marcello7 Posted May 14, 2018 Posted May 14, 2018 Does someone know why the kernel reverted to 4.9.* from 4.14.*?
Igor Posted May 14, 2018 Posted May 14, 2018 2 hours ago, marcello7 said: Does someone know why the kernel reverted to 4.9.* from 4.14.*? Because 4.14.y has issues - some people ended up in a boot loop - which is not resolved yet.
marcello7 Posted May 14, 2018 Author Posted May 14, 2018 Ty Igor. I inadvertitly updated my HC1 (going from 4.14 to 4.9) and and the HC1 didn't boot anymore. Maybe I end in such a loop, but I have a headless configuration and I cannot troubleshoot it. The HC1 is not a production one (I have an OrangePI+2 for that) and I were experimenting its power requirements. Before flashing the sd with a fresh install, I don't remember if the 4.9.* has UAS support. Otherwise I will wait for kernel fixing on the 4.14 branch.
Igor Posted May 14, 2018 Posted May 14, 2018 2 hours ago, marcello7 said: I inadvertitly updated my HC1 (going from 4.14 to 4.9) and and the HC1 didn't boot anymore. Problems were detected when upgrading from 4.9 to 4.14 ... kernel 4.9 is just fine for and has UAS support.
chwe Posted May 14, 2018 Posted May 14, 2018 6 hours ago, marcello7 said: Maybe I end in such a loop, but I have a headless configuration and I cannot troubleshoot it. Yes We Can! (hmm, the slogan is a bit 2008 , and USB-UART is maybe also a bit outdated but still a must when you want to figure out what happens) https://wiki.odroid.com/accessory/development/usb_uart_kit 3 hours ago, Igor said: Problems were detected when upgrading from 4.9 to 4.14 someone figured out why? ___ _ _ _ __ ___ _ _ _ / _ \ __| |_ __ ___ (_) __| | \ \/ / | | | || | | | | |/ _` | '__/ _ \| |/ _` | \ /| | | | || |_ | |_| | (_| | | | (_) | | (_| | / \| |_| |__ _| \___/ \__,_|_| \___/|_|\__,_| /_/\_\\___/ |_| Welcome to ARMBIAN 5.44.180506 nightly Debian GNU/Linux 8 (jessie) 4.14.39-odroidxu4 System load: 0.05 0.10 0.09 Up time: 6 days I updated mine 6 days ago (from 4.9 to 4.14, OMV) The only errors I have is an IO error related to the HDD attached to it (might need some attention, but definitively not a bootloop issue).
JMCC Posted May 23, 2018 Posted May 23, 2018 Is there a thread where this issue is being followed? Has any dev been able to reproduce the bug? I didn't have problems either when upgrading from 4.9 to 4.14, so it has to be something very specific. 4.14 has many advantages over 4.9, such as crypto acceleration (at least twice as fast) and GPU/VPU support. It would be very good if we could narrow down what the problem is.
JMCC Posted May 25, 2018 Posted May 25, 2018 Another crazy idea could be changing the board name in the kernel packages. I mean: change "linux-image-next-oroidxu4" for something like "linux-image-next-5422", taking opportunity in the fact that there are different boards based on that SoC, and now armbian-config reflects it by providing the option to do board-based dtb tweaks (BTW that feature is an excellent addition). That way, new images would be based in the 4.14 kernel, while the existing images would not be updated automatically. In the release notes, there can be some explanation telling the users that they can upgrade to the new kernel, but warning them to do a full backup before, just in case they are one of the rare cases where the update causes a bootloop.
zador.blood.stained Posted May 25, 2018 Posted May 25, 2018 34 minutes ago, JMCC said: Another crazy idea could be changing the board name in the kernel packages. I mean: change "linux-image-next-oroidxu4" for something like "linux-image-next-5422", taking opportunity in the fact that there are different boards based on that SoC, and now armbian-config reflects it by providing the option to do board-based dtb tweaks (BTW that feature is an excellent addition). There are a lot of ways to improve the current situation if we don't care about maintaining easy upgrades and backwards compatibility. I.e. instead of hardcoded kernel branch names "default", "next" and "dev" and an artificial 3 branch limit we could have more than 3 branches with unique names, i.e. "hardkernel-4.9", "lts-4.14", "mainline", "hardkernel-3.10".
JMCC Posted May 25, 2018 Posted May 25, 2018 13 minutes ago, zador.blood.stained said: There are a lot of ways to improve the current situation if we don't care about maintaining easy upgrades and backwards compatibility Yeah, definitely, leaving existing 4.9 kernel users without updates doesn't fit "the Armbian way" of doing things, even though Hardkernel themselves don't update their 4.9 kernels anymore. That would be a dirty trick that I would do, taking excuse on the fact that new 5422-based boards have been released and so justifying the change in the package names, but doing it really to get rid of the kernel upgrade problem that is getting so old. I'm glad that you guys are more serious than that
Blue_Ninja0 Posted July 6, 2018 Posted July 6, 2018 Odroid XU4 seems to have a bug in the USB 3.0 driver that causes it to fail occasionally on high load on kernel 4.9 (there's no power delivery problem). It seems to have been fixed in subsequent kernels so it is important for XU4 to get the mainline kernel. Are there any updates on this? Or is there a way I can install newer kernels on my XU4?
Igor Posted July 6, 2018 Posted July 6, 2018 25 minutes ago, Blue_Ninja0 said: Are there any updates on this? Or is there a way I can install newer kernels on my XU4? It's getting ready. You can install it from a nightly repository, kernel name DEV. 1
Blue_Ninja0 Posted July 6, 2018 Posted July 6, 2018 2 hours ago, Igor said: It's getting ready. You can install it from a nightly repository, kernel name DEV. Good to know it's in the pipeline! Is it easy for me to downgrade back to NEXT kernel from DEV when it gets stable?
Igor Posted July 6, 2018 Posted July 6, 2018 27 minutes ago, Blue_Ninja0 said: Good to know it's in the pipeline! Is it easy for me to downgrade back to NEXT kernel from DEV when it gets stable? According to Hardkernel folks, it's stable. Last time I want to make a switch (again) I made a few tests and sometimes I could not boot. If you know what you are doing is easy. Everything is packed (kernel, dtb, u-boot, root, headers, firmware) and it was explained how to do it manually in case armbian-config way of switching kernels and repositories fails. You can try, but if you fail you are on your own to do some forum searching and fix the problem. 1
prof Posted July 8, 2018 Posted July 8, 2018 The Hardkernel folks say 4.14 is stable. You find that it's not ("I made a few tests and sometimes I could not boot."). Is there a resolution in the offing? Kernel bugs to be reported upstream? Issues in the Armbian distro (not present in Hardkernel's distro)? Are we waiting for the release of 18.04.1 ?
Igor Posted July 9, 2018 Posted July 9, 2018 5 hours ago, prof said: Are we waiting for the release of 18.04.1 ? No, that is irrelevant and btw. Ubuntu 18 is also packed with bugs but Canonical sell it as stable. Do we fix their bugs? Few of them we did but we can't fix them all. We are just a few people and we can't afford to do that in such scale. As you can see even a big company with a bunch of full-time engineers cannot do that. A switch is in the pipeline. We are waiting for slots of our free time. This is very limited lately and we have little options to do anything in this area. I would like to do or see more tests before pushing out an update. We can't afford that it doesn't boot after an upgrade since support is also extremely limited and our direct cost. That is also the reason why Ubuntu 18 is labeled testing.
Recommended Posts