached target reboot.target - System Reboot. [44967.125952] systemd-shutdown[1]: Failed to set timeout to 10min: Invalid argument systemd-shutdown[1]: Failed to set timeout to 10min: Invalid argument [44967.761002] reboot: Restarting system U-Boot SPL 2024.01-armbian-2024.01-S866c-P7738-Ha5c2-Vd711-Bb703-R448a (Aug 16 2025 - 02:09:48 +0000) DRAM: 512 MiB Trying to boot from MMC1 ns16550_serial serial@1c28000: pinctrl_select_state_full: uclass_get_device_by_phandle_id: err=-19 U-Boot 2024.01-armbian-2024.01-S866c-P7738-Ha5c2-Vd711-Bb703-R448a (Aug 16 2025 - 02:09:48 +0000) Allwinner Technology CPU: Allwinner H3 (SUN8I 1680) Model: Xunlong Orange Pi Zero DRAM: 512 MiB Core: 69 devices, 18 uclasses, devicetree: separate WDT: Not starting watchdog@1c20ca0 MMC: mmc@1c0f000: 0, mmc@1c10000: 1 Loading Environment from FAT... Unable to use mmc 0:1... In: serial,usbkbd Out: serial Err: serial Net: eth0: ethernet@1c30000 starting USB... Bus usb@1c1a000: sun4i_usb_phy phy@1c19400: External vbus detected, not enabling our own vbus USB EHCI 1.00 ...... You can see that immediately after the log entry "[44967.761002] reboot: Restarting system", the system begins booting.

Hello @Going, I mean If uboot is available on a sdcard plugged in, it will boot on it else on emmc, Sometimes when UUID are the same, it is the last device discovered that overlaps. Wehen cloning emmc UUID back to sdcard do'nt change. My mind was to have a multiboot like refind or ventoy that helps if a system crashed. having a failsafe system aside to fix it graphically can be a must. Changing /boot/boot.cmd that affect u-boot may do the work

Apologies for being late to the party, and please excuse me for commenting on something which may already have been fixed. What is the difference between enabling the flag, and allowing whatever apps to work if they can, and disabling it so they certainly can't work? Seems to me that enabling it would make most sense, since H3 SOCs support it.

Thank you all for your input. And a big thank to @djurny I will double check if end user can add static route on their Windows machine or whatever PC1 is. route add 10.10.10.2 mask 255.255.255.255 192.168.100.101 As this solution appears to work. Alternative way is bridging the NICs however Device1 requires separate subnet from NIC1.... Thank you again.