Active threads

Pami, I have not personally found any LLM capable of generating working DTS files no matter what. It seems to be too niche for the LLM to understand. This DTS could potentially be a good starting ground for the pi, replacing "bcm2835" which refers to the pi zero 1, with "bcm2710" which refers to the pi zero 2. The github repository I linked is a good guide to getting the red displays working for a voron printer using panel-mipi driver, and I referenced it to create the DTS I use for the rock pi 4b. Also, try lowering the SPI speed to 20, 16, or 10 MHz, then raise it when you verify it works. I was frustrating myself for weeks thinking my DTS was broken, when it was fine the entire time, just my SPI speed was too high. robertoj: I appreciate your help, but I believe the issue I have with the red display (I have the same one as pami sent) is a combination of SPI commands being sent at the wrong interval (at the end of the clock cycle instead of in line), as well as insufficient SPI speed - where the display freezes above 24 MHz, and shows extreme visual tearing below it. At RGB666, its too much data for a measly 20MHz SPI rate, and the board does not support any other RGB modes in SPI. Since the TE pin is not exposed on the board, I cannot sync the SPI writes to the display. At least, that is my theory. I've given up on using an SPI display, and instead transitioned to an RGB Parallel display, which uses one gpio pin per color bit. Note that, at least in the display datasheet I received, the display actually does have the pins needed to drive it over DPI RGB Parallel. You can drive the board in 16, 18, or 24 bit mode over DPI RGB Parallel. These pins are not exposed on the board, but you could use an FPC breakout board to interface with these pins, and drive the display in DPI RGB parallel instead (accounting for the backlight voltage needed). The advantage of using RGB Parallel is it natively uses the GPU (RGB Parallel is the same technology as VGA) and has very fast refresh rate. For anyone here in future, any DPI hats that exist, like the adafruit kippah or othermod dpi topper, do not have the correct pinout to interface with this specific ili9486/8, but if you were to rearrange the pins on either the HATS or with an intermediary FPC PCB, then they would work plug and play.

Hello, i tried "Armbian_community_26.2.0-trunk.606_Radxa-cubie-a5e_trixie_edge_6.19.8_minimal.img" and wlan didn't work, i think the device tree is broken Radxa-cubie-a5e kernel 6.19.x as a quick fix, I copy the kernel 6.18.x "sun55i-a527-cubie-a5e.dtb" file to the /boot/dtb/allwinner directory to get wlan working.

@sven-ola there are 3 ways to boot the R2S: 1. Press the boot button before applying power. White USB port connected to a PC. Apply power. It goes into fastboot mode like an android. Theoretically this should let you flash the on-board eMMC. I can successfully put it in fastboot mode but the flashing has not worked. 2. Boot from eMMC 3. Boot from USB. The eMMC has two boot partitions that are fixed size and one large main partition. My project uses the u-boot environment but also does a few minor things. It prioritizes USB if it finds a bootable medium there. It also safely flashes the boot blocks on the eMMC and then the main blocks. It adds some safety checks before flashing. Nothing else.

hi, I installed in usb the last version and boot. something wrong happen here, only left these two upgrades, I tried again with backup-img and falied to boot again

[ 5.092655] CPU1: failed to come online [ 5.092666] CPU1: failed in unknown state : 0x0 [ 10.217311] CPU2: failed to come online [ 10.217324] CPU2: failed in unknown state : 0x0 [ 15.341948] CPU3: failed to come online [ 15.341961] CPU3: failed in unknown state : 0x0 Yes we're aware of this, since it affects all H5 and H6 devices. We have a few ideas about what could be the cause like added logging (yes, sounds odd but..oh well) or certain ATF version. https://github.com/armbian/build/issues/9555

moved to off-topic since not directly related to armbian.

Ok after trying a speeker switcher, switching from cinnimon to the kde version, finnally I tried flashing with the offical armbian flasher, and I tried a bunch of old online solutions that looked similar. Still my headphone doesn't work. I think it is due to armbian thinking the built in sound card is the headphones. But on a positive note the weird unmounting of armbian when my chromebook sleeps or when the lid is closed does not occur on armbian kde. So far kde is more stable than cinnimon.

@José Manuel Márquez Luque try these images https://github.com/NickAlilovic/build/releases/tag/20250306

@Bob-the-great try these images https://github.com/NickAlilovic/build/releases/tag/20250306

Thank a lot !

At a glance, unless you see your board at https://armbian.com/download, chances are that no one else has gotten that board working. Might be better to post under "Other Families" for unsupported board stuff. @Werner may move later ... Might be crashing. Might be redirecting kernel to a different UART. May be worth checking earlycon and console settings in kernel args? You can try to pass break=premount or similar arguments to pause boot and allow you to poke around and see what's going on?

HI, It looks like you have an RK3518 chip from your photo, and not the 3528? Or was that a typo? In any case I also bought a box from AliExpress. It looks similar to mine with the 3518 chip. If there is a way to extract the ROM Image from my box I can maybe upload it for you somewhere?

You may want to check out this thread. There's someone else doing the same with an Orange Pi 5. Might be worth comparing notes with @KhanhDTP

seems like openwrt already figured this and the fix is actually upstream: https://github.com/torvalds/linux/commit/c3fe7071e196e25789ecf90dbc9e8491a98884d7 (also see https://github.com/immortalwrt/immortalwrt/commit/7ef19bb9cd59f5018fc96b01a5465ed84038cfe8) This may tell that one or more Armbian patches make things worse , no clue though which one if any.

Hi @Werner, I bought an UART USB device and followed your video here https://www.youtube.com/watch?v=UpVMO7gbnYM to read boot logs. I'm not a huge expert but looks like there is a problem finding /boot when two nvme are connected. Please find attached the following logs: uart-bootcorrectnvme.log -> one nvme plugged (the one with os installed in slot 1) - works! uart-bootdualnvme.log -> two nvme plugged (the one with os installed in slot 1) - fails! uart-bootmicrosd.log -> two nvme plugged + microsd plugged - works! Tests have been executed with official 30W Radxa Power adapter ( same result with POE+ 25W hat) Please, let me know if there is anything else I can provide. Thanks uart-bootcorrectnvme.log uart-bootdualnvme.log uart-bootmicrosd.log

moved

@Werner I have tried your command but no success. Here is what a clean boot of Armbian shows. I also ran some commands with the outputs below.

https://docs.armbian.com/Process_Contribute/#adding-a-new-board

I ran into what appears to be the same issue (see my log). I have added the suggested qemu-user-static package, which forced removal of another qemu- family package I had previously installed to work around problems with binfmt. Not only did it compile correctly, it also ran so much faster, as you can see in the number of log lines for this build! Thank you for the assistance and getting me on the right track with the right package selection! Best Regards, Lobosito

@tin harden A7A mainline is still a WIP.

@Thur_Wander For debugging purposes, UART is highly recommended. Use balena Etcher to write the compiled image to your SD card.

Hi, I am here in 2026 with the Amrbian 26.2.1, I've tried the proposed fix from @c0rnelius but it's not working on my OS version. If anyone has any updated solution that'd be amazing. Thank you

I will try downgrading u-boot next once I find the older package somewhere. Kernel does not seem to be the cause given that it has been working fine with both my own build 6.6.122 and 6.6.128 before and rolling back not helping. New device is something I had on my mind for a while. But given the RAM and SSD prices these days, I do not really want to overspend. Not to mention that the SBC landscape has not been that great lately when it comes to compact NAS offerings. HC4 is five years old at this point. Many even recommend N100 but this is a whole different size category.

hello ,my box is mortal t1 2g8rom and motherboard show mortal t2 v2.0 it's broken and i wanna make linux boot what's the most version to be able for my box

Good morning I have big problem I have tv box Transpeed-H616_ I update with orange pi But i can't back to old firmware