Bug 142483

Summary: powersave deamon causes computer to enter powerstate without user action
Product: [openSUSE] SUSE LINUX 10.0 Reporter: Armin Mohring <arminmohring>
Component: KernelAssignee: Pavel Machek <pavel>
Status: RESOLVED INVALID QA Contact: E-mail List <qa-bugs>
Severity: Normal    
Priority: P5 - None CC: trenn
Version: Final   
Target Milestone: ---   
Hardware: x86-64   
OS: SuSE Linux 10.0   
Whiteboard:
Found By: System Test Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---
Attachments: acpidmp.log
dmseg.log

Description Armin Mohring 2006-01-11 05:51:52 UTC
powersavedaemon (update last day) causes computer to enter powerstate without user action after completed boot.
Comment 1 Olaf Kirch 2006-01-11 07:31:24 UTC
Thomas, Pavel - this one is for you...
Comment 2 Pavel Machek 2006-01-11 10:09:11 UTC
It may be for Stefan ... if it is actually userland asking for suspend
I need more info here. (1) What is "enter powerstate"? Suspend-to-disk? Suspend-to-RAM? Something else?

(2) If you boot without powersaved, will running it manually cause that, too?
Comment 3 Armin Mohring 2006-01-11 11:45:48 UTC
Yesterday I made the update. Powersave daemon was updated, too.
After the boot process completed, the monitor went black and the only way to reawake the computer, was to press the power button. In the bios: S3 is enabled.
Comment 4 Forgotten User ZhJd0F0L3x 2006-01-11 12:32:26 UTC
and what happened when you pressed the power button? Did it resume or did it boot normally again?
How do you conclude that this was powersaved switching off your monitor?
Comment 5 Thomas Renninger 2006-01-11 14:11:38 UTC
Maybe a critical temp shutdown?
Or maybe a hot temp shutdown, S4 should be initiated by spec then, no idea whether it's really implemented that way, hot trip points are not widespread.
Do you get something with: grep -i critical /var/log/messages?
Comment 6 Armin Mohring 2006-01-11 15:01:15 UTC
Pressing the power button caused the computer to restart (into Bios).
Monitor LED (indicates off/on/DMPS) was blinking (=DMPS).

Jan 11 06:46:17 Sol kernel:     ACPI-0212: *** Warning: Device is not power manageable
Jan 11 06:46:17 Sol kernel: ACPI: PCI Interrupt Link [APCF] enabled at IRQ 22
Jan 11 06:46:17 Sol kernel:     ACPI-0212: *** Warning: Device is not power manageable
Jan 11 06:46:17 Sol kernel: ACPI: PCI Interrupt Link [APCG] enabled at IRQ 21
Jan 11 06:46:17 Sol kernel:     ACPI-0212: *** Warning: Device is not power manageable
Jan 11 06:46:17 Sol kernel: ACPI: PCI Interrupt Link [APCH] enabled at IRQ 20
Jan 11 06:46:17 Sol kernel:     ACPI-0212: *** Warning: Device is not power manageable
Jan 11 06:46:17 Sol kernel: ACPI: PCI Interrupt Link [APCL] enabled at IRQ 23
Jan 11 06:46:17 Sol kernel: shpchp: shpc_init : shpc_cap_offset == 0
Jan 11 06:46:17 Sol kernel: shpchp: shpc_init : shpc_cap_offset == 0
Jan 11 06:46:17 Sol kernel:     ACPI-0212: *** Warning: Device is not power manageable
Jan 11 06:46:17 Sol kernel: ACPI: PCI Interrupt Link [APCJ] enabled at IRQ 22
Jan 11 06:46:17 Sol kernel: ACPI: PCI Interrupt Link [APC2] enabled at IRQ 17
Jan 11 06:46:17 Sol kernel: ACPI: PCI Interrupt Link [APC4] enabled at IRQ 19
Jan 11 06:46:17 Sol kernel: microcode: CPU0 not a capable Intel processor
Jan 11 06:46:17 Sol kernel:     ACPI-0252: *** Warning: Error evaluating _PSS
Jan 11 06:46:17 Sol kernel:     ACPI-0252: *** Warning: Error evaluating _PSS
Jan 11 06:46:17 Sol [powersave]: ERROR (CPUFreq_Kernel:21) Can not set ondemand governor, maybe your cpufreq driver is too slow.
Jan 11 06:46:17 Sol [powersave]: ERROR (CPUFreq_Kernel:23) Try CPUFREQ_CONTROL=userspace in /etc/sysconfig/powersave/cpufreq.
Jan 11 06:46:17 Sol [powersave]: ERROR (PM_Interface:231) Cannot create cpufreq objects
Comment 7 Thomas Renninger 2006-01-11 15:26:42 UTC
This seem to be an ACPI interpreter problem.
Could you please attach acpidmp output.
Full dmesg output could also be helpfull.
Comment 8 Armin Mohring 2006-01-11 17:04:45 UTC
Created attachment 62976 [details]
acpidmp.log
Comment 9 Armin Mohring 2006-01-11 17:05:30 UTC
Created attachment 62977 [details]
dmseg.log
Comment 10 Pavel Machek 2006-01-11 21:18:54 UTC
Can you answer questions from comment #2? How long does it take to enter powersave state? Do you see some disk activity, or some fast console switches? Does your power LED blink or is it only monitor going powersave? Or does it look like regular (but fast) poweroff?
Comment 11 Thomas Renninger 2006-01-12 09:51:49 UTC
Is this is an AMD64 machine or an Intel EM64T?
In first case it is really strange that cpufreq states are not supported via ACPI... Even this should not cause the shutdown/suspend (the _PSS message should be harmless in this regards), it may indicate that you run on an old BIOS. Please search for BIOS updates for this machine before proceeding.
Comment 12 Armin Mohring 2006-01-12 13:47:36 UTC
1. The boot screen apears (vesafb mode) and directly after switching to graphics mode (nvidia 1.0-7676) the monitor screen went black. The power led of the monitor 
was blinking (=DMPS mode). The power led of the computer was on (no flashing). The 
LED for the disk activity was flashing from time to time.
powersave -U is not working on this computer, since after powerdown, the computer tries to load the pages from swap (1GB RAM, 1.5GB swap), completes loading the pages and reboots then in to Bios.

2. It is a AMD64 with a nForce 3 250GB Mainboard. I have installed the lastet (working) bios for this mainboard (Version 1.7 and not 2.1).
Comment 13 Armin Mohring 2006-01-12 14:07:22 UTC
1. If I enable "cool and quiet" in the mainboard bios (MSI K8Neo Platinum) powersave -U causes the computer to save pages to swap. Reawake with keyboard, mouse and pressing the powerbutton fails. THe only way to restart the computer is to press the Reset button.
Comment 14 Thomas Renninger 2006-01-12 14:11:53 UTC
comment #13 -> You mean the system freezes at suspend to disk?
Does this machine have an SATA disk? IIRC there where some issues with those.

Is this also what you see when booting? Pages are saved to swap, etc.?
If yes, have you played with some scripts to try to initiate a suspend to disk and added a powersave -U somewhere?
If yes, then this is an invalid...

If not:
Please try what Pavel suggested:
Disable powersave daemon: chkconfig powersaved off.
Best you also don't load any acpi modules:
set ACPI_MODULES="NONE" in /etc/sysconfig/powersave/common
and remove "thermal fan reiserfs" in INITRD_MODULES variable in /etc/sysconfig/kernel
then invoke mkinitrd and reboot.
Comment 15 Armin Mohring 2006-01-12 15:09:07 UTC
1. If have two SATA disks installed.
2. I haven´t changed any scripts. 
3. powersave -U: I see on console: freeing RAM, pages are saved to disk. 
4. I cannot remove reiserfs from the initrd_modules list, since /boot and / are formatted with reiserfs.
Comment 16 Armin Mohring 2006-01-12 15:32:30 UTC
1. ACPI module weren´t loaded
2. initrd_modules: removed thermal and fan

powersave -U: pages are saved to disk, computer shuts down and pressing the power button causes the computer to reawake. Pages are loaded from disks and then the computer hangs.
Comment 17 Forgotten User ZhJd0F0L3x 2006-01-12 20:12:11 UTC
(In reply to comment #12)
> 1. The boot screen apears (vesafb mode) and directly after switching to
> graphics mode (nvidia 1.0-7676) the monitor screen went black. The power led of
> the monitor 
> was blinking (=DMPS mode). The power led of the computer was on (no flashing).
> The 
> LED for the disk activity was flashing from time to time.

This means that the machine is perfectly working, just that the X server is not configured or coming up good. This has nothing to do with powersave at all.
Retry without the binary NVidia driver and open a bugreport against component X if it fails again.

> powersave -U is not working on this computer, since after powerdown, the
> computer tries to load the pages from swap (1GB RAM, 1.5GB swap), completes
> loading the pages and reboots then in to Bios.

This is a different bug. Please open one report per issue. Also, since you are using the binary NVidia graphics drivers, there is nothing we can do about this.