Qubes OS Live USB Installation Bootloop

After running Qubes OS for a while, I decided to buy a new laptop since mine is a bit old and unfortunately its RAM is soldered so I can’t upgrade.

So I bought this new Lenovo Yoga Pro 7 Gen 9, but after booting from USB and selecting either of the install options (default, test and install, verbose, latest kernel…) I see the boot log for a few seconds and then the pc reboots.

I’ve used this USB stick and its contents to install Qubes OS on two other devices, both a desktop and my old laptop, without problems.

I also tried another USB device, I tried using rufus instead of dd, but the problem is still the same. I can get to the boot menu, but then the OS installer doesn’t boot.

Now, I’d try to boot and install in legacy mode, as I read on the docs that there’s some bug with Lenovo and UEFI , but unfortunately Lenovo removed that option altogether for this new shiny laptop of theirs.

Am I cooked? Should I just return this pc and buy something else to run Qubes OS on?

1 Like

Yes.

After fiddling around a bit I found out that by adding acpi=off it will boot, but of course the internal mouse and keyboard won’t work.

Also, X will fail to start, but by adding both the above kernel parameter and blacklist_module=amdgpu I can reach the installer.

Any way to narrow down what acpi module to try and blacklist to avoid putting off the entire thing?

1 Like

My suggestion is to proceed with installation, then after installation is complete, boot the kernel with those parameters, and perform first boot setup.

Why? Because then journalctl will tell you everything you need to know if you get a black screen :wink:

If you are comfortable posting it here, we can help.

1 Like

I believe the Lenovo Yoga Pro 7 Gen 9 is available with both AMD and Intel CPUs … if that’s correct, can you tell a bit more about the hardware?

I know the blacklist_module=amdgpu is a hint - but it’s better to know than guessing …

:slight_smile:

1 Like

I have the version with

AMD Ryzen AI 9 365
AMD Radeon 880M

1 Like

Try to edit the boot options to include:

console=vga vga=,keep loglvl=all guest_loglvl=all noreboot=1 earlyprintk=xen

to the Xen command line (the one with console=none) and replace the quiet on the kernel command line to have:

blacklist_module=amdgpu nomodeset

and see if that helps.

:slight_smile:

1 Like

So without putting acpi=off?

I tried that way, adding the parameters you suggested. The boot process goes on seemingly forever with logs that get printed out constantly and very slowly.

1 Like

Ok, now I got a kernel panic:

1 Like

Yes - without acpi=off – as I understod it, the machine boots fine when ACPI is disabled and now the question is “why doesn’t it boot when ACPI is enabled?”

The options to Xen makes it very verbose and makes sure to print messages to the sceen for debugging (resulting in a slooooow boot). At the same time, Xen is instructed to not reboot when something crash, so we can get a hint about “Who killed the kernel?” … and with the noreboot=1, we get a chance to take pictures of the messages.

I also have a Lenovo laptop (T14, Gen3) with AMD CPU/GPU that I would like to run Qubes on … and I’ve made it to this point as well and is trying to decode the kernel stack trace to figure out what I need to do, to make it run with ACPI on … but I’m stuck at this point. :-/

So I hope we can exchange suggestions/ideas about what could be the next step - and maybe someone with more insight from the forum will chip in with things we can dive into.

:slight_smile:

2 Likes

I already made my suggestion clear nearly a month ago.