Unplugging the AC adapter causes the computer to suspend

[Niran Babalola]

This doesn't occur every time, and only started happening a few days ago.

[Daniel Silverstone (dsilvers)]

Can you please confirm what version of gnome-power-manager you have installed

Also if it occurs on a fresh boot, or only after a suspend via lid close

Also can you please confirm your settings in terms of actions, so lid-close-on-ac, lid-close-on-battery, sleep-button-action etc.

Thanks

[Richard Hughes (richard-hughes)]

Daniel, you might want to grab the poke-gnome-screensaver-on-hardware-event from CVS, and see if that fixes things. See http://bugzilla.gnome.org/show_bug.cgi?id=333525

[Niran Babalola (niran)]

I have version 2.13.93-0ubuntu1 installed.

The bug has never occurred on a fresh boot, but it doesn't occur on every subsequent suspend either. For instance, today I went about two suspends without it happening. Over the past few days, I have also run into bugs #32143 and #33991.

lid-close-on-ac: Blank screen
lid-close-on-batt: Suspend
critical battery: Shutdown
sleep-type-when-inactive: Do nothing
sleep-button-action: Suspend

[Paul Sladen (sladen)]

Is this something like:

  On unplug from AC, battery-level goes to "unknown" which is represented by '0'.

  Zero battery level is less than $crtical_level, therefore g-p-m performs $action

[Richard Hughes (richard-hughes)]

So telling g-p-m not to do the critical action at 0% would fix the problem?

[Paul Sladen (sladen)]

Or, if it is zero, take the result with a pinch of salt and try a minute later.

[Niran Babalola (niran)]

My critical action is set to shutdown, so if this was the cause of the bug, wouldn't it be shutting down rather than suspending?

I hope the suggested change somehow fixes it anyway.

[Paul Sladen (sladen)]

Funky... Can you try:

  sudo tail -0f /var/log/acpid

when you unplug the AC if it is still reproducible.

[Niran Babalola (niran)]

This was the second time I unplugged the power and it suspended. Output follows.

Laptop mode enabled, not active.
[Fri Mar 24 16:32:32 2006] received event "ac_adapter ADP1 00000080 00000000"
[Fri Mar 24 16:32:32 2006] notifying client 3947[111:111]
[Fri Mar 24 16:32:32 2006] executing action "/etc/acpi/power.sh"
[Fri Mar 24 16:32:32 2006] BEGIN HANDLER MESSAGES
[Fri Mar 24 16:32:32 2006] END HANDLER MESSAGES
[Fri Mar 24 16:32:32 2006] action exited with status 0
[Fri Mar 24 16:32:32 2006] completed event "ac_adapter ADP1 00000080 00000000"
[Fri Mar 24 16:32:32 2006] received event "battery BAT0 00000080 00000001"
[Fri Mar 24 16:32:32 2006] notifying client 3947[111:111]
[Fri Mar 24 16:32:32 2006] executing action "/etc/acpi/power.sh"
[Fri Mar 24 16:32:32 2006] BEGIN HANDLER MESSAGES
[Fri Mar 24 16:32:32 2006] END HANDLER MESSAGES
[Fri Mar 24 16:32:32 2006] action exited with status 0
[Fri Mar 24 16:32:32 2006] completed event "battery BAT0 00000080 00000001"
yes
Laptop mode enabled, active.
[Fri Mar 24 16:32:51 2006] received event "button/power PWRF 00000080 00000001"
[Fri Mar 24 16:32:51 2006] notifying client 3947[111:111]
[Fri Mar 24 16:32:51 2006] executing action "/etc/acpi/powerbtn.sh"
[Fri Mar 24 16:32:51 2006] BEGIN HANDLER MESSAGES
[Fri Mar 24 16:32:51 2006] END HANDLER MESSAGES
[Fri Mar 24 16:32:51 2006] action exited with status 0
[Fri Mar 24 16:32:51 2006] completed event "button/power PWRF 00000080 00000001"[Fri Mar 24 16:32:52 2006] received event "battery BAT0 00000081 00000001"
[Fri Mar 24 16:32:52 2006] notifying client 3947[111:111]
[Fri Mar 24 16:32:52 2006] executing action "/etc/acpi/power.sh"
[Fri Mar 24 16:32:52 2006] BEGIN HANDLER MESSAGES
[Fri Mar 24 16:32:52 2006] END HANDLER MESSAGES
[Fri Mar 24 16:32:52 2006] action exited with status 0
[Fri Mar 24 16:32:52 2006] completed event "battery BAT0 00000081 00000001"

[Richard Hughes (richard-hughes)]

Can you try CVS HEAD for me please? It now ignores values == 0.

[Paul Sladen (sladen)]

The useful stuff out of that log is:

16:32:32 received event "ac_adapter ADP1 00000080 00000000"
16:32:32 received event "battery BAT0 00000080 00000001"
16:32:51 received event "button/power PWRF 00000080 00000001"
16:32:52 received event "battery BAT0 00000081 00000001"

Do you remember pressing the PowerButton on the laptop? There's an event coming in, which is the laptop sending a power-button event; 'hald-addon-acpi' then punts this across the D-Bus, which g-p-m then presumbly catches and starts the suspend from.

So... something is making the laptop send a spurious PowerButton press. (Or was that you pressing "POWER" to wake the laptop up again after it had gone to sleep?)

[Niran Babalola (niran)]

That was me pressing POWER to wake the laptop up after suspending.

[Matt Zimmerman (mdz)]

Are the relevant changes from CVS now in Dapper? Are you still having the problem?

[Niran Babalola (niran)]

I was about to reply saying that I hadn't noticed it recently, then I unplugged the cable and my computer suspended. I have the latest package (2.14.0-1) installed, but I haven't restarted in almost three days, and g-p-m's about window says 2.13.93. I'll reboot soon and see if the bug still persists.

[Richard Hughes (richard-hughes)]

I've fixed the hal side of the problem:

See the first entry here: http://live.gnome.org/GnomePowerManager/Faq

[Daniel Silverstone (dsilvers)]

gnome-power-manager (2.14.3-0ubuntu2) dapper; urgency=low

  * Patches added in this version:
    - 20-enable-xfce-startup.patch
      Enable XFCE in the .desktop file so that it starts up for xubuntu too.
      Closes: launchpad #43077
    - 30-transparent-notification-icon.patch
      Enable the eggtrayicon stuff to do transparency.
      Closes: launchpad #40446
    - 95-lid-state-tracking.patch
      Augment lid state tracking to use hal if hal reports it is capable
      of tracking the lid state.
      If hal reports an acpi_LID with button.has_state == true then we use
      that state information whenever we want to consider the lid switch
      This should deal with LP #33072 and thus its duplicates #42988,
      #42823, #40730, #40662, #37442, #36459 and #33952
    - 96-disable-session-save-on-shutdown.patch
      Disable the saving of the session during critical power shutdown.
      This confuses too many people for now.
      Closes: launchpad #35691
  * Patches changed in this version:
    - 40-ubuntu-schema-defaults.patch
      Alter screen lock defaults to always lock, regardless of screensaver.
      Closes: launchpad #39448
  * Related fixes: #35591 -- We believe this is caused by the hal and
    gnome-power-manager interaction wrt. lid status. The new hal and the
    2.14.3 gnome-power-manager should fix it all.