[hardy] resume displays "Policy timeout is not valid. Please wait a few seconds and try again"

[Emil Sit]

Binary package hint: gnome-power-manager

On a Thinkpad X40, on resume, a tooltip is displayed indicating:
    Action Forbidden: Policy timeout is not valid. Please wait a few seconds and try again

This may possibly also have also been observed in bug #176544 but presumably not due to the same root cause as #189445. Some possibly relevant log messages through a suspend/resume cycle:

$ gnome-power-manager --no-daemon --verbose --debug
[gpm_debug_init] gpm-debug.c:236 (09:53:14): Verbose debugging enabled
[main] gpm-main.c:219 (09:53:14): GNOME Power Manager 2.22.1
[gpm_ac_adapter_init] gpm-ac-adapter.c:217 (09:53:14): using /org/freedesktop/Hal/devices/computer_power_supply_ac_adapter_AC
[gpm_control_init] gpm-control.c:657 (09:53:14): Using a supressed policy timeout of 5 seconds
...
Saving to syslog: Suspending computer. Reason: The suspend button has been pressed.[gpm_info_event_log] gpm-info.c:594 (09:53:28): Adding 5 to the event log
[gpm_control_get_lock_policy] gpm-control.c:389 (09:53:28): Using custom locking settings (1)
...
Saving to syslog: Resuming computer[gpm_info_event_log] gpm-info.c:594 (09:53:54): Adding 7 to the event log
...
[gpm_control_is_policy_timout_valid] gpm-control.c:116 (09:53:54): Skipping suppressed action
[gpm_notify_create] gpm-notify.c:126 (09:53:54): libnotify: Power Manager : Policy timeout is not valid. Please wait a few seconds and try again.

$ apt-cache policy gnome-power-manager
gnome-power-manager:
  Installed: 2.22.1-1ubuntu2
  Candidate: 2.22.1-1ubuntu2
  Version table:
 *** 2.22.1-1ubuntu2 0
        500 http://us.archive.ubuntu.com hardy/main Packages
        100 /var/lib/dpkg/status

[Gerv (gerv)]

I see exactly the same thing on my X40. (Rather unhelpful as an error message...) Let me know if I can help debug.

[Emil Sit (emilsit)]

It'd be useful to determine if this is X40 specific ... does anyone not on an X40 see this error? I don't know enough about gnome-power-manager to offer much other debugging suggestions.

[Ted Gould (ted)]

Could you readd the messages in the last "...."? It would be useful to know what actions it is surpressing.

[christooss (matic-ahacic)]

I have HP compaq 6720s and I get same problem. Suspend is taking rather long time. I don't quite remember how this was in Gutsy but I think it was much faster. Even in Hardy waking from suspend was quick

It takes about 10 sec.

[Gerv (gerv)]

Taking my cue from the original reporter, here is a full log of

gnome-power-manager --no-daemon --verbose --debug power > gpm.txt 2>&1

It contains two suspends. On the first, the error popped up just as the screen was fading out to suspend. On the second, it appeared when the desktop came back on the resume.

Let me know if there's more information you need.

Gerv

[Gerv (gerv)]

Having looked at the code in gnome-power-manager, the first error in my log file could be bogus. It looks like this error comes up if it receives two commands too quickly (inside the timeout value). The first suspend may have been initiated pretty quickly after the previous resume, which is why the message appeared on suspend rather than resume.

It seems odd that the error comes up telling you to try again, but then the suspend/resume succeeds anyway...

I'll let someone who knows what they are doing take it from here :-)

Gerv

[Ted Gould (ted)]

It looks to me from that log like we're getting two sleep button events for everytime you're pressing the sleep button. I'm unsure of why in one case they're both before the suspend and the second they're split across it. I'm reassigning this to hal as we shouldn't be getting those dual events (though it's good we're blocking them).

[Emil Sit (emilsit)]

Here's an unfiltered gpm log for the X40, with a single suspend event.

I have occasionally seen messages saying that suspend failed, even though the machine suspended, which seems to align with getting two sleep events.

[ZanoZik (zanozik)]

So in other words, the policy insists that there should be no two APM events in 5 seconds gap and blocks them?
If unplugging the AC adapter is considered as APM event, then the fact that I don't get a "lid closed" (or "button pressed" for that matter) suspend after I unplug the adapter - is not a bug, but rather a feature? :)

[dn (nobled)]

Oops. Didn't mean to press the "also needs a fix here" button for gnome-power-manager.

[lobner (soren-lobner)]

Just reporting I have Hardy on an X60 and experience the same bug whenever I resume from suspend, while having AC power connected.

[LOKe (lbharti)]

I have the same problem on X61.

[Chaskiel Grundman (cg2v)]

I don't actually use ubuntu, but based on what I'm seeing on debian lenny, I would guess that the extra event is coming from acpi-support, as it invokes 'acpi_fakekey $KEY_SLEEP' when it processes a sleep button acpi event and finds that gnome-power-manager or klaptopdaemon is running. (this will be in /etc/acpi/sleep.sh, /etc/acpi/sleepbtn.sh, or /usr/share/acpi-support/suspendorhibernate, depending on the version of the acpi-support package)

[lobner (soren-lobner)]

Same problem on my X60 with Hardy.

Seems quite a lot is having this problem - hope someone smart finds a way to fix it...

[Michael Gauthier (mike-silverorange)]

Same problem with a T43 on 8.04.1

[balak (balak)]

I see the same issue on an averatec 2100 series laptop.

uname:
Linux abc 2.6.24-19-generic #1 SMP Fri Jul 11 21:01:46 UTC 2008 x86_64 GNU/Linux

In my laptop, the first suspend/resume cycle happens fine. However, after 1 cycle, it is unable to resume from suspend or hibernate. I have to do a hard-reboot.

Also attaching the output of
$ gnome-power-manager --no-daemon --verbose --debug all
for what is is worth

[Michael Gauthier (mike-silverorange)]

balak,

Your problem could be a different bug. Suspend-resume always works for my T43, I just see the notification message when resuming.

[Thomas Novin (thomasn80)]

I get this problem from time to time aswell on my 8.04.1 system.

[Kasper J. (kasperj)]

Iæm affected by this too (aren't everybody?). I understand it could be a feature, but it's more annoying than helpful. Is it difficult fix or is just forgotten on the TO DO list? ;)

[Charlie Daly (cdaly)]

I noticed this message briefly appear (almost flicker - could just make out the text 'policy timeout') on my Toshiba Satelite pro (U200) running Hardy.

It's the first time I noticed it and it probably was because, I opened the lid and then pressed the power button. Opening the lid is normally enough to trigger resume, but I accidently pressed the power button which was probably the cause of the alert.

[Michael Shulman (shulman)]

I used to get this message frequently, but I haven't seen it appear in quite some time, certainly never with karmic. Maybe it's been fixed?

[dino99 (9d9)]

http://www.freedesktop.org/wiki/Software/hal/