going

The changes that are used in this test. https://github.com/armbian/build/pull/3737

Just test the operation of the equipment for a few days. And tell me if everything is okay. Fixed it

@chouchouUse this link and install only the kernel.

Yes. Everything is right. That's what needs to be done.

@mdrmdr https://github.com/The-going/PKG_test/blob/master/nanopineo/README.txt

Also please provide information about your OS.

Please publish information about your current working operating system: cat /boot/armbianEnv.txt lsb_release -rc find /sys/ -name '*spi*' df -h Are the overlay files that you use files that are part of the dtb system package? And simple curiosity. Which device do you connect via SPI? I will make all the packages and provide a download link. There will be instructions.

A little wrong. This is the official way to load this kernel module. You can write complete nonsense, as I did for the test. diff --git a/drivers/spi/spidev.c b/drivers/spi/spidev.c index 852dcbb2d..2c53e2b5a 100644 --- a/drivers/spi/spidev.c +++ b/drivers/spi/spidev.c @@ -682,6 +682,7 @@ static const struct spi_device_id spidev_spi_ids[] = { { .name = "m53cpld" }, { .name = "spi-petra" }, { .name = "spi-authenta" }, + { .name = "elephant" }, {}, }; MODULE_DEVICE_TABLE(spi, spidev_spi_ids); @@ -696,6 +697,7 @@ static const struct of_device_id spidev_dt_ids[] = { { .compatible = "menlo,m53cpld" }, { .compatible = "cisco,spi-petra" }, { .compatible = "micron,spi-authenta" }, + { .compatible = "hello,elephant" }, {}, }; MODULE_DEVICE_TABLE(of, spidev_dt_ids); And call this module to load and initialize by replacing the compatible = "spidev"; string with compatible = "hello,elephant"; in the dts file. I have discovered this and will fix it in the 5.17 kernel soon. And I need someone to test the fixes on real hardware. Anyone interested?

https://lore.kernel.org/all/20210921192149.50740-1-broonie@kernel.org/ In other words, the core has become smarter. If it is correctly spelled out in the device tree, then the kernel does the mapping and the mapped driver is used. For example in orangepi-pc2 &spi0 { status = "okay"; spi-flash@0 { #address-cells = <1>; #size-cells = <0>; compatible = "jedec,spi-nor"; reg = <0>; /* Chip select 0 */ spi-max-frequency = <10000000>; status = "okay"; partitions { compatible = "fixed-partitions"; #address-cells = <1>; #size-cells = <1>; partition@0 { label = "uboot"; reg = <0x0 0x100000>; }; partition@100000 { label = "env"; reg = <0x100000 0x100000>; }; }; }; }; compatible = "jedec,spi-nor"; The kernel uses a mapped driver (spi-nor) and creates the corresponding device, not /dev/spidevX.X

This no longer works in version v5.15 and above. All we can do for the user is to expand the warning to a more detailed one if the kernel receives such an outdated way of overlaying the device tree from the user. @IgorIgor, should we delete the files of outdated overlays for the device tree from the corresponding patches? If this SPI channel is already in use, why are you trying to use it in another way?

I just wanted to make a link to the documentation, which explains which OS Armbian supports, so as not to discuss the problem further. But I didn't find it.

@Werner Can I make a request to you? Somewhere on this page, write in large green letters the names of the operating systems that are currently supported. https://docs.armbian.com/#what-is-supported And in the section "How to report bugs" Ask to specify the minimum information about the OS: lsb_release -rc Or in some other way, the main thing is that users and you and I can understand that the question relates to a supported operating system.

Thank you for your feedback

A negative result is also a result. You have a choice. Update the system. Use the linux-image-legacy-sunxi system version of the package on the current system. It should be 5.10.y and the kernel modules are not compressed

Do not reinstall u-boot On board: git clone https://github.com/The-going/PKG_test cd PKG_test/cubietruck-bionic ls sudo dpkg -i linux-dtb-edge-sunxi_22.05.0-trunk_armhf.deb sudo dpkg -i linux-image-edge-sunxi_22.05.0-trunk_armhf.deb Packages must be installed from this directory. Non-system packages. If everything went without errors. Reboot

Igor, I didn't say the main thing. Spring has come in our area where I live. Every day adds new worries. I will switch off from the development process in 10 days. I will only be able to support the process of applying patches for sunxi. This is the imperfection of our own algorithms. We upload the same thing to the build server many times. This is the first in line for a fix. I've been trying to fix something in my repository, but I don't understand this code well and I'm doing even worse than it was. But perhaps we can discuss the algorithm in words? I do not know how to defeat the monster. Probably just to get away from him on github? Once a few years ago I was tracking this: https://mirrors.edge.kernel.org/pub/linux/kernel/v5.x/ChangeLog-5.17.4 285K https://mirrors.edge.kernel.org/pub/linux/kernel/v5.x/linux-5.17.4.tar.xz 122M Downloading 120-150 MB, checking and unpacking is fast. This is a ready-made archive and it is distributed for download.

@rgala Can you confirm or deny the operability of the kernel package?

I made the first attempt to test the branch https://github.com/armbian/build/tree/armbian-next . Unsuccessful. ... [🌱] Kernel build starting [ kernel/arm64__5.15__sunxi64 ] [🌱] Getting sources from Git [ unused:set via GIT_FIXED_WORKDIR linux-5.15.y ] [🌿] Creating local copy [ unused:set via GIT_FIXED_WORKDIR linux-5.15.y ] [🌱] Downloading Git cold bundle via HTTP [ https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/clone.bundle ] ... [👉] Resolving mirrors.edge.kernel.org (mirrors.edge.kernel.org)... 147.75.101.1, 2604:1380:2001:3900::1 [👉] Connecting to mirrors.edge.kernel.org (mirrors.edge.kernel.org)|147.75.101.1|:443... connected. [👉] HTTP request sent, awaiting response... 200 OK [👉] Length: 2530673996 (2.4G) [application/octet-stream] [👉] Saving to: ‘/home/leo/build/cache/gitbundles/cold/dda92618d5138e315cf9b5031c20a936.gitbundle’ [👉] 0K ........ ........ ........ ........ 1% 1.83M 21m44s [👉] 32768K ........ ........ ........ ........ 2% 2.37M 18m58s [👉] 65536K ........ ........ ........ ........ 3% 2.35M 17m58s ... The connection is broken on my side by the provider. Three subsequent attempts were unsuccessful. An attempt to download 460 megabytes already downloaded ends with a server failure and the program waits 10 seconds. Then again an unsuccessful replay. I had to delete a partially downloaded file. A 2.4GB file is excruciatingly long for my connection. Fantastic! On the fifth attempt, he coped. The compilation process has started [👉] E: Failed getting release file http://localhost:3142/ports.ubuntu.com/dists/focal/Release [👉] -->--> command failed with error code 1 after 1 seconds [👉] --> 🐛 4430: 24605 - 24605 - 24605: ehBE: debug: Error-related files found [ ] [👉] --> 💥 4430: 24605 - 24605 - 24605: ehBE: err: error: Debootstrap first stage failed The log file brought special joy: [🌿] Build log file [ /home/leo/build/output/logs/armbian-logs-4775c623-741d-4662-87fd-f4d1da533e6f.html ] From git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable * [new tag] v2.6.12 -> v2.6.12 * [new tag] v2.6.12-rc2 -> v2.6.12-rc2 * [new tag] v2.6.12-rc3 -> v2.6.12-rc3 * [new tag] v2.6.12-rc4 -> v2.6.12-rc4 * [new tag] v2.6.12-rc5 -> v2.6.12-rc5 * [new tag] v2.6.12-rc6 -> v2.6.12-rc6 * [new tag] v2.6.13 -> v2.6.13 * [new tag] v2.6.13-rc1 -> v2.6.13-rc1 * [new tag] v2.6.13-rc2 -> v2.6.13-rc2 * [new tag] v2.6.13-rc3 -> v2.6.13-rc3 * [new tag] v2.6.13-rc4 -> v2.6.13-rc4 * [new tag] v2.6.13-rc5 -> v2.6.13-rc5 * [new tag] v2.6.13-rc6 -> v2.6.13-rc6 * [new tag] v2.6.13-rc7 -> v2.6.13-rc7 * [new tag] v2.6.14 -> v2.6.14 For me, this branch is a very difficult and expensive case in terms of support.

sudo modinfo name_module | grep 'filename:'

First update u-boot and reboot. In a simple way: dpkg -i linux-u-boot-*.deb Then the dtb and the core and better on some kind of test flash drive.

I have a build environment for ubuntu bionic, so the first test for this OS. The build was done in VM-qemu-x86_64: @rgala https://github.com/The-going/PKG_test/tree/master/cubietruck-bionic

I want to test my own assumption. I will create a package for your specific OS in the environment of this OS. And I will provide you with a download link. You will install this kernel package into your system and check its operability. Ubuntu 18.04 and Debian 9. Wait for the link here.

Debian 10 and 11 are supported

I think we should do exactly the opposite. As soon as the author realizes that his codebase has begun to create stable images for the family\branch, he can invite the most active members of this group to test and fix bugs. That is, errors, changes must be made in this branch. While the master should remain stable and not create problems for the current work.

That's the problem. And I think the problem is not that it's a different OS. And the fact is that the kernel is initially assembled in the environment of another OS. Can we check it out if you agree? I will build a kernel package in this OS 18.04. Or update the operating system.