Bug 30083

Summary: Shutdown is not complete with mga8
Product: Mageia Reporter: Pierre Jarillon <jarillon>
Component: RPM PackagesAssignee: Kernel and Drivers maintainers <kernel>
Status: NEW --- QA Contact:
Severity: normal    
Priority: Normal CC: davidwhodgins, johnltw
Version: 8   
Target Milestone: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Source RPM: CVE:
Status comment:

Description Pierre Jarillon 2022-02-23 14:43:29 CET
Two laptops ASUS X70iD series were working fine with MGA7. After upgrading to MGA8, shutdown cannot end properly. All process are killed, the screen is black but the LED of power button and the wifi LED are still on.
It is necessary to press the power button during several seconds to have a complete shutdown.
Shutting off in console mode, the last lines are:
[ OK ] Finished Power-Off.
[ OK ] Reached target Power-Off.
There is a regression, mga7 was able to finish the job, mga8 cannot...
Comment 1 sturmvogel 2022-02-23 15:53:07 CET
This is most likely a BIOS setting. It's related to APM. Search for power management settings and turn off ErP/EuP.
Lewis Smith 2022-02-23 21:03:56 CET

CC: (none) => lewyssmith

Comment 2 Pierre Jarillon 2022-02-23 23:46:10 CET
There is no power management in the BIOS...

With dmidecode, I have found the same release of the BIOS in both PC:

BIOS Information
	Vendor: American Megatrends Inc.
	Version: 203    
	Release Date: 02/01/2010
BIOS Revision: 1.84
	Firmware Revision: 153.153

As USB is not managed from the BIOS, I have not found how to upgrade the BIOS at a newer version.
Comment 3 John L. ten Wolde 2022-02-24 22:19:39 CET
Hi Pierre.  Are you running the latest available kernel (5.15.23)?

  ┌────
  │ $ uname -r
  │ 5.15.23-desktop-1.mga8
  └────

I ask because I saw exactly the same behaviour you're describing a couple of months back on an Acer Aspire 3.  That machine had already been running Mageia 8 since its release, but after a kernel update it also "Reached target Power-Off" without ever actually powering *off*.  This was completely out of the blue.

A few days later another kernel update dropped and the problem -- whatever had caused it -- was resolved without my ever getting around to intervening with or reporting about it.  I swear it seems like kernel updates have been dropping practically one a week for the past four months.  Even now, I see 5.15.25 sitting in testing...

Sorry I can't be more specifically helpful.

CC: (none) => johnltw

Comment 4 Pierre Jarillon 2022-02-25 00:32:48 CET
Before returning the PC to the owner, I did this:
- I have booted on the latest kernel of Mga7 still available after upgrade and the problem is the same.
- I have tried to restart (instead of shutdown) and the process goes in the same state: power and wifi leds on, all process killed.

I am not sure that the kernel is involved. It seems that an instruction is missing, but I don't know how to find it.
Comment 5 Lewis Smith 2022-02-26 10:09:36 CET
> It seems that an instruction is missing
I imagine you mean an instruction to the hardware.

Can you please attach just the end of the journal of a session that fails to power down?
Doing it from the following session:
 # journalctl -b-1 --no-hostname

Mine shows (desktop PC with kernel 5.15.23-desktop-1.mga8):
...
Chw 25 22:28:36 systemd[1]: Reached target Unmount All Filesystems.
Chw 25 22:28:36 systemd[1]: Stopping Monitoring of LVM2 mirrors, snapshots etc.>
Chw 25 22:28:36 systemd[1]: systemd-remount-fs.service: Succeeded.
Chw 25 22:28:36 systemd[1]: Stopped Remount Root and Kernel File Systems.
Chw 25 22:28:36 systemd[1]: systemd-tmpfiles-setup-dev.service: Succeeded.
Chw 25 22:28:36 systemd[1]: Stopped Create Static Device Nodes in /dev.
Chw 25 22:28:37 systemd[1]: lvm2-monitor.service: Succeeded.
Chw 25 22:28:37 systemd[1]: Stopped Monitoring of LVM2 mirrors, snapshots etc. >
Chw 25 22:28:37 systemd[1]: Reached target Shutdown.
Chw 25 22:28:37 systemd[1]: Reached target Final Step.
Chw 25 22:28:37 systemd[1]: systemd-poweroff.service: Succeeded.
Chw 25 22:28:37 systemd[1]: Finished Power-Off.
Chw 25 22:28:37 systemd[1]: Reached target Power-Off.
Chw 25 22:28:37 systemd[1]: Shutting down.
Chw 25 22:28:37 systemd-shutdown[1]: Syncing filesystems and block devices.
Chw 25 22:28:37 systemd-shutdown[1]: Sending SIGTERM to remaining processes...
Chw 25 22:28:37 systemd-journald[489]: Journal stopped

CC'ing DaveH for his view.

CC: (none) => davidwhodgins

Comment 6 Dave Hodgins 2022-02-26 11:18:08 CET
Does it do any better if the user is logged out and then shutdown from the
login screen?

Assigning to the kernel team.

Assignee: bugsquad => kernel

Lewis Smith 2022-02-26 20:50:17 CET

CC: lewyssmith => (none)