Strange ACPI errors related to deactivating devices in BIOS.

[mandawebs]

Good evening everyone,

I have installed the last Proxmox version, with kernel /boot/vmlinuz-5.19.17-1-pve. This kernel was needed to get a Radeon GPU properly passthrough.

My motherboard is an Asus Z690 Creator Wifi. Bios Version 1760.

When I deactivate some devices I will not use in Proxmox VMs (v.g. 10G NIC, Bluetooth, Wifi Card, Serial Port) I start getting huge amounts of ACPI-related logs.



When I reactivate these devices in BIOS, all messages are gone.

My /etc/default/grub looks like:



Any clue about how this can be happening? Any solution that makes possible to have these devices deactivated and not getting that huge storm of ACPI-related logs.

Many thanks!,
Mandawebs

 

[mandawebs]

Hello,

Reviewed again and, even when I am quite convinced the errors started appearing after changing the BIOS, it does not seem it is only related to activate/deactivate devices.

In fact, I found traces of similar bugs found in the kernel.

http://jhshi.me/2015/11/14/acpi-error-method-parseexecution-failed-_gpe_l6f/index.html#.Y7YaInnMK3J

Will keep on digging.

Thx,
Mandawebs.

 

[paulkingnz]

I have similar errors on a fresh install on a Toshiba Satellite C55 laptop. Proxmox runs however the screen is full of ACPI errors and keeps scrolling while you are trying to choose options during the install. Hope this info helps

 

[mandawebs]

Hello,

Validated my own assumptions. Did the following:

- Upgraded to last version the BIOS.
- Dmesg was quite clean, much cleaner than previously.
- Deactivated the same said components.
- The log started to become full at top speed again.

Let´s see if someone is able to point in any direction.

Thanks!,
Mandawebs.

 

[paulkingnz]

I fixed mine with some help from Reddog1 and it is now working without the error messages:

To condense it all into simple form, you need to change the log level by editing /etc/default/grub (which some distributions do by default, now, to suppress the messages).

#sudo nano /etc/default/grub

arrow down and change GRUB_CMDLINE_LINUX_DEFAULT="quiet splash"

to:

GRUB_CMDLINE_LINUX_DEFAULT="quiet splash loglevel=3"

An alternative, is to print all the boot messages at startup, which will lose the acpi errors in the scrum by removing the quiet splash (some distributions print all boot messages by default, and it is what I prefer to do, myself, making the line look like this)

GRUB_CMDLINE_LINUX_DEFAULT=""

Whichever method you choose, you must save and run #sudo update-grub

Reboot

 

[keeka]

I fixed mine with some help from Reddog1 and it is now working without the error messages:

To condense it all into simple form, you need to change the log level by editing /etc/default/grub (which some distributions do by default, now, to suppress the messages).

#sudo nano /etc/default/grub

arrow down and change GRUB_CMDLINE_LINUX_DEFAULT="quiet splash"

to:

GRUB_CMDLINE_LINUX_DEFAULT="quiet splash loglevel=3"

An alternative, is to print all the boot messages at startup, which will lose the acpi errors in the scrum by removing the quiet splash (some distributions print all boot messages by default, and it is what I prefer to do, myself, making the line look like this)

GRUB_CMDLINE_LINUX_DEFAULT=""

Whichever method you choose, you must save and run #sudo update-grub

Reboot

I like your 'fix' ;-)
These ACPI method warnings are fairly common IME. More so on older boards and when you disable an onboard device. Most of the time they are harmless and usually down to the board vendor's ACPI implementation. I see them on my PVE host every time i power on a SATA hotplug drive and it annoys me every time I see them.

 

[mandawebs]

Ok, many thanks.

So I understand that main takeaway is that they are harmless and can be ignored, correct?.

Thanks!,
Mandawebs.

 

[keeka]

In my experience, yes. And if these only appear when you disable devices on the mobo, I'm guessing even more harmless.
It seems many of these boards' firmware are written with windows in mind rather than a complete and correct implementation. Usually, all the annoying errors/warnings appear at boot then don't resurface unless, as in my case, you plug some hardware in that relates to those acpi functions. Search for those acpi symbols/methods, it might turn up something to explain it. However only solution is an updated BIOS.
I think these ACPI warnings will apear when you boot any recent linux kernel/distro.

 

[keeka]

Hello,

Reviewed again and, even when I am quite convinced the errors started appearing after changing the BIOS, it does not seem it is only related to activate/deactivate devices.

In fact, I found traces of similar bugs found in the kernel.

http://jhshi.me/2015/11/14/acpi-error-method-parseexecution-failed-_gpe_l6f/index.html#.Y7YaInnMK3J

Will keep on digging.

Thx,
Mandawebs.

By the way that is not a kernel bug. It's a BIOS issue.
Actually that post mentions a way of disabling specific ACPI methods. There might lie your solution.

 

[mandawebs]

Ok. BIOS was already updated to the last version, so I guess there is little more to do in that front.

Thx!,
Mandawebs.