Active threads

Did you try loading the "pinctrl-rk805" module? This was once directly compiled into the kernel and later externalized into a module. And IIRC that happened when switching to 6.12. If loading that module helps, then adding it to the MODULES line of the board's config like this should fix it.

moved

Yes, shorting the pins I was switch the device into MaskRom mode and install Android. Now I trying to identify and use the debug console, not much success so far

Will check it when i back to PC, will be away from pc for around a week

I do get multiple warnings while compiling the DTS file, about 50-100 warnings similar to this: ttt.dts:1752.3-32: Warning (gpios_property): /usb0-vbus:gpio: cell 0 is not a phandle reference ttt.dts:1764.3-32: Warning (gpios_property): /usb1-vbus:gpio: cell 0 is not a phandle reference ttt.dts:1776.3-32: Warning (gpios_property): /usb2-vbus:gpio: cell 0 is not a phandle reference ttt.dts:1823.4-34: Warning (gpios_property): /leds/led-3:gpios: cell 0 is not a phandle reference ttt.dts:1828.4-34: Warning (gpios_property): /leds/led-4:gpios: cell 0 is not a phandle reference The phandles are in this kind of format on the dts file: phandle = <0x20>; phandle = <0xc2>; phandle = <0x1c>; phandle = <0xc3>; phandle = <0xc4>; phandle = <0x16>; phandle = <0xc5>; example: Note: the usb0 vbus was originally disabled but I enabled it as the pcduino3 nano has a single regulator for all ports and it is attached to usb0 apparently. In the closing section of __symbols__, I have redirected all 3 regulators to usb0. I believe that the fact it is disabled is because they don't recommend powering the board through the OTG port as it can cause voltage drops when connecting devices as host. My setup will be a gadget with a wifi dongle in a USB A port, so voltage drops should not be a concern, this is just my assumption... usb0-vbus { compatible = "regulator-fixed"; regulator-name = "usb0-vbus"; regulator-min-microvolt = <0x4c4b40>; regulator-max-microvolt = <0x4c4b40>; enable-active-high; gpio = <0x17 0x01 0x09 0x00>; status = "okay"; phandle = <0xc2>;

Ok, in case anyone's got any ideas about how to fix this, I'm happy to help if I can, but kernel repairs are a bit out of my league... It seems to reboot for different reasons - when it doesn't like something connected to the USB port, when doing a lot of disk I/O, etc. Here's a couple of the error logs from the terminal: INFO: task sgdisk:1567 blocked for more than 120 seconds. Tainted: G O 6.18.10-edge-sunxi64 #1 "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. task:sgdisk state:D stack:0 pid:1567 tgid:1567 ppid:1566 task_flags:0x400000 flags:0x00000018 Call trace: __switch_to+0xcc/0x180 (T) __schedule+0x348/0xaf8 schedule+0x38/0x110 io_schedule+0x40/0x60 folio_wait_bit_common+0x15c/0x390 folio_wait_bit+0x1c/0x30 folio_wait_writeback+0x4c/0xb8 __filemap_fdatawait_range+0x70/0xb0 filemap_fdatawait_keep_errors+0x24/0x58 sync_bdevs+0xb4/0x1a8 ksys_sync+0x64/0x98 __arm64_sys_sync+0x14/0x28 invoke_syscall.constprop.0+0x54/0xe8 do_el0_svc+0x44/0xc8 el0_svc+0x38/0x140 el0t_64_sync_handler+0x98/0xe0 el0t_64_sync+0x170/0x178 Kernel panic - not syncing: hung_task: blocked tasks CPU: 6 UID: 0 PID: 61 Comm: khungtaskd Tainted: G O 6.18.10-edge-sunxi64 #1 PREEMPT Tainted: [O]=OOT_MODULE Hardware name: Radxa Cubie A5E (DT) Call trace: show_stack+0x1c/0x30 (C) dump_stack_lvl+0x30/0x80 dump_stack+0x14/0x20 vpanic+0x2d4/0x308 panic+0x50/0x58 watchdog+0x278/0x718 kthread+0x134/0x1f8 ret_from_fork+0x10/0x20 SMP: stopping secondary CPUs Kernel Offset: disabled CPU features: 0x080000,00008000,48006281,0400701b Memory Limit: none ---[ end Kernel panic - not syncing: hung_task: blocked tasks ]--- INFO: task systemd-udevd:285 blocked for more than 120 seconds. Tainted: G O 6.18.10-edge-sunxi64 #1 "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. task:systemd-udevd state:D stack:0 pid:285 tgid:285 ppid:1 task_flags:0x400100 flags:0x00000809 Call trace: __switch_to+0xcc/0x180 (T) __schedule+0x348/0xaf8 schedule+0x38/0x110 io_schedule+0x40/0x60 folio_wait_bit_common+0x15c/0x390 folio_wait_bit+0x1c/0x30 folio_wait_writeback+0x4c/0xb8 __filemap_fdatawait_range+0x70/0xb0 filemap_write_and_wait_range+0x90/0xc0 sync_blockdev+0x24/0x40 bdev_disk_changed+0x64/0x5c8 blkdev_get_whole+0xa4/0xf8 bdev_open+0x278/0x3b0 bdev_file_open_by_dev+0xdc/0x148 disk_scan_partitions+0x6c/0x150 blkdev_ioctl+0x664/0xfc0 __arm64_sys_ioctl+0x470/0xaa0 invoke_syscall.constprop.0+0x54/0xe8 do_el0_svc+0xa4/0xc8 el0_svc+0x38/0x140 el0t_64_sync_handler+0x98/0xe0 el0t_64_sync+0x170/0x178 INFO: task sgdisk:1465 blocked for more than 120 seconds. Tainted: G O 6.18.10-edge-sunxi64 #1 "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. task:sgdisk state:D stack:0 pid:1465 tgid:1465 ppid:1464 task_flags:0x400000 flags:0x00000008 Call trace: __switch_to+0xcc/0x180 (T) __schedule+0x348/0xaf8 schedule+0x38/0x110 io_schedule+0x40/0x60 folio_wait_bit_common+0x15c/0x390 folio_wait_bit+0x1c/0x30 folio_wait_writeback+0x4c/0xb8 __filemap_fdatawait_range+0x70/0xb0 filemap_write_and_wait_range+0x90/0xc0 bdev_release+0x19c/0x1b0 blkdev_release+0x18/0x30 __fput+0xd0/0x2e8 fput_close_sync+0x44/0x108 __arm64_sys_close+0x3c/0x88 invoke_syscall.constprop.0+0x54/0xe8 do_el0_svc+0x44/0xc8 el0_svc+0x38/0x140 el0t_64_sync_handler+0x98/0xe0 el0t_64_sync+0x170/0x178 Kernel panic - not syncing: hung_task: blocked tasks CPU: 1 UID: 0 PID: 61 Comm: khungtaskd Tainted: G O 6.18.10-edge-sunxi64 #1 PREEMPT Tainted: [O]=OOT_MODULE Hardware name: Radxa Cubie A5E (DT) Call trace: show_stack+0x1c/0x30 (C) dump_stack_lvl+0x30/0x80 dump_stack+0x14/0x20 vpanic+0x2d4/0x308 panic+0x50/0x58 watchdog+0x278/0x718 kthread+0x134/0x1f8 ret_from_fork+0x10/0x20 SMP: stopping secondary CPUs Kernel Offset: disabled CPU features: 0x080000,00008000,48006281,0400701b Memory Limit: none ---[ end Kernel panic - not syncing: hung_task: blocked tasks ]---

@billymore The "Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(0,0)" error indicates that the Linux kernel could not locate or mount the root file system during startup. . This is commonly caused by a missing or corrupted initial RAM filesystem (initramfs), an incorrect root= boot parameter write 0x4ff00000 rootfs.cpio.lzma.uboot

@Jain Ziad I'm currently working on hardware acceleration, though I can't say yet if it will outperform the official Radxa builds since it uses the same drivers. I'll keep the community updated on the results.

Hello Armbian-Community! TL;DR I've successfully enabled the Rockchip RK3568 NPU (0.8 TOPS INT8) on the ODROID-M1 with 8GB RAM running Armbian 6.18.9-current-rockchip64. The only thing needed to make this work for everyone is a 1-line kernel patch in Armbian build that I've already submitted: https://github.com/armbian/build/pull/9403 Once this patch is merged into Armbian, the DKMS modules will work out-of-the-box on all RK3568 boards. The Problem 1. Hardkernel and Rockchip provide NPU patches for quite old kernel versions: rockchip <= 6.6 hardkernel <= 5.10 2. IOMMU page table allocation bug — On systems with >4GB RAM, the kernel's IOMMU allocates page tables above 4GB, but the NPU can only access the first 4GB of physical memory, causing DMA mapping failures and inference timeouts The Solution 1. Kernel Patch (1 line!) - .gfp_flags = 0, + .gfp_flags = GFP_DMA32, This forces IOMMU page tables to be allocated below 4GB, fixing NPU operation on 8GB boards. 2. DKMS Driver Package I've created a complete DKMS package that: - Includes DT overlays for NPU, IOMMU, power-domain, and clock configuration - Supports dynamic frequency scaling (100 MHz - 1000 MHz) - Provides `/dev/dri/renderD129` (DRM/GEM interface) for RKNN Runtime - Would compile against Armbian's stock kernel headers with the above patch merged The Questions 1. What would be the best way to provide those modules for installation in the armbian system? 2. Has someone a 4 GB version of the M1 and could test the modules?

Hello, 4GB version. - Detecting meteors in GMN network (https://globalmeteornetwork.org/) - monitor other x64 machine and reset by power cycle (via gpio). rock64 is rock stable

The sysfs GPIO interface does not allow you to create clock outputs as-is. You'd do better seeing if that pin can be exposed through the sysfs PWM interface. I know the NanoPi M4V2's fan controller works this way (RK3399), and you can set the duty cycle. But you'd have to check if PWM is supported on that pin, and can be used in the way you're hoping. Failing that, it is possible to bit-bang the GPIO lines yourself if you're willing to write some C code. You can use Kernel Timers for fairly accurate timing, as long as you set High Scheduler Priority and Real Time Scheduler Class. Note that kernel timers only have a user specified resolution of 1ns, so you might not hit that frequency exactly.

Probably a new product. Images are not compatible unless specifically marked as such.

Regarding the vendor kernel I also managed to get the GPU going with some AI help, but the HDMI hotplug issue persists. Will probably revisit it in the future. (the dtb should go in packages/blobs/h96-m9_original.dtb and the defconfig in patch/u-boot/legacy/u-boot-radxa-rk35xx/deconfig) h96-m9.csc h96-m9_original.dts h96-m9-rk3576_defconfig h96-m9_original.dtb

the tft is working well on stock orangepi image but when I try the prebuilt armbian images or build it myself it look like this my dts file: 35Lcd.dts

I've gotten it working. I'll detail the steps soon, but it's not difficult. I just want to confirm that running it this way I can boot from sd/usb/nvme/eMMC/Network, and see if changing the boot order works.

This is not working for me either. I had audio initially working with the "Radxa OS" (Ubuntu 24.04), but after a few reboots, seems to lose audio and reverts to "Dummy Output" audio placeholder for the Output Device. No luck at all with sound on Armbian even after following the instructions laid out here. Would be nice to get audio working reliably for the Q6A on Armbian as its my favorite SBC to date, but no sound is a show stopper.

It appears there would have been no problem with the Logitech Unifying Dongle that uses Bluetooth. I ended up buying a Logitech Keyboard/Mouse combination that is practically identical except uses WiFi protocol rather than Bluetooth as per the suggestion of the vendor of the HamClock Quadra4K. PeeSHaw.... Prices have come down I guess... I found it at the local Wally World for $19.95 on sale... Crazy! I was asking because I had an older Logitech Bluetooth mouse and dongle and was thinking I would just use them and get the keyboard... Thank you for the help!

It works, I'm soooo happy with this. Big thank you!!!

Not excited AT ALL about Allwinner, but there are few (actually none that I've found?) choices for low-priced, "higher performance" SBCs with WiFi 6, BT 5.4, USB3, GB Ethernet, PoE and a small size and factory case. I'll take anything rockchip in a heartbeat I much prefer their chips. My go-to has been the RockPI-S, but it's low performance CPU, WiFi 4, BT 4, 100M Ethernet, USB2, and getting pretty old. AFAIK, there's nothing out there that checks all the boxes, sadly.

There was a commit 4 days ago on that repo metioned above: https://github.com/joilg/x88pro/commit/a6630814a98b39ec39dc22108c01e35f5d456395 Anyone noticed? Anyone tried it? Any change?

OK, so it actually was cold in FL for a couple weeks and I did a power hack. I ported Java UIO to Java UIO 2. Java UIO (HawtJNI) This was the "Old School" high-performance approach. It relied on a custom C wrapper and JNI to bridge the gap. How it works: You have to write a C helper file, compile it into a shared library (.so) for on ARM architecture (armv7l, aarch64), and bundle it in your JAR. The Overhead: Every call from Java to C has a "JNI transition cost." To minimize this, you used moveJavaToNative to bulk-transfer buffers, which added complexity to the Java code. Java UIO 2 (FFM / Project Panama) The "Modern" approach. It treats native memory and functions as first-class citizens in Java. I now cross compile on x86_64 for arm32 and arm64. How it works: Java uses Linker and SymbolLookup to find functions in u8g2 or libc directly. No custom C wrapper is strictly required unless you want to simplify complex macros. The Overhead: Transitions are heavily optimized by the JVM (often inlined). By using MemorySegment, you can point Java directly at the UIO hardware registers or the U8g2 tile buffer. The Armbian Advantage: It is "Write Once, Run Anywhere" for native code. As long as the system library (like libSDL2.so) exists on the Armbian filesystem, the same JAR will work on an Orange Pi 5, a Pine64, or even a Raspberry Pi without re-compiling C code.

Actually not as I am human - I need time and focus to be able to understand what the code does. Same as everyone else. BASH is list of commands and if you are average tech person you can interpret those command, one after another to understand what is going on. AI can help. You can, but you can not expect that I will answer them. Perhaps someone else will, perhaps nobody. I show you how you can understand it without destroying my brain - I need to look into the code to answer questions ... Its Friday evening here and I already answered at maximum effort.

Thank you. Basic functionality, few non critical bugs = release ready. Moving to main download location. Imager will have them after they are moved to primary download location. Test images are not there. We don't provide that. We will only provide images with mainline derived kernel which is hard to compare - its a very different SW stack - with Radxa vendor kernel. Basic (!) functions for headless usage are already working. The rest in next 6-12-> month.

This would be amazing. I don't suppose anything came of this? I'm not sure if armbian-install is still actively maintained, but I'm not having any luck using it with several boards I have now.