DPMS Suspend doesn't work unless lid is open on laptop.

[Troy R.]

Binary package hint: gnome-power-manager

Expected:
GPM should allow my external monitor to enter DPMS suspend while the laptop lid is closed.

Observed:
GPM doesn't allow my external monitor to suspend while the lid is closed on the laptop. If the lid is open, it works fine.

Steps To Reproduce:
Restart Ubuntu with the laptop lid closed, log in and wait for the timeout period. The external monitor shows the screensaver, but does not enter DPMS suspend.

No combination of GConf options has allowed me to effectively work around this issue. In another bug, someone wrote a perl script that works, but only if the screensaver is enabled. I use the machine with the screensaver both on and off. So, I started digging.

Here's some snipped GPM logs (gnome-power-manager --verbose):

(Lid down, 00:01 timeout for screensaver and dpms blanking, does not work as expected)
8<--- snip --->8
304-TI:22:50:53 TH:0x9a788b8 FI:gpm-session.c FN:gpm_session_presence_status_changed_cb,145
305- - emitting idle-changed : (1)
306-TI:22:50:53 TH:0x9a788b8 FI:gpm-idle.c FN:gpm_idle_session_idle_changed_cb,354
307- - Received gnome session idle changed: 1
308-TI:22:50:53 TH:0x9a788b8 FI:gpm-idle.c FN:gpm_idle_evaluate,187
309- - session_idle=1, idle_inhibited=0, suspend_inhibited=0, x_idle=1
310-TI:22:51:03 TH:0x9a788b8 FI:gpm-idle.c FN:gpm_idle_set_mode,108
311- - Doing a state transition: blank
312-TI:22:51:28 TH:0x9a788b8 FI:gpm-idle.c FN:gpm_idle_idletime_reset_cb,391
313- - idletime reset
314-TI:22:51:28 TH:0x9a788b8 FI:gpm-idle.c FN:gpm_idle_evaluate,187
315- - session_idle=1, idle_inhibited=0, suspend_inhibited=0, x_idle=0
316-TI:22:51:28 TH:0x9a788b8 FI:gpm-idle.c FN:gpm_idle_set_mode,108
317- - Doing a state transition: normal
318-TI:22:51:28 TH:0x9a788b8 FI:gpm-manager.c FN:gpm_manager_idle_changed_cb,804
319: - lid is closed, so we are ignoring ->NORMAL state changes
320-TI:22:51:28 TH:0x9a788b8 FI:gpm-idle.c FN:gpm_idle_evaluate,192
321- - X not idle
8<--- snip --->8

(Lid up, 00:01 timeout for screensaver and dpms blanking, works as expected)
8<--- snip --->8
TI:23:04:23 TH:0x88638b8 FI:gpm-session.c FN:gpm_session_presence_status_changed_cb,145
 - emitting idle-changed : (1)
TI:23:04:23 TH:0x88638b8 FI:gpm-idle.c FN:gpm_idle_session_idle_changed_cb,354
 - Received gnome session idle changed: 1
TI:23:04:23 TH:0x88638b8 FI:gpm-idle.c FN:gpm_idle_evaluate,187
 - session_idle=1, idle_inhibited=0, suspend_inhibited=0, x_idle=1
TI:23:04:33 TH:0x88638b8 FI:gpm-idle.c FN:gpm_idle_set_mode,108
 - Doing a state transition: blank
TI:23:04:41 TH:0x88638b8 FI:gpm-idle.c FN:gpm_idle_idletime_reset_cb,391
 - idletime reset
TI:23:04:41 TH:0x88638b8 FI:gpm-idle.c FN:gpm_idle_evaluate,187
 - session_idle=1, idle_inhibited=0, suspend_inhibited=0, x_idle=0
TI:23:04:41 TH:0x88638b8 FI:gpm-idle.c FN:gpm_idle_set_mode,108
 - Doing a state transition: normal
TI:23:04:41 TH:0x88638b8 FI:gpm-idle.c FN:gpm_idle_evaluate,192
 - X not idle
8<--- snip --->8

I also went digging in gpm-manager.c from GPM to figure out what that 'lid is closed...' stuff meant:

8<--- snip --->8
 /* Ignore back-to-NORMAL events when the lid is closed, as the DPMS is
  * already off, and we don't want to re-enable the screen when the user
  * moves the mouse on systems that do not support hardware blanking. */
 if (gpm_button_is_lid_closed (manager->priv->button) &&
     mode == GPM_IDLE_MODE_NORMAL) {
  egg_debug ("lid is closed, so we are ignoring ->NORMAL state changes");
  return;
 }
8<--- snip --->8

... but I haven't been able to relate it directly to my problem, because I want it to blank in the first place. It was the only big difference I could spot in much of the logging.

Thanks!

ProblemType: Bug
DistroRelease: Ubuntu 10.10
Package: gnome-power-manager 2.32.0-0ubuntu1
ProcVersionSignature: Ubuntu 2.6.35-24.42-generic 2.6.35.8
Uname: Linux 2.6.35-24-generic i686
NonfreeKernelModules: wl
Architecture: i386
Date: Thu Dec 23 23:14:00 2010
GnomeSessionIdleInhibited: No
GnomeSessionInhibitors: None
GnomeSessionSuspendInhibited: No
InstallationMedia: Ubuntu 10.10 "Maverick Meerkat" - Release i386 (20101007)
MachineType: Dell Inc. Vostro 1000
ProcCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.35-24-generic root=UUID=3f91319e-e82a-4ca6-9a44-9fc44213128d ro quiet splash
ProcEnviron:
 LANG=en_CA.utf8
 SHELL=/bin/bash
SourcePackage: gnome-power-manager
dmi.bios.date: 07/30/2006
dmi.bios.vendor: Dell Inc.
dmi.bios.version: 2.5.2
dmi.board.name: 0WY383
dmi.board.vendor: Dell Inc.
dmi.chassis.type: 8
dmi.chassis.vendor: Dell Inc.
dmi.chassis.version: Not Specified
dmi.modalias: dmi:bvnDellInc.:bvr2.5.2:bd07/30/2006:svnDellInc.:pnVostro1000:pvrNotSpecified:rvnDellInc.:rn0WY383:rvr:cvnDellInc.:ct8:cvrNotSpecified:
dmi.product.name: Vostro 1000
dmi.product.version: Not Specified
dmi.sys.vendor: Dell Inc.

[Tom Star (cstamas77)]

I think that considering how many people use laptop these days, this should be a major issue. Seriously, these kind of bugs affect user experience a lot. I am aware that security and stability problems are of highest priority, but how can anyone expect Linux to become more widespread on desktops if more than a year is not enough to even make a decision whether an issue like this will be fixed or not.

Anyway, if someone needs a workaround, then in this forum (https://bugs.launchpad.net/ubuntu/+source/gnome-power-manager/+bug/550054) search for 'nightflow' and follow his/her instructions.