Updating to v6.6.20 now my VM hangs during boot

[k5map]

This weekend I updated my RPi OS VM with "apt upgrade" which installed v6.6.20.  But now when they boot, they hang at:

Loading Linux 6.6.20+rpt-rpi-v8 ...
Loading initial ramdisk ...

Anyone know how to fix this so they will complete the boot process?

[cyprienlaplace]

There is one thing that I find odd, that is the name of the linux kernel "6.6.20+rpt-rpi-v8". The "rpi" part suggests that it is a Raspberry Pi customized kernel, but it is booted inside the VM, using EFI+grub... Would that mean that RPi is going to support UEFI natively in the future??

You can try to reboot and choose the previous kernel image, usually they are kept and grub should show up the boot menu if the previous boot failed. Otherwise try to press the ESCape key during early boot.

[cyprienlaplace]

I did reproduce it, and to my surprise yes, the RPi kernel now comes with EFI support enabled. So grub-efi picks it up and make it the default boot option. But that RPi linux kernel does not support ACPI tables, and is failing to boot.

Please delete the rpi kernels: sudo apt remove linux-image-*rpi*

If you are looking for a more recent kernel from debian, you can use the debian-backports repo (https://backports.debian.org/Instructions/), but only 6.6.13 is available right now.

Debian testing has 6.6.15 and sid has 6.7.9.

[k5map]

I can't get the VM to boot in order to enter the apt remove command.  Any ideas how to get it to boot?

[cyprienlaplace]

Can't you use the grub menu to use "advanced" options to start an older kernel?

If there are no other kernel, then you will have to use a rescue ISO from debian.

[k5map]

Ok I was able to interrupt the boot process and select the previous kernel.  I then deleted the images per your command and the VM is back up.  

I upgraded to the bad kernel via the apt update & apt upgrade commands.  If I run this in the future will I have the same issue?

[BarryGrowler]

The issue with your RPi OS VM not booting after updating to kernel 6.6.20 is due to incompatibilities between the RPi-specific kernel and the VM environment. The RPi kernel now includes EFI support, which grub-efi selects as the default, but it fails to boot because of missing ACPI table support. To fix the issue, interrupt the boot process and choose an older, working kernel from the grub menu.

Once booted, remove the problematic RPi kernels with: sudo apt remove linux-image-rpi

To prevent this in the future, use standard Debian kernels instead of RPi-specific ones. Consider using kernels from debian-backports, testing, or sid repos based on your stability needs. In case a bad kernel update prevents booting, use the grub menu to select an older kernel or boot from a Debian rescue ISO to remove the faulty kernel. Stick with standard Debian kernels for your RPi OS VM to ensure compatibility and avoid boot problems.