Some DUTs can't boot up after installing the hwe kernel in the proposed on Focal

[Kevin Yeh]

During the SRU testing for 5.15.0-112 kernel, I found some machines freeze at very early stage of booting process.
The last messages displayed on the screen are:
Loading Linux 5.15.0-112-generic...
Loading initial ramdisk...

It looks like it's very similar to this bug.
https://bugs.launchpad.net/ubuntu/mantic/+source/linux/+bug/2061940

Here is the list of the impacted machines that I found so far
https://certification.canonical.com/hardware/202107-29332
https://certification.canonical.com/hardware/202107-29331
https://certification.canonical.com/hardware/202105-29007
https://certification.canonical.com/hardware/202104-28970
https://certification.canonical.com/hardware/202104-28965
https://certification.canonical.com/hardware/202104-28966
https://certification.canonical.com/hardware/202104-28968
https://certification.canonical.com/hardware/202104-28969
https://certification.canonical.com/hardware/202104-28970

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Roxana Nicolescu (roxanan)]

It is not the same issue because jammy:linux (the parent of focal:linux-hwe-5.15) do not have the commits that broke mantic

[Andrew Bonkowski (watchernine)]

I had this issue just now as well. I rolled back to the previous kernel and it works. I am new to Linux, so I know zero about how to capture errors and all that, but I'm running an Asus laptop and haven't even figured out how to list my hardware using Mint yet. I have a long way to go before I can replace windows with THIS, especially if it's prone to bugging out where I can't boot.

I am using kernel 5.15.0-107. That is the one that works for me. I've tried 5.15.0-112 several times and no-go, it locks on boot. After 2, sometimes 3 attempts it drops to a menu that offers options for choosing which kernel to boot from. I rolled it back to 107 a few days ago and have not tried it again since because I don't know if anything is being done on it without changing the version number or if a new version number will supercede -112.

Question: what happens if I go to 6.5? I'd ask about 5.19 and 6.2, but the updater says they're "end of life". Should I try 6.5.0-41? That one says it's supported to August 2024 (and what happens after August?)

Sorry if the question is dumb, but like I said, I'm new to Linux.

[Ivan Hu (ivan.hu)]

Do some tests with the HP machine,

1. with flash OEM image, kernel version 5.10.0-1044-oem, directly install the proposed kernel version(5.15.0-112.122~20.04.1), it can boot up without issue.

2. with flash OEM image, kernel version 5.10.0-1044-oem, directly do the disk-upgrade(all update packages on proposed), this is cert lab do.
  $sudo apt disk-upgrade
  then the kernel "cannot" boot up.

3. after following 1 step, boot up with 5.15.0-112 kernel, then do the proposed disk-upgrade
  reboot, then it cannot boot up
  on this procedures, it informed that the grub-efi-amd64.bin will not be updated.

4. force reboot after 3, and boot with bootable old kernel(5.10.0-1044-oem), then install grub-efi-amd64.bin, then reboot, the 5.15.0-112.122~20.04.1 can boot up.

5. after 2 on the kernel can not boot up and then reboot to bootable old kernel(5.10.0-1044-oem), do the disk-upgrade(all update packages on proposed). Then 5.15.0-112 kernel can boot up.

From the above tests, directly install the 5.15.0-112.122~20.04.1 kernel has no problem, it cannot boot up only when flashing the oem image the directly do the disk-upgrade with all the packages on proposed. It seems some update procedures left the grub not be updated that causes boot up issue. If after flashing image do the disk-ugrade without proposed, then do the disk-upgrade with proposed, then the grub and kernel(5.15.0-112.122~20.04.1) will be updated, it can boot up without problems. So, I believe this is not the kernel issue.

[Kevin Yeh (kevinyeh)]

Cert team has already fixed the test procedure to prevent this issue happen, so I think this bug can be closed.