Active threads

Hello, I'm new to embedded linux installation, so please bear with me. So I purchased an android tv box labeled h96max S905L3 from TEMU. Specifications (if the label is trustworthy): RAM: 2GB, ROM:16GB, running android 14.0. On the pcb it says: s905_zx02_v01. It features an ethernet port, 2 USB, HDMI, AV port (looks like a headphone jack) and no hidden reset button or sd card slot. I was trying to install any kind of linux to it using a bootable usb, I just need a simple shell capable of running an http server. I have tried using the remote control to trigger boot to linux, but nothing works. I have then opened the thing, and connected uart to a RP2040 and wrote a simple script as a uart bridge. I now have access to u-boot shell, interupted setup. So far I used chatgpt to guide me, but couldn't get it to boot properly. It told me how to boot manually the image from the usb, but it failed every time. It told me to set these env vars, and over time it showed me different addresses. It said the issues were related to overlapping of the addresses, or using address range that was reserved. It also said the dtb which I used, the meson-gxl-s905x-p212.dtb won't work, since the device always threw an error bad_magic, even though I checked fdt header, and the magic was correct. setenv kernel_addr_r 0x20000000 setenv ramdisk_addr_r 0x24000000 setenv fdt_addr_r 0x21000000 I would then load the files fatload usb 0:1 ${kernel_addr_r} linux-android.img fatload usb 0:1 ${ramdisk_addr_r} uInitrd fatload usb 0:1 ${fdt_addr_r} vendor-expanded.dtb Notice: I used different images, firstly the Image file, and the meson-gxl-s905x-p212.dtb. After failed boots, chatgpt told me to expand, i.e. add padding to the dtb (first with the meson...dtb, and after that with the vendor.dtb). It didn't work, so it told me to generate an android like image using mkbootimg \ --kernel Image \ --dtb vendor-expanded.dtb \ --cmdline "console=ttyAML0,115200 console=ttyS0,115200 root=/dev/sda1 rootwait rootdelay=10 rw" \ --base 0x0 \ --pagesize 2048 \ --kernel_offset 0x1080000 \ --dtb_offset 0x1000000 \ -o linux-android.img and boot using the extracted vendor.dtb (pulled from android) and using bootm ${kernel_addr_r} instead of booti ${kernel_addr_r} ${ramdisk_addr_r} ${fdt_addr_r} The bootargs I used are setenv bootargs "root=/dev/sda1 rootfstype=ext4 rw rootwait console=ttyAML0,115200n8 console=tty0" Please, any help would be appreciated as I already spent over 6 hours with this little cheap bastard that wouldn't boot anything but that chinese android. I will post output of anything, just let me know if something is needed. I'm sorry for the messiness of this post, but I tried a lot of things that chatgpt told me, explained the errors, and supplied commands for the u-boot shell.

Thanks @Werner this is fixed, it works. In my case auto selected mirror is fi.mirror.armbian.de .

Hi, what image for this board

Workaround solution: downgrade to 2.3.2 apt install zfs-dkms=2.3.2-2 zfsutils-linux=2.3.2-2 libzpool6linux=2.3.2-2 libzfs6linux=2.3.2-2 libuutil3linux=2.3.2-2 libnvpair3linux=2.3.2-2 # block unwanted upgrade: apt-mark hold zfs-dkms=2.3.2-2 zfsutils-linux=2.3.2-2 libzpool6linux=2.3.2-2 libzfs6linux=2.3.2-2 libuutil3linux=2.3.2-2 libnvpair3linux=2.3.2-2

Armbian 25.11.2 Noble XFCE (BSP Kernel: 6.1.115) + PanVk - mesa 26.0 (https://launchpad.net/~ernstp/+archive/ubuntu/mesaaco) + box64 4.1 (https://ryanfortner.github.io/box64-debs/) + wine-10.20-staging-tkg-ntsync-amd64-wow64 (https://github.com/Kron4ek/Wine-Builds/releases/tag/10.20) + dgVoodoo2 (https://github.com/dege-diosg/dgVoodoo2/releases) + DXVK-stripped v2.1 (more likely to be crash-proof than other Dxvk versions) ~60fps@720p (Medium/High settings) box64 environment variables: Need for Speed - Underground 2

Hello, the link you shared has expired. Could you please re-analyze it

Hello, im trying to get my rk322x board with SV6051P chip to work, i tried every LED config but wpa_supplicant crashes on boot every time. Here are the dmesgs: dmesg_led8.logdmesg_led7.logdmesg_led6.logdmesg_led5.logdmesg_led4.logdmesg_led3.logdmesg_led2.logdmesg_led1.logdmesg_default_dto.log Anyway to fix this?

First testing on similar device (Odroid N2) didn't reveal any troubles. Need to locate VIM3 and try to recreate.

@chris178 SmartSolution on discord says his camera works. SmartSolution Hi, i had problem with camera and isp color quality After check and test download radxa-a733 t2 rsdk , image file and upload on mmc and after load system all camera works good

Duplicate

That's advanced as hell for build, but you can use any Rockchip SDK to make it, Some factoryes share the SDK online Some don't you will just compile a copy of the same file that you can find in the android image

And others: https://docs.armbian.com/Mirrors/#current-mirrors

It works with this old drivers in windows. CP2102 doesn't reach high baud rates in linux. Baud rate: 1500000 Data bit: 8 Stop bit: 1 Parity check: none Flow control: none ☑️ CP2104 TTL Tested the original one! ☑️ CP2102 TTL Tested the chinese fake one! with this Driver

I can confirm that booting Odroid M1 from SD card does not work with U-Boot versions newer than 2024.04. Many thanks to IAV for providing a workaround. Perhaps most people are booting from NVMe now, so this issue has flown under the radar? # Install U-Boot 24.5.1 (maps to upstream v2024.04) apt-get install -y linux-u-boot-odroidm1-edge=24.5.1 # Freeze the package at that version "printf 'linux-u-boot-odroidm1-edge hold\n' | sudo dpkg --set-selections" # Flash SPI directly from the package artifacts (or via armbian-install) flashcp -v /usr/lib/linux-u-boot-edge-odroidm1/idbloader.img /dev/mtd0 flashcp -v /usr/lib/linux-u-boot-edge-odroidm1/u-boot.itb /dev/mtd2 # Confirm strings /dev/mtd0 | grep -m1 "U-Boot" # U-Boot SPL 2024.04 (Jan 10 2026 - 10:54:44 +0100) strings /dev/mtd2 | grep -m1 "U-Boot" # FIT image for U-Boot with bl31 (TF-A) Now I can boot from sdcard again w/o RCY button. To use only ZFS on my system I compiled zfs into U-Boot. That worked but the problem was still with U-Boot's handling of the mmc1 somehow. Even though zfsls/zfsload were present in the SPI-flashed U-Boot, that build immediately reinitialized mmc1 in HS200 (1.8 V) mode and hit the “Card did not respond to voltage select! : -110” failure before the boot script could run. So the inability to boot the ZFS v28 pool was due to the SPI U-Boot’s bad MMC tuning, not the ZFS commands themselves. For future reference I’ll leave here on how to built U-Boot with ZFS support from the upstream sources: # Workspace & Dependencies apt install git gcc-aarch64-linux-gnu bison flex device-tree-compiler libssl-dev make python3 ninja-build build-essential mkdir -p ~/work/u-boot && cd ~/work/u-boot # Upstream U-Boot git clone --depth=1 --branch v2024.04 https://source.denx.de/u-boot/u-boot.git # Firmare blobs git clone --depth=1 https://github.com/rockchip-linux/rkbin.git # Prepare .config cd u-boot cp configs/odroid-m1-rk3568_defconfig .config scripts/config --file .config --enable CONFIG_CMD_ZFS yes "" | olddefconfig # Compile make -j$(nproc) \ BL31=../rkbin/bin/rk35/rk3568_bl31_v1.45.elf \ ROCKCHIP_TPL=../rkbin/bin/rk35/rk3568_ddr_1560MHz_v1.23.bin # Flash SPI directly from the package artifacts (or via armbian-install) cp spl/u-boot-spl.bin /usr/lib/linux-u-boot-edge-odroidm1/idbloader.img cp u-boot.itb /usr/lib/linux-u-boot-edge-odroidm1/u-boot.itb flashcp -v /usr/lib/linux-u-boot-edge-odroidm1/idbloader.img /dev/mtd0 flashcp -v /usr/lib/linux-u-boot-edge-odroidm1/u-boot.itb /dev/mtd2

If that worked for you, then you're lucky. For me it did not (other board tested). I guess the files on there have an older state which is still good. Again, as mentioned, repo is being worked on and we're aware of the issues and trying our best to resolve this asap.

Why is there no image for RK3288?

After a ton of time porting the dts for modern kernels, i have a working release for the Bobcat 300 29X Variant To compile https://github.com/sicXnull/armbian-build/tree/Bobcat-29x . ./compile.sh build BOARD=bobcat-29x BRANCH=current BUILD_DESKTOP=no BUILD_MINIMAL=yes EXPERT=yes KERNEL_BTF=no KERNEL_CONFIGURE=NO KERNEL_GIT=shallow RELEASE=bookworm Link to Prebuilt Armbian-unofficial_26.02.0-trunk_Bobcat-29x_bookworm_current_6.18.4.img.xz First check to see if you have a compatible version. Flip the unit over and check the serial number. It will start with G290 or 295. This does not work on 28 or 285. How to Flash 1) Install Rockchip Flash Tools 2) Download bootloader (rk356x_spl_loader_ddr1056_v1.10.111.bin) attached here. 3) Boot device into LOADER mode. Plug USB Cable in to back of the unit, powered off. Open the device. With it powered on hold the reset button for 1 second, then immediately let off and hold the recovery button down until the PC recognizes you in LOADER mode 4) Follow the flash procedure found here. WARNING While I2C/Spidev is functional, i did not install Helium or the Lora packet forwarder for testing. Flash this image at your own risk if you still wish to use helium mining. rk356x_spl_loader_ddr1056_v1.10.111.bin

Yep Already tried that:- "chatgpt+gemini have given lots of ideas like patches, hooks, extensions but none of them work. (grrrrrr!)." Didn't work probably because the source is not an ASCII text file but a large CPIO binary. But....I am just running a make right now which looks like it might work..... don't want to jinx it....

duplicate

This isn't a DMA problem. There are too many inconsistencies in the audio driver. The driver needs to be fixed first.

Unfortunately, after multiple tries the image provided didn't work. Here is the outputs of the commands asked to be runned with there respective tested images: dmesg | grep -iE "stmmac|sun8i|dwmac|emac|phy|mdio|motorcomm|yt8531" dmesg | grep -i "registered.*phy\|phy.*connect" ls -l /sys/bus/mdio_bus/devices/ cat /sys/bus/mdio_bus/*/phy_id 2>/dev/null find /sys/class/net -name "end0" -o -name "eth0" | head -1 | xargs -I {} cat {}/phys_switch_id 2>/dev/null ls /sys/firmware/devicetree/base/soc/ethernet*/mdio/ Armbian_community_25.11.0-trunk.437_Orangepi3-lts_trixie_current_6.12.47_minimal.img [NOT WORKING] https://pastebin.com/v4bqxkeU Armbian_community_26.2.0-trunk.100_Orangepi3-lts_trixie_current_6.12.63_minimal.img [NOT WORKING] https://pastebin.com/fG0YKepz Armbian_community_26.2.0-trunk.151_Orangepi3-lts_trixie_current_6.12.63_minimal.img [NOT WORKING] https://pastebin.com/4eEC9mMg Armbian-unofficial_26.02.0-trunk_Orangepi3-lts_trixie_current_6.12.63_minimal_phy-io-supply.img [NOT WORKING] https://pastebin.com/DnRgP4k5 Armbian-unofficial_26.02.0-trunk_Orangepi3-lts_trixie_edge_6.18.2_minimal.img [WORKING] https://pastebin.com/8qp2X9wC armbianmonitor -U Armbian_community_25.11.0-trunk.437_Orangepi3-lts_trixie_current_6.12.47_minimal.img [NOT WORKING] https://pastebin.com/iUKFu1u0 Armbian_community_26.2.0-trunk.100_Orangepi3-lts_trixie_current_6.12.63_minimal.img [NOT WORKING] https://pastebin.com/SUDx7yme Armbian_community_26.2.0-trunk.151_Orangepi3-lts_trixie_current_6.12.63_minimal.img [NOT WORKING] https://pastebin.com/QVU9q2La Armbian-unofficial_26.02.0-trunk_Orangepi3-lts_trixie_current_6.12.63_minimal_phy-io-supply.img [NOT WORKING] https://pastebin.com/qiE4u2PQ Armbian-unofficial_26.02.0-trunk_Orangepi3-lts_trixie_edge_6.18.2_minimal.img [WORKING] https://pastebin.com/2iWbYgqK For some reason, Armbian_community_25.11.0-trunk.437_Orangepi3-lts_trixie_current_6.12.47_minimal.img stopped working when it did before. I tested it 5 times, and none of those boots worked. I don't know if the times I tested it before worked because of the small chance of booting with an ethernet connection, but it stopped working after booting. Anyway, I just want to clarify the inconsistency, but Armbian-unofficial_26.02.0-trunk_Orangepi3-lts_trixie_edge_6.18.2_minimal.img is still consistently working after numerous boots. Also, while testing the images to provide the outputs, I realized something interesting, which for some reason I didn't notice before. After booting for the first time, the images downloaded from the official place, https://www.armbian.com/orangepi3-lts/, showed a message that explains that the image used is a “Developer Preview Build” and that it should not be used for production. I don't know if this is expected or is a mistake in publicizing the incorrect images, but now that I realized it, I felt that I needed to disclose it. Maybe that is the reason the Ethernet connection is not working. Armbian_community_26.2.0-trunk.151_Orangepi3-lts_trixie_current_6.12.63_minimal.img Welcome to Armbian_community! Documentation: https://docs.armbian.com | Community support: https://community.armbian.com/ IP address: Network connection timeout! Create root password: **** Repeat root password: **** Developer Preview Build This Armbian image was generated automatically for development and testing purpose. It may include unfinished features or unstable components. If you are not here to report issues or just test it, please do not use this image in production. Expect things to change - or even break - as improvements are made. Creating a new user account. Press <Ctrl-C> to abort Please provide a username (eg. your first name):_ Armbian_community_25.11.0-trunk.437_Orangepi3-lts_trixie_current_6.12.47_minimal.img Welcome to Armbian_community! Documentation: https://docs.armbian.com | Community support: https://community.armbian.com/ IP address: Network connection timeout! Create root password: **** Repeat root password: **** WARNING! You are using an automated build meant only for developers to provide construtive feedback to improve build system, OS settings or UX. If this does not apply to you, STOP NOW! Especially don't use this image for production since things might not work as expected or at all. They may break anytime with next update. Creating a new user account. Press <Ctrl-C> to abort Please provide a username (eg. your first name):_

A quite easy workaround until it's resolved up-upstream is to install and enable the Cinnamenu applet and disable the "vanilla" menu applet. Right-click on the taskbar/panel, choose "Applets" and the rest is straightforward. To place the Cinnamenu in the bottom left corner, right click on the taskbar/panel and enter Edit mode to be able to move stuff around, finish Edit mode and you're all set.

moved to off-topic since op is not using Armbian.

All clear for the PCIe MSI: it is something with that Intel 7260 card. Today, I got two other M.2 Wifis: MT7922 and RTL8822ce, both seem to work on current and also on edge. Thus, I consider this as a success. Here's some screen blurb. // Sven-Ola

Minor update, I hardcoded usage of uboot debugging option and removed cache to force compiling uboot. But output from bootloader remained the same: U-Boot SPL 2024.01_armbian-2024.01-S866c-P7738-Ha5c2-Vcc17-Bf389-R448a (Jan 08 2026 - 10:33:14 +0000) DRAM: 512 MiB Trying to boot from MMC1 Maybe detailed logging was already enabled? But it still does not give me any hints. Especially my board does not contain any additional slots like emmc, only for micro sd card (I don't understand why bootloader tries to use mmc1) UPD: I find out question with a little bit similar problem (hang out on mmc1) but need time to understand the reason, maybe it has similar issue