All Activity

I think this is @Igor's signing key aml-s9xx-box:~:# gpg --keyserver keyserver.ubuntu.com --recv-keys 93D6889F9F0E78D5 gpg: key 93D6889F9F0E78D5: 1 duplicate signature removed gpg: key 93D6889F9F0E78D5: public key "Igor Pecovnik <igor@armbian.com>" imported gpg: Total number processed: 1 gpg: imported: 1

also getting this on latest community builds of aml-s9xx-box at https://www.armbian.com/amlogic-s9xx-tv-box/ e.g. https://dl.armbian.com/aml-s9xx-box/Bookworm_current_xfce ``` PRETTY_NAME="Armbian_community 25.8.0-trunk.90 bookworm" ``` wondering if something is broken or missing on the build

this soldering is visionally not very attractive 😛 , but at lest the result is what counts 👍

We don't deal with Android. Better check xda developers forums.

Description As per title How Has This Been Tested? [ ] hasn't Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [ ] I have commented my code, particularly in hard-to-understand areas [ ] My changes generate no new warnings [ ] Any dependent changes have been merged and published in downstream modules View the full article

Description This is a draft/wip for adding https proxy caching. This is especially designed for machines hosting lots of github runners to reduce their bandwidth usage by sacrificing disk space. http caching has been merged already and only needed minor changes. Caching https however is way more complicated. Encryption needs to be broken and rewrapped and then obviously downloads will fail due to cert issues. Classic MITM. So a new certificate authority (CA) needs to be put in place to allow these "fake" certs. Why is this even necessary? Well one major part is downloads from upstream apt repositories. The other major part is downloading 3rd party repositories and artifacts from OCI/ORAS/whatever the correct name for that system is. They do not allow plain http downloads but redirect to https. So in order to cache these encryption must be broken on the fly. On custom runs this works already with the mentioned extension. However when the build framework is used within a GH Actions Runner environment it would need to detect this and enable the extension automatically. No clue how to do that or overall if it is worth diving even deeper into this topic. Therefore this RFC to collect feedback. Also lots of fail-safe checks are missing. Sometimes I had to mix run_host_command_logged and chroot_sdcard since pipes don't work using latter. Documentation summary for feature / change Yes, this would need documentation. How Has This Been Tested? [x] custom builds with ENABLE_EXTENSION [ ] Test B Checklist: [no idea] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] My changes generate no new warnings [ ] Any dependent changes have been merged and published in downstream modules View the full article

Description as per title How Has This Been Tested? [x] build [ ] nope, no hw on hand right now Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [ ] I have commented my code, particularly in hard-to-understand areas [x] My changes generate no new warnings [ ] Any dependent changes have been merged and published in downstream modules View the full article

Okay I will try that patch and the build switch that you mentioned. About the RAM: It is SKHynix 966.3MB(989480kB) https://linux-sunxi.org/File:A088_Board_Front_3.jpg I then tried an archived image for the cubieboard and so far no errors and runs quite stable: https://armbian.atomonetworks.com/archive/cubieboard/archive/ Is this a kernel issue then? Maybe I should compile with an older kernel. The stable one is 6.1.63 Thanks a lot for your continued support, Shivam

Hi, Tried this Opi5-Plus with kernel-6.15.2-edge-rockchip64 but mpv doesn't seems to have vpu hardware acceleration Does this repository works on RK3588 - Orange Pi 5 Plus Debian Trixie with mainline kernel 6.15.2-edge-rockchip64?

I think @amazingfate made a custom Chromium package at some point that included optimizations for rk35xx but no clue about its status. Might be even included in mesa-vpu extension? Firefox probably won't work since that would need adjustments as well and I think messing with browsers is a pita already. topics merged as duplicates.

I don't know. I don't have this board nor is the Armbian team aware of its state since it is community supported. This looks like some upstream (Debian to say) uboot package which certainly does not support this board. IIRC u-boot versions for rk3588, depending on family are something like 2017.09 or 2024.01 (both vendor cases) and 2025.01 or .04 for current/edge.

I installed 25.04 with kernel 6.1 on the orange pi 5 plus, then I installed vanilla gnome. Then I installed the GPU drivers & GStreamer, then I enabled the GPU in the Armbian config. This works great, hardware acceleration works on GNOME. However, chrome/firefox are both super laggy and do not have hardware acceleration. What are the proper steps to get hardware acceleration (browser & video player) on this platform? Yes I'm aware this is considered a developmental build, I'm just asking for general advice as I'm fairly new to the OPI5+ and don't know the general steps of doing so. Thanks!

Right, I was thinking what if you try forcing a different mirror via DOWNLOAD_MIRROR: https://docs.armbian.com/Developer-Guide_Build-Switches/#hidden-options-for-advanced-users-default-values-are-marked-bold I cant guarantee this will work though. While A2 cards offer better performance, they are not as well supported. That is still the case to the best of my knowledge. Right one of the caveats I did find with getting HDMI working was the likelihood of increased taints so I think this patch is also required: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/drivers/gpu/drm/sun4i/sun4i_hdmi_enc.c?h=v6.14.11&id=ae048fc4f96d716a2ef326bd6e894694057c078c It is a possibility that either the RAM could be faulty or just unstable and need fine tuning. The problem is that really old system images tend to be less picky about RAM issues and will still try to boot. Normally I would say try installing memtester but that seems to detect failures in the most recent kernel yet my 6.6 build found no memory issues. Besides the originally Android image, have you tested with one of the archived images? More importantly what exactly are the RAM chips used on the board? Can you confirm the actual RAM capacity? From my experience, my second pcduino2 board with the Skhynix RAM had these sorts of problems. Sorry I cant be of more help Ryzer

So I kept trying to build inside VirtualBox and on my 3rd try I was able to get an image built (along with the patch that you shared). I tried that on the Class 10 card and got some "Bad Page Map in process xxx" errors, although it did boot and I was able to set it up. On running any command I keep getting these errors and most commands fail. Then I got an A2 card and tried the same image on that, with the same results. Towards the end it tries to turn HDMI on but fails because of said errors. I am attaching the terminal output from the A2 card run here: https://pastebin.com/7H70dUNN What do you think the errors are related to? The RAM? I want to mention that since I have the original Android image on the NAND, I can confirm that it runs okay. Thanks,' Shivam

Same problem here, but enabling the audio codec on armbian-config was not sucessfull on this version: v25.5.1 for Orange Pi Zero running Armbian Linux 6.12.30-current-sunxi Any solution?

@Werner You suggested using armbian-install. You meant to update boot loader in current installation? If I'll do this will it break currently working booting from nvme?

In that case I would strongly recommend purchasing an A1 rated SD card before trying anything else. If you are getting MMC related errors these are definitely related to the SD card rather than RAM. Once you have done this then we can look into DRAM config if need be. Trying to enable the display post build is kinda tricky. If it was just a matter of adding the HDMI display nodes, then yes that is normally possible however the display driver needs a slight patch in order to work. Now if you had kernel headers installed you could try to build the driver however another current problem means that custom loaded modules seem to not load/unload properly. Seeming as you are using the pre-built Cubieboard images, it already has the display nodes in place but not the recent patches to sun4i-drm. Looking again at the build log you provided earlier it looks like something is blocking the download of bookworm distro which is why it is failing. what happens if you trying ping apt.armbian.com? Given that this issue happens with bookworm, I wonder if it will also happen if trying build with Ubuntu Noble instead. I am currently working on formally including the display patches on my latest development branch so please bear with me. Best of luck Ryzer

i need help too.... please

Hello, @sicxnull, Thanks for all your work on the H96X box. I'm trying to install Armbian on my x96q V5.1, your image (Armbian-unofficial_24.11.0-trunk_X96q-v5-1_bookworm_current_6.6.44_mate_desktop.img.xz) works great apart from I'd really like a server instead of desktop. I'm just wanting to use it for a VPN (Tailscale) exit node. Would it be hard to build a server version? I have no Linux experience.

I don't know what else to do... I usually research and learn on the go Search the 88xxau.c in your downloaded linux source. Try to find any link to documentation or github. Search 8812au in the respberry pi forum

Update maintainers and board status synced status from the database rename to .csc where we don't have anyone If you want to become a board maintainer, adjust data here. Ref: Board Maintainers Procedures and Guidelines Contribute View the full article

Hi! On the Orange Pi 5 plus with the 6.1 image (KDE), with a 4k monitor @ 60FPS and 200% scaling, when opening discord on Chrome it gets incredibly laggy. I'm just curious, is this a limitation with the GPU itself, or is this related to something in the software stack? Thanks!

The actions in the dts for setting the GPIO pin are is based on the boot log line. marlin chip en dummy pull up -- need manually set GPIO You can check the whole discussion here and the reasoning https://chatgpt.com/share/68466d15-e5bc-8013-b200-17ae7f7c89fc However, After your last answer, Chatgpt suggests this is not really the issue, but a delayed initialisation would help. When the bt fails to initialise, try the following: sudo hciattach /dev/ttyS0 any 115200 flow sudo hciconfig hci0 up If this works, you can put this in a service to run at every boot. You can try yourself to continue the above conversation with chatgpt and see if you find the final solution. Then you can post it here.

I have decompiled the dtb file but am unable to incorporate the patch. How do i incorporate the patch or change it? In terms for a newbie please! I should say it is for debian 12 Cinnamon 6.12.

NO_PUBKEY 93D6889F9F0E78D5 while using apt (e.g. apt update) symptom: > apt update ... Err:8 https://github.armbian.com/configng stable InRelease The following signatures couldn't be verified because the public key is not available: NO_PUBKEY 93D6889F9F0E78D5 ... W: Failed to fetch https://github.armbian.com/configng/dists/stable/InRelease The following signatures couldn't be verified because the public key is not available: NO_PUBKEY 93D6889F9F0E78D5 observed in Armbian image for Orange pi zero 3 Armbian_community_25.8.0-trunk.90_Orangepizero3_bookworm_current_6.12.30_minimal.img build date May 28, 2025 fix: - run this as root > su - ^ login as root > wget -O - https://apt.armbian.com/armbian.key | gpg --dearmor -o /usr/share/keyrings/armbian.gpg you should find a file /usr/share/keyrings/armbian.gpg about 2 KB in size repeat apt update etc should have resolved the error