All Activity

@Nick AIs it possible to configure the front display using the Linux header you provided and the pin details I shared above? I know this is a big request, but if it's possible, it would make this setup complete. Since I am not a programmer, I tried using Google's AI, but it just took me in circles and landed me nowhere. If you require any additional log or configuration details please let me know.

It seems there is a 4GB limit. Someone forked llama.cpp (so you can run GGUF models) and found a way to get rid of the 4GB limit.

If you are looking for image generation, someone converted Stable Diffusion to run on the NPU.

Here is an IMG with just u-boot flash to it; https://github.com/pyavitz/binary/releases/download/images/k1-spacemit-boot.img Using it will bypass the SPI and allow you to boot from USB NVME EMMC. Granted, an Armbian IMG would need to have already been flashed or transferred to one of them. It could be useful if say, you b0rked the SPI and wanna attempt to fix it or to use the SD card as a chainload mech to boot into the NVMe.

Hi everyone, I'm running a few services on an Armbian-based SBC and need a reliable SMTP Service Provider for system alerts, backup reports, and application notifications. Has anyone here used SMTP Service Provider iDealSMTP with Armbian, or compared it with other SMTP solutions for lightweight ARM servers? I'm interested in deliverability, ease of setup, and long-term reliability. I'd appreciate any experiences or recommendations. Thanks!

I don't own an OPI RV2, but I do have a few K1's. I suggest trying without that NVMe attached as it may be a contributing factor to boot fails. Technically correct. But I've seen some peps attempt some odd things and then wonder why it doesn't work. With serial attached, stop auto boot with the SPACEBAR. From there run `nvme` and you'll be provided with options. Are you suggesting you wanna bypass SPI by using an SD card to boot into say an NVMe or USB? Should be easy enough. Partition an SD card with an 4MiB offset and flash the u-boot bins to it. U-boot should then check for what ever is available to boot from; USB SD NVME MMC. If there is no extlinux.conf, boot.scr or *.efi to boot from on the SD, it will look else where.

I remember reading somewhere that the top slot doesn't support booting from NVMe. If it's non-boot then it shouldn't matter.

@c0rnelius Thanks for your input. The SSD booting was not the only issue on my system. For you, does booting https://sven-ola.commando.de/privat-in/Armbian-unofficial_26.05.0-trunk_Orangepirv2_trixie_edge_7.1.0-rc3_minimal.img from a MicroSD card work? With or without any hardware in any M.2 slot. When I tested, it did not work. However I think the SSD was connected when I tried. I'd need to try again with it disconnected. As there is only one NVMe SSD on my RV2, it should not matter in which M.2 slot my SSD is? From my boot log I can't see if the SSD is on "Bus0" "Bus2", is there some u-boot command to tell? Regarding making a MicroSD as boot disk, do you have any thought how do it. With hours of testing my way I could probably figure it out, any pointer to speed up the process would be much appreciated.

Patched the .dtb for v6.1, currently not yet in final format. I will probably create a commit to Armbian with a .dts for our board. I am attaching the dtb/dts just to keep track of the progress. Here are results of the RKNN test launch: rknn_api/rknnrt version: 2.3.2 (429f97ae6b@2025-04-09T09:09:27), driver version: 0.9.8 model input num: 1, output num: 1 input tensors: index=0, name=input, n_dims=4, dims=[1, 224, 224, 3], n_elems=150528, size=150528, fmt=NHWC, type=INT8, qnt_type=AFFINE, zp=0, scale=0.007812 output tensors: index=0, name=MobilenetV1/Predictions/Reshape_1, n_dims=2, dims=[1, 1001, 0, 0], n_elems=1001, size=2002, fmt=UNDEFINED, type=FP16, qnt_type=AFFINE, zp=0, scale=1.000000 custom string: Begin perf ... 0: Elapse Time = 8.06ms, FPS = 124.12 ---- Top5 ---- 0.935059 - 156 0.057037 - 155 0.003881 - 205 0.003119 - 284 0.000172 - 285 So: CPU: OK GPU: OK via Panfrost NPU: OK, RKNN inference works on RKNPU rk3566-box-X88PRO20-npu.dtb rk3566-box-X88PRO20-npu.dts

Can you give more details about the limitations of this NPU? I thought that they would be using shared system RAM. because all models are multiple megabytes, Can you comment about the ability to convert any model to your desired NPU target with "AI model conversion" https://tinycomputers.io/posts/rockchip-rk3588-npu-benchmarks.html (search for "Conversion Performance (x86_64)") Maybe its possible to run LCM-Lora in an armbian device, if it has 16GB. It is already running in a 16GB laptop without NPU or GPU. https://forums.raspberrypi.com/viewtopic.php?t=398225&sid=2b66e2cebe37506dd74215ead3dd264c I am thinking about buying an Orange PI zero 3W (allwinner NPU), but if you have troubles with a 6 TOPs NPU, I will have more troubles with a 2 TOPs NPU

Refer to the experiences here: I think someone wrote that the mainline kernel has NPU support. Is this NPU going to load only models that are in the Rockchip model repository?

This week's updates center on new board enablement, Rockchip platform refinements, and tooling and kernel maintenance. Board support expanded across multiple silicon families, with the addition of Seeed Studio reComputer RK3576/RK3588 DevKits and the Anbernic RG DS handheld image. The EasePi A2/R2 received substantial revisions to its board configurations and device trees, alongside a vendor logo transition to Linkease. SpacemiT K1 boot support was updated, and per-SoC LINUXCONFIG separation was introduced for the TQ family to better isolate kernel configurations. Rockchip received the bulk of low-level improvements. Notable changes include AUX recovery for USB-C DP Alt Mode in the dw-dp driver, device-tree-based LED configuration for the r8169/r8125 controllers, and an updated patch ensuring stable PCIe Ethernet MAC addresses across many boards. Additional fixes resolve slow WiFi on the NanoPi R76S via SDIO SDR104, enable Bluetooth on the Orange Pi 5 Ultra edge kernel, and restore the tm16xx driver on current kernels. On the tooling and maintenance side, Armbian Imager 2.0 was released, the mainline kernel was bumped to 7.1-rc6, and the rtl8192eu driver was re-enabled following a cleanup of compilation warnings. A previously merged USB gadget NULL pointer fix was reverted pending further evaluation. #Armbian #EmbeddedLinux #Rockchip #SBC #KernelDevelopment Changesmainline: bump to 7.1-rc6. by @EvilOlaf in armbian/build#9922Add Anbernic RG DS board image and Anbernic vendor logo. by @crackerjacques in armbian/armbian.github.io#326Add jellyfin-ffmpeg release constraints and ip-wrapper for easepi-r2. by @ifroncy01 in armbian/build#9942arm64: dts: overlay: add radxa display 8hd for rock5t. by @Ken-Vamrs in armbian/linux-rockchip#498board-vendor-logos: update EasePi vendor logo to Linkease. by @ifroncy01 in armbian/armbian.github.io#321drm: rockchip: dw-dp: add AUX recovery for USB-C DP Alt Mode. by @mingzhangqun in armbian/linux-rockchip#494easepi-r2/a2: fix device tree. by @jjm2473 in armbian/linux-rockchip#491feat(boards): Add Seeed Studio reComputer RK3576/RK3588 DevKit support. by @baorepo in armbian/build#9719Give each TQ SoC family its own LINUXCONFIG. by @schmiedelm in armbian/build#9924Introducing Armbian Imager 2.0. by @SuperKali in armbian/imager#143Milk-V Jupiter disable EEPROM node. by @pyavitz in armbian/build#9944nanopi-r76s: fix slow WiFi by enabling SDIO SDR104. by @SuperKali in armbian/build#9929net: r8169: add device tree based LED configuration for r8125. by @mingzhangqun in armbian/linux-rockchip#497orangepi5-ultra: enable Bluetooth on edge kernel via hci_bcm. by @pdapandapda in armbian/build#9697Revert "rockchip64: fix USB gadget NULL pointer crash in eth_get_drvinfo (6.18 + 7.0)". by @EvilOlaf in armbian/build#9926rockchip64: enable back tm16xx driver for current kernel. by @paolosabatino in armbian/build#9941rockchip64: update "describe PCIe Ethernet controllers" patch for stable MAC (many boards). by @rpardini in armbian/build#9933rtl8192eu: re-enable and fix tons of compilation warnings. by @EvilOlaf in armbian/build#9927SpacemiT: Update K1 boot support. by @pyavitz in armbian/build#9930Treat imx93 as slow hardware. by @schmiedelm in armbian/armbian.github.io#323Update EasePi A2/R2 board configs and device trees for improved hardware support. by @ifroncy01 in armbian/build#9907Update VERSION. by @EvilOlaf in armbian/build#9938View the full article

This week's updates center on new board enablement, Rockchip platform refinements, and tooling and kernel maintenance. Board support expanded across multiple silicon families, with the addition of Seeed Studio reComputer RK3576/RK3588 DevKits and the Anbernic RG DS handheld image. The EasePi A2/R2 received substantial revisions to its board configurations and device trees, alongside a vendor logo transition to Linkease. SpacemiT K1 boot support was updated, and per-SoC LINUXCONFIG separation was introduced for the TQ family to better isolate kernel configurations. Rockchip received the bulk of low-level improvements. Notable changes include AUX recovery for USB-C DP Alt Mode in the dw-dp driver, device-tree-based LED configuration for the r8169/r8125 controllers, and an updated patch ensuring stable PCIe Ethernet MAC addresses across many boards. Additional fixes resolve slow WiFi on the NanoPi R76S via SDIO SDR104, enable Bluetooth on the Orange Pi 5 Ultra edge kernel, and restore the tm16xx driver on current kernels. On the tooling and maintenance side, Armbian Imager 2.0 was released, the mainline kernel was bumped to 7.1-rc6, and the rtl8192eu driver was re-enabled following a cleanup of compilation warnings. A previously merged USB gadget NULL pointer fix was reverted pending further evaluation. #Armbian #EmbeddedLinux #Rockchip #SBC #KernelDevelopment Changesmainline: bump to 7.1-rc6. by @EvilOlaf in armbian/build#9922Add Anbernic RG DS board image and Anbernic vendor logo. by @crackerjacques in armbian/armbian.github.io#326Add jellyfin-ffmpeg release constraints and ip-wrapper for easepi-r2. by @ifroncy01 in armbian/build#9942arm64: dts: overlay: add radxa display 8hd for rock5t. by @Ken-Vamrs in armbian/linux-rockchip#498board-vendor-logos: update EasePi vendor logo to Linkease. by @ifroncy01 in armbian/armbian.github.io#321drm: rockchip: dw-dp: add AUX recovery for USB-C DP Alt Mode. by @mingzhangqun in armbian/linux-rockchip#494easepi-r2/a2: fix device tree. by @jjm2473 in armbian/linux-rockchip#491feat(boards): Add Seeed Studio reComputer RK3576/RK3588 DevKit support. by @baorepo in armbian/build#9719Give each TQ SoC family its own LINUXCONFIG. by @schmiedelm in armbian/build#9924Introducing Armbian Imager 2.0. by @SuperKali in armbian/imager#143Milk-V Jupiter disable EEPROM node. by @pyavitz in armbian/build#9944nanopi-r76s: fix slow WiFi by enabling SDIO SDR104. by @SuperKali in armbian/build#9929net: r8169: add device tree based LED configuration for r8125. by @mingzhangqun in armbian/linux-rockchip#497orangepi5-ultra: enable Bluetooth on edge kernel via hci_bcm. by @pdapandapda in armbian/build#9697Revert "rockchip64: fix USB gadget NULL pointer crash in eth_get_drvinfo (6.18 + 7.0)". by @EvilOlaf in armbian/build#9926rockchip64: enable back tm16xx driver for current kernel. by @paolosabatino in armbian/build#9941rockchip64: update "describe PCIe Ethernet controllers" patch for stable MAC (many boards). by @rpardini in armbian/build#9933rtl8192eu: re-enable and fix tons of compilation warnings. by @EvilOlaf in armbian/build#9927SpacemiT: Update K1 boot support. by @pyavitz in armbian/build#9930Treat imx93 as slow hardware. by @schmiedelm in armbian/armbian.github.io#323Update EasePi A2/R2 board configs and device trees for improved hardware support. by @ifroncy01 in armbian/build#9907Update VERSION. by @EvilOlaf in armbian/build#9938View the full article

Hello everyone! I'm using this patch set from github to run armbian on x88pro13 (which has RK3528 with Mali 450 MP2). Everything seems to work fine (even wifi), but i'm not able to enable hw acceleration on kodi interface (it shows llvmpipe, so it's sw rendering) Do you have any hints on how to do that? Because at the moment i get 80% cpu just for the kodi gui. Thanks!

I see similar issues with some s905x2 and some s905c3 boxes I have. I've never dug into the issue to see what is happening, but the workaround that works for my cases is that when the boot from emmc fails, I boot from an SD card (I have an SD card ready just for this). Then reboot from emmc, and then it usually will boot. For some reason a successful boot from SD clears up something that then allows the emmc to boot properly. This is what works for me, don't know if your situation is the same or different but you can try.

Thank you for sharing this detailed guide and the partition layout. Getting Debian Trixie up and running on a Teclast tablet is an impressive piece of engineering, and your warning about the 1.8V logic levels will definitely save some hardware from getting fried. Have you made any progress with the aic8800 Wi-Fi driver yet? Also, how is the rest of the hardware behaving with the Radxa kernel?

I got a small progress here with a RKNN (TPU). 1. note from previous posts: rk3566-box-demo.dtb looks like does not support our USB. Use rk3566-box-X88PRO20.dtb that other shared. I tried to use the NPU, looked across the forum. Managed to get our .dtb running with the correct settings - however the Kernel didn't had the RKNN packages + it could not link the 'rocket' to our NPU. While the NPU was recognized after my tweaks. Read that it runs only on kernel v. 6.1 from VENDOR - recompiled the kernel for our board, installed it over the existing installation - .dtb didn't run... When I switched to box-demo from the kernel the NPU and RKNN were recognized instantly. I will not attach the kernel itself for now, as it is ~300 Mb, however you can compile it from Armbian: ./compile.sh kernel BOARD=station-m2 BRANCH=vendor \ KERNEL_BTF=no KERNEL_GIT=full it have to be installed like that: dpkg -i linux-image-vendor-rk35xx_*.deb linux-dtb-vendor-rk35xx_*.deb update-initramfs -c -k 6.1.115-vendor-rk35xx Probably will be a task for today to tweak .dtb for this kernel version and get board running with RKNN. I want to try immich with a local image classification.

Some NVMe's just don't play nice with some units. I have one such NVMe, which is why I rarely try to use it. Also u-boot is not set up to look for an infinite amount of $devnum. Currently it is hard coded to look for nvme 0:1. So I suggest moving the NVMe to the first available. According to the boot log you have more than one PCIE link up [ 1.438] PCIE-0: Link up (Gen2-x2, Bus0) <-- Thats where you want to attach the NVMe [ 1.600] PCIE-2: Link up (Gen2-x2, Bus2) Honestly though. It looks like the NVMe you are trying to use isn't compat.

@sven-ola For my next debugging: To make a separate boot MicroSD, how should I go about? What about this pragmatic approach: Flash the image to a MicroSD card. Then, on that MicroSD card, edit some file in /boot/ which instructs the booting system to use /dev/nvme0n1p1 as root partition. Then flash the SSD with the same image. If wanting to change any boot settings, just mount /dev/sd0a /mnt; cd /mnt/boot , and edit there. Does this work, or would you go through more hoops? Like..: Boot from the MicroSD but in single-user mode (how?) and then mount /dev/nvme0n1p1 / . Then, I re-flash the MicroSD with the image from you: dd if=Armbian.. of=/dev/sd0c bs=1M And then create a new Ext4 filesystem on the now-system partition, for use as boot: mkfs.ext4 /dev/sd0a And then move all the boot files there: mv /boot /boot.orig; mkdir /boot; mount /dev/sd0a /boot; cp -rf /boot.orig/* /boot/ And in addition to this, what files should I change and what commands should I run, for the MicroSD to boot to the SSD i.e. /dev/nvme0n1p1 ? (I'd think some file in /boot , perhaps /etc/fstab ) Thoughts? Aside from trying other power supplies, any thoughts about how to get it going would be much appreciated. One thing I should test is disconnect the SSD from the computer and then try to boot the 7.1 kernel Armbian from MicroSD and see if that works.

@Werner here is the URL: https://paste.armbian.com/yoyimafolu

Let us know @johlnx

I did more of a dance with AI and eventually got it working. This was the real repo: https://github.com/radxa-pkg/aic8800 Got the AI to write a summary of the dance. Phase 1: Resolving the Tooling and Environment Dependencies We started with a clean, lightweight system image missing common Linux development tools. We methodically installed the software compilation and packaging toolchain required to handle vendor drivers: Kernel Headers: Replaced the generic linux-headers-$(uname -r) command with the specific architecture branch package (linux-headers-current-arm64) to give the driver access to the Linux kernel API blueprints. Line Ending Conversions: Installed dos2unix to fix internal cross-platform formatting issues within the raw source files. Packaging Toolchain: Installed devscripts, debhelper, and fakeroot to fulfill the minimum environment demands of the dpkg-buildpackage engine. DKMS Engine Helpers: Installed dh-dkms to handle the missing modern virtual package mapping (dh-sequence-dkms). Phase 2: Resolving Kernel Version API Incompatibilities Because the hardware driver code was originally written for older Linux builds, trying to compile it directly against a modern 6.18 kernel threw standard compilation crashes. We bypassed the rigid Debian patch architecture and fixed the source code directly: Signature Alignment: Modified the function argument signature for .get_tx_power in rwnx_main.c to support the newer 5-parameter layout required by the upstream kernel (struct wiphy *wiphy, struct wireless_dev *wdev, int bss_idx, unsigned int link_id, int *mbm). Variable Alignment: Updated the interior variable name pointer within that function definition from dbm to mbm to match modern wireless power unit metrics used in current Linux network subsystems. Phase 3: Bypassing Packaging & Direct Kernel Compilation Rather than battling failing packaging lint tests or restrictive file verification scripts, we pivoted to an elegant, direct implementation: Dropped straight into the target interface folder (src/USB/driver_fw/drivers/aic8800/). Ran a direct raw build command targeting the standalone USB interface module flag: make CONFIG_AIC8800_USB=m. Manually pushed the generated binaries directly into the system kernel storage directory tree using sudo make install and synchronized the dependency layout using depmod -a. Phase 4: Correcting Firmware Pathing & Module Sequence The driver successfully built and registered with the kernel, but the physical USB bus threw initialization timeouts (bus is not up=0). We fixed the underlying hardware communication pipeline: Firmware Relocation: Traced the location of the compiled firmware folder hidden within the repository's package blueprints, and copied the raw firmware files directly into the absolute hardware search path at /lib/firmware/aic8800D80/. Sequential Probing: Cleared out the broken module states and forced the hardware interface modules to load in their structural order—giving the hardware bus loader helper (aic_load_fw) a 2-second sleep window to awaken the USB links before initializing the operational wireless adapter engine (aic8800_fdrv).

You will have to make a choice which bootloader firmware you use. And then also which boot method and/or bootmanager. Armbian images for Rockchip SBC's are as simple as possible, so only 1 OS, no ad-hoc kernel selection. That means 1 partition for rootfs, no others. Different methods have different advantages and disadvantages. If you want other boot method and/or bootmanager, you will need to add a boot partition yourself, usually FAT formatted and type 0xEF00 (called ESP). The Armbian rockchip64 supports it all, no problem. Especially the Rock5B, which is more or less the blue-print of high-end ARM64 SBC. But all this own custom action also make it not Armbian anymore, but just a generic UEFI ARM64 computer. Where you could also boot from a CD-ROM .iso image and run a traditional Linux distro installer like used on x86 PC's.

I did a apt upgrade that messed up my set up, so I have recently (as in two hours ago) reinstalled armbian on on my dragon q6a. Unlike the previous set up, I have not been able to connect to my wifi. Did a bit of a dance with AI help and still failed, but it looks like the wifi module is directly not found. nmcli device returns ``` DEVICE TYPE STATE CONNECTION enp1s0 ethernet unmanaged -- lo loopback unmanaged -- ``` lsusb returns Bus 001 Device 005: ID a69c:8d80 aicsemi AIC Wlan So I guess that's the wifi hardware and AI claims my system does not have the drivers. AI then sent me on a search for the drivers, but it was to a repo that does not exist, so kind of looking for help here now!

I had once put EDK2 on the SPI flashed and it seems to run fine, I was able to enter the menu. > i used this specific armbian image But Armbian images use U-boot right? So if it boot from the SD-card EDK2 shouldn't launch right. Armbian images are UEFI compatible?