Ubuntu mate 22.04 shutdown acts as log out

Hi.

In the past weeks I can shutdown my precision m4800 only with sudo poweroff and then click on power button to wake it up and then it shuts down.

Sudo shutdown doesn't do anything besides displaying a scheduled message.

Shut down from the menu acts as logout.

I ran sudo snap refresh --hold to rule out snap backgroumd updates.

I have latest bios A26 installed.
Also Dual boot with win 11 in UEFI mode both mate and win.

Any ideas?

On win it is shutting down properly. I ran a dell diagnostic and the hardware is ok.

I use:

sudo shutdown now

you can learn more with:

man shutdown

1 Like

now or after 2 minutes is the same command. I use it like this to schedule shutdown after a period. Still did not work. I wrote a post in here somewere with a bash file for that.

Is this behavior from a clean install?

If you boot the machine with an install iso and choose to try it out rather than install MATE, can you power it off?

Just trying to rule out something odd with the software.

no. it is with old install. I'll try with a usb and get back to you.

sudo shutdown now is putting it into sleep and when I click power again it wakes up and shuts down... so same as
sudo poweroff

Most likely not problem, just a thought maybe check settings in attached image. Image shows dropdown choices.

2 Likes

Power button pressed while sleeping ->wakes the laptop and then it is shutting down. The setting from the picture is ask me in my case, and if i click shut down, then it behaves like log off.

2 Likes

@mdd12

I can try ubuntu from usb stick only in legacy mode.
I got an error just after I tried ubuntu. clicked send.
Then shut down from the menu.
Please remove usb and press enter.
Then it shuts down.

Update. I shut it down just after I started the laptop. it worked. Retrying after I open firefox.

Ha... it reproduced.... snap is the issue again... even if snap refresh says all snaps are up to date.

This sounds a lot like your BIOS having a buggy ACPI implementation.
Or at least its ACPI table is not following the ACPI standard.

You might try to update the BIOS or, in case there are no updates to find, use a boot parameter (like 'acpi=off' to see if APM takes over and does a better job of powering down the machine or 'acpi_osi=Linux' to change the ACPI behaviour)

You can interactively test the sanity of your ACPI:

If all else fails:
Some people manually edit the ACPI tables.
(be warned though. Editing the ACPI table is a daunting adventure: You'd have to take out your dsdt and disassemble it to review the code and get a better understanding of what is broken. You probably don't want to go there)

MS-Windows will have ACPI patches from Dell.
It doesn't mean that the Dell BIOS is following the ACPI standard.
Neither does MS-Windows by the way (they don't adhere to any standard...not even to their own beloved OOXML).

1 Like

@tkn I have i8kutils to avoid noisy fans. could that be the issue?

I don't think so, but i don't know enough about i8kutils so I really can't say for sure.


with acpi off...

Just a side clarifying question to help my understanding. Unit was working correctly prior to problem. If you start laptop and use other apps without starting Firefox does the shutdown work properly and only fails after or when Firefox is open?

Now I have to revert this acpi off thing so I can start it...

Ah yes, USB2 errors. This is not something to worry about. Does it boot ? And after that: does it shut down properly ?

Tried now acpi_osi=Linux.

acpi=off resulted in that screen and nothing else

So without opening firefox it shut down even if I started caja and power statistics with acpi_osi=Linux.

Retrying with firefox.. it shut down... but I managed to turn it on from the second try only.

yeap... it does not start. reverting the grub changes....

Here are some other ACPI= strings that you can try.

ACPI="Windows 2013" for instance, makes the BIOS believe that it is talking to Windows 8.1 which is probably the windowsversion that was installed originally on this computer.
(And the ACPI was probably only tested against that)

More background about this ACPI problem:
https://www.kernel.org/doc/html/latest/firmware-guide/acpi/osi.html

To be fair, it is actually a miracle that most PC's work without ACPI problems.

2 Likes