Discussion:
Hibernation not working
Add Reply
Peter Nabbefeld
2016-12-31 06:38:05 UTC
Reply
Permalink
Raw Message
Hello,

since some time, hibernation does not work, dmesg contains following
messages:

[ 738.404244] ACPI Warning: \_SB.PCI0.PEG0.PEGP._DSM: Argument #4 type
mismatch - Found [Buffer], ACPI requires [Package] (20160422/nsarguments-95)
[ 738.404353] ACPI: \_SB_.PCI0.PEG0.PEGP: failed to evaluate _DSM
[ 738.404356] ACPI Warning: \_SB.PCI0.PEG0.PEGP._DSM: Argument #4 type
mismatch - Found [Buffer], ACPI requires [Package] (20160422/nsarguments-95)

Some time ago, when I closed my laptop, Linux automatically hibernated,
now it does not. As my laptop becomes really warm and I don't want to
stress my display, I'm using this rarely, so I cannot tell, since when
hibernation doesn't work :-/

Kind regartds
Peter
Andreas Innerlohninger
2016-12-31 09:42:08 UTC
Reply
Permalink
Raw Message
I noticed that too since some time. System doesn't hibernate and xorg
causes 100% cpu load.

Process with 100% cpu load was
/usr/lib/xorg-server/Xorg :0 -seat seat0 -auth /run/lightdm/root:0
-nolisten tcp vt7 -novtswitch
Carlchristian Eckert
2017-01-20 15:12:41 UTC
Reply
Permalink
Raw Message
Post by Andreas Innerlohninger
I noticed that too since some time. System doesn't hibernate and xorg
causes 100% cpu load.
Process with 100% cpu load was
/usr/lib/xorg-server/Xorg :0 -seat seat0 -auth /run/lightdm/root:0
-nolisten tcp vt7 -novtswitch
I have the same problem since several months. Suspend to disk causes a
blank screen (as supposed to happen) and then freezes the system with
100% cpu load. However, no useful log messages could be found, so I
dropped the issue.

Happens on my desktop machine, no matter how I initiate the hibernation.
(pm-hibernate, systemctl hibernate, echo "disk" > /sys/power/state)
Andreas Innerlohninger
2017-01-22 13:55:43 UTC
Reply
Permalink
Raw Message
In my case it had to do something with lightDM/light-locker. I solved
it by switching to gdm.
I have the same problem since several months. Suspend to disk causes a blank
screen (as supposed to happen) and then freezes the system with 100% cpu
load. However, no useful log messages could be found, so I dropped the
issue.
Happens on my desktop machine, no matter how I initiate the hibernation.
(pm-hibernate, systemctl hibernate, echo "disk" > /sys/power/state)
Carlchristian Eckert
2017-02-14 08:50:30 UTC
Reply
Permalink
Raw Message
Two days ago, I started looking into this again. After some adjustments,
hibernation works for now. Dropping the solution here in case it helps.
The likely culprit seems to be a RAM upgrade from 16 to 32 GiB. During
this process, I forgot to adjust my SWAP partition, which was also just
16GB in size.

I use ZFS, which fills a large portion of the RAM with L2ARC data, so
that might have been just too big for the swap partition.

TL;DR the size of the swap partition was too small.
Post by Andreas Innerlohninger
In my case it had to do something with lightDM/light-locker. I solved
it by switching to gdm.
I have the same problem since several months. Suspend to disk causes a blank
screen (as supposed to happen) and then freezes the system with 100% cpu
load. However, no useful log messages could be found, so I dropped the
issue.
Happens on my desktop machine, no matter how I initiate the hibernation.
(pm-hibernate, systemctl hibernate, echo "disk" > /sys/power/state)
Bruno Pagani
2016-12-31 12:42:53 UTC
Reply
Permalink
Raw Message
Hi,
Post by Peter Nabbefeld
Hello,
since some time, hibernation does not work, dmesg contains following
[ 738.404244] ACPI Warning: \_SB.PCI0.PEG0.PEGP._DSM: Argument #4 type
mismatch - Found [Buffer], ACPI requires [Package]
(20160422/nsarguments-95)
[ 738.404353] ACPI: \_SB_.PCI0.PEG0.PEGP: failed to evaluate _DSM
[ 738.404356] ACPI Warning: \_SB.PCI0.PEG0.PEGP._DSM: Argument #4 type
mismatch - Found [Buffer], ACPI requires [Package]
(20160422/nsarguments-95)
Some time ago, when I closed my laptop, Linux automatically hibernated,
now it does not. As my laptop becomes really warm and I don't want to
stress my display, I'm using this rarely, so I cannot tell, since when
hibernation doesn't work :-/
Kind regartds
Peter
These messages are related to Optimus system (Nvidia implementation of dGPU power management handling is not compliant with ACPI specs). So that’s not related to your hibernation issue.

Regarding this one, I’d like a precision that might be useful: is this about hibernation or suspend?

Other than that, what DE do you use if applicable? What does handle lid close event? I had the issue that the lid close event wasn’t handled because KDE PM daemon isn’t inhibiting systemd one correctly, and once fixed it worked again.

Bruno
Peter Nabbefeld
2017-01-02 05:01:46 UTC
Reply
Permalink
Raw Message
Post by Bruno Pagani
Hi,
Post by Peter Nabbefeld
Hello,
since some time, hibernation does not work, dmesg contains following
[ 738.404244] ACPI Warning: \_SB.PCI0.PEG0.PEGP._DSM: Argument #4 type
mismatch - Found [Buffer], ACPI requires [Package]
(20160422/nsarguments-95)
[ 738.404353] ACPI: \_SB_.PCI0.PEG0.PEGP: failed to evaluate _DSM
[ 738.404356] ACPI Warning: \_SB.PCI0.PEG0.PEGP._DSM: Argument #4 type
mismatch - Found [Buffer], ACPI requires [Package]
(20160422/nsarguments-95)
Some time ago, when I closed my laptop, Linux automatically hibernated,
now it does not. As my laptop becomes really warm and I don't want to
stress my display, I'm using this rarely, so I cannot tell, since when
hibernation doesn't work :-/
Kind regartds
Peter
These messages are related to Optimus system (Nvidia implementation of dGPU power management handling is not compliant with ACPI specs). So that’s not related to your hibernation issue.
Hm, I'm using the Nouveau driver - is this message hardware-related?
Post by Bruno Pagani
Regarding this one, I’d like a precision that might be useful: is this about hibernation or suspend?
Hm, I'm usually not distinguishing those. As System doesn't need much
time to wake-up, I'd assume it just suspends to RAM - but as computers
are getting more power, this is probably not really visible. How can I
detect, which mode it had to switch to?
Post by Bruno Pagani
Other than that, what DE do you use if applicable? What does handle lid close event? I had the issue that the lid close event wasn’t handled because KDE PM daemon isn’t inhibiting systemd one correctly, and once fixed it worked again.
Using XFCE. Don't know, what handles the lid close evenet - how can I
find that out?

Kind regards
Peter
Bruno Pagani
2017-01-02 14:52:30 UTC
Reply
Permalink
Raw Message
Post by Peter Nabbefeld
Hi,
Le 31 décembre 2016 07:38:05 GMT+01:00, Peter Nabbefeld
Post by Peter Nabbefeld
Hello,
since some time, hibernation does not work, dmesg contains following
[ 738.404244] ACPI Warning: \_SB.PCI0.PEG0.PEGP._DSM: Argument #4 type
mismatch - Found [Buffer], ACPI requires [Package]
(20160422/nsarguments-95)
[ 738.404353] ACPI: \_SB_.PCI0.PEG0.PEGP: failed to evaluate _DSM
[ 738.404356] ACPI Warning: \_SB.PCI0.PEG0.PEGP._DSM: Argument #4 type
mismatch - Found [Buffer], ACPI requires [Package]
(20160422/nsarguments-95)
Some time ago, when I closed my laptop, Linux automatically hibernated,
now it does not. As my laptop becomes really warm and I don't want to
stress my display, I'm using this rarely, so I cannot tell, since when
hibernation doesn't work :-/
Kind regartds
Peter
These messages are related to Optimus system (Nvidia implementation
of dGPU power management handling is not compliant with ACPI specs).
So that’s not related to your hibernation issue.
Hm, I'm using the Nouveau driver - is this message hardware-related?
Yes, it’s hardware related, and that’s nothing to worry about, there is
nothing to be done here.
Post by Peter Nabbefeld
Regarding this one, I’d like a precision that might be useful: is
this about hibernation or suspend?
Hm, I'm usually not distinguishing those. As System doesn't need much
time to wake-up, I'd assume it just suspends to RAM - but as computers
are getting more power, this is probably not really visible. How can I
detect, which mode it had to switch to?
You can see it in dmesg for instance:
PM: Preparing system for sleep (mem)
PM: Suspending system (mem)

I suppose mem is replaced with disk or something like that in case of
hibernation. But looking at the wiki[0], if you don’t know it’s unlikely
to be hibernation given the steps involved to make this one work.
Post by Peter Nabbefeld
Other than that, what DE do you use if applicable? What does handle
lid close event? I had the issue that the lid close event wasn’t
handled because KDE PM daemon isn’t inhibiting systemd one correctly,
and once fixed it worked again.
Using XFCE. Don't know, what handles the lid close evenet - how can I
find that out?
Not sure how to know for sure, but see the wiki[1]. Apparently, XFCE
should properly inhibit systemd action. But this article says KDE does
too, however that’s not true for LidSwitch (at least it wasn’t last time
I tried). So you should check that, and also what settings you have in
XFCE power manager. One thing that might have happened is XFCE now
properly inhibiting systemd and not going to sleep on LidSwitch while
systemd was.

Hope it might help,
Bruno

[0]
https://wiki.archlinux.org/index.php/Power_management/Suspend_and_hibernate#Hibernation
[1] https://wiki.archlinux.org/index.php/Power_management#ACPI_events
Neven Sajko via arch-general
2017-01-02 21:32:29 UTC
Reply
Permalink
Raw Message
Regarding the difference between hibernation (suspension to disk)
and suspension to RAM:
Suspension to RAM just means turning off peripherals like disk and
display while keeping the RAM on. Thus from your perspective you
should just see the screen go black (and hear the disk spin down,
if you use a HDD).

Hibernation, on the other hand, is saving the system state to disk
and powering off the device, so you should be able to see some
messages while powering down, and again during boot up you could
see bootloader, initramfs, and kernel messages. There should be a
message about the Resume mkinitcpio hook being activated, although
it may be gone too quickly for you too see it.

Happy New Year 2017!
Tim Ye via arch-general
2017-01-05 14:56:15 UTC
Reply
Permalink
Raw Message
I read a web page sometime ago says these would work:

- put your computer into "sleep" (to RAM):
$ sudo su -c 'echo "mem" > /sys/power/state'
- put your computer into "hibernate" (to disk):
$ sudo su -c 'echo "disk" > /sys/power/state'

I just tried both, but "hibernate" doesn't work on my Archlinux but
works on Debian (with an older kernel) on my laptop.

- Tim
Post by Peter Nabbefeld
Hello,
since some time, hibernation does not work, dmesg contains following
[ 738.404244] ACPI Warning: \_SB.PCI0.PEG0.PEGP._DSM: Argument #4
type mismatch - Found [Buffer], ACPI requires [Package]
(20160422/nsarguments-95)
[ 738.404353] ACPI: \_SB_.PCI0.PEG0.PEGP: failed to evaluate _DSM
[ 738.404356] ACPI Warning: \_SB.PCI0.PEG0.PEGP._DSM: Argument #4
type mismatch - Found [Buffer], ACPI requires [Package]
(20160422/nsarguments-95)
Some time ago, when I closed my laptop, Linux automatically
hibernated, now it does not. As my laptop becomes really warm and I
don't want to stress my display, I'm using this rarely, so I cannot
tell, since when hibernation doesn't work :-/
Kind regartds
Peter
ber532k
2017-01-22 11:16:33 UTC
Reply
Permalink
Raw Message
I had the same issue with my laptop some days ago. At the moment it
seems to be working most times though, no idea why. I only made some
seemingly unrelated changes in BIOS settings (fan speed, battery charging).

The screen went blank, came back a second (as it usually does when about
to hibernate) and went blank again. After that the laptop would still
not turn off until the battery was fully drained. If I turned it off
manually after the screen went blank the second time, however, it would
later still resume all right. I guess it's something about triggering
the correct acpi-event for poweroff or something.
Loading...