Igor

Main repository? Recompile rockchip-next from development branch and push up?

Yes. The simplest way is to add eMMC support to Neo image and both boards are covered.

You want to have ethernet in u-boot or in a kernel? Or both?

Our system for creating patches stops right before the compilation starts, but at that moment defconfig was already applied. You need to: 1. Edit this config and proceed that patch is created 2. Copy patch from output/patch/u-boot-sunxi*.patch to userpatches/u-boot/u-boot-sunxi 3. Recompile with this patch included Then it should be applied. Perhaps best to add eMMc support (CONFIG_MMC_SUNXI_SLOT_EXTRA=2) to Nanopi Neo u-boot config and use that image?

Hmm. We had some urgent updates from development branch which is already at 5.42 Splitting master and development branch, in general, brought some confusion and this is a direct consequence. Not that critical but it shell be fixed soon.

Rainer, thank you. Fixed: https://github.com/armbian/build/commit/28feb0269a8f084e23f0760ab5a19ab6b12b174f Will show up in next nightly update.

Worth to add here that we host kernel sources as well. Those are present in a stable repository, while beta we don't since it takes too much space.

You need to edit/add configs in cache/sources/u-boot-source/u-boot-branch/configs/ More precisely, add whatever you need to /configs/nanopi_neo_defconfig ... in theory, you could have there the whole .config, but its not needed. You will need to do this patch in two runs. First, run "compile.sh" with CREATE_PATCHES option, edit this nanopi_neo_defconfig ... save, proceed and a patch to the Neo config will be created. Then run again and eventually put this patch to your userpatches/u-boot/sunxi-u-boot-sunxi ...

Theoretically you could put things together on 4.14.y but not recommended. 4.17.y ... in a few months to get theoretical grounds, then few more to get things operational.

Go for the legacy kernel. There is no MALI in a modern kernel .. yet.

I would need to download and burn the image to be sure I am telling you correct info, but you can do that on your own as well. Waste of time. Dietpi is a BASH creation on top of Ayufan image. Download Armbian and see what's in and what's missing. If something is missing and only need to be enabled ... no problem. Submit a patch (to this file) and the very next day this feature is enabled in a (beta/nightly) repository.

You need yet another switch LIB_TAG="development" / untested under Docker to switch into the development branch. I am not sure master is updated for Rock64.

I am not sure if it exists, but it would be cool. There are some older boards which are ready that you can plug in LTE mPCI modem - Clearfog and Hummingboard ... none is small. There is also an Orange 4G IOT but with questionable software support. @Stanislav Sinyagin https://txlab.wordpress.com/ is playing with 3G/4G connectivity and had a similar wish some time ago. Perhaps he made some progress?

That is good since I (hopefully) fixed the problem. Now, general userspace problems. We already spotted that network is not up, more packages might be missing, ...

This means boot process is somehow and sometimes broken due to wrong u-boot settings and/or hardware issue ... and fails to network boot and since there is no network drivers nor any configurations it ends like this. We don't support this board and this Armbian build is not ours.

This used to work on older images with old kernel and nobody adjust/fix this for a modern kernel. I think this is the problem.

What about this image: https://dl.armbian.com/orangepipcplus/Debian_stretch_next_desktop.7z BTW. My test setup is an old 19" 1280x1024 monitor with this HDMI 2 VGA converter. It works.

Check this script: https://github.com/armbian/build/blob/development/packages/bsp/common/etc/init.d/armhwinfo ... and apply your fix after determining the diff between XU4 and HC1? I am not sure we want this but we can add this as a toggle function to armbian-config O.K.. Tnx. Tomorrow changes - upstream + all images are rebuild by using upgraded deboostrap ... which is a must to build Ubuntu Bionic.

No problem, at least we find the source of troubles. My hostile reaction was due to the fact you were technically not running Armbian at all, claiming it doesn't run and asking for help even we actually dont provide any support for this kind of builds. Wrote on mobile

Aha. This is a problem of a system which is on eMMC. This device is not officially supported yet so you won't find a lot of specifics information yet. I tried eMMC install (with our general install tool called nand-sata-install or from armbian-config) and it is broken, but boot from SPI works ... There are rough edges and that's why we keep this image labeled WIP.

This image is officially still WIP/testing phase. Add parameter EXPERT=yes

https://www.armbian.com/rock64/ Both are tested.

Not possible. You are not listening and reading what is written. Everything you claim, it is not working, works. If you are a noob why are you installing a kernel that is not fully compatible? Where did we put such instructions? If you are using armbian, you should only use kernels we prepare ... of course you can use other (compatible) kernels, but in that case, you have to understand what you are doing. This forum is support for armbian builds, anything else is ignored. We don't support any 3rd party kernels since supporting own creations is already expensive enough. All this is written in the documentation, which is written with a reason that we don't waste time repeating things over and over again. Again. This is your first start: https://docs.armbian.com/

Either way, one file is o.k. for me.

Yes, the same. And only with ethernet. Wireless works but I have troubles with host preparation. On a clean host, deboostrap of any system fails completely: I: Extracting libmount1... I: Extracting libsmartcols1... I: Extracting libuuid1... I: Extracting mount... I: Extracting util-linux... I: Extracting liblzma5... I: Extracting zlib1g... [ error ] ERROR in function create_chroot [ chroot-buildpackages.sh:43 ] [ error ] Create chroot first stage failed [ o.k. ] Process terminated That bothers me atm. Still investigating why is this happening. On my main build machine, it works.