lid shut laptop = high cpu load?

[Eric S. Johansson]

Zach wrote:
> Would be interesting to see complete top output, to see what process
> is hogging resources.  Does it do it every time you close the lid, or
> just this once?  If it's every time, then likely it's something that
> one of the acpi scripts is trying to do.

here is something I captured.

   116 root      15   0     0    0    0 D  0.7  0.0   4:55.65 kswapd0
14151 root      17   0  2132  788  632 R  0.1  0.2   0:02.91 top
  7692 esj       17   0 23816 1056  960 D  0.1  0.2   0:13.96 
gweather-applet
11774 esj       17   0 13816  760  720 D  0.1  0.1   2:52.15 emacs
  7690 esj       17   0 21564  716  692 D  0.1  0.1   0:44.38 clock-applet
  7725 esj       16   0 35984  912  852 D  0.1  0.2   1:04.60 gnome-terminal
14210 root      26  10  2044  512  448 D  0.1  0.1   0:00.71 updatedb
14321 root      18   0  3900  816  560 D  0.1  0.2   0:00.06 maxlifetime
  6894 hplip     17   0  8968   80   20 D  0.0  0.0   0:13.73 python
  9037 esj       17   0 53916  880  768 D  0.0  0.2   0:10.39 
gnome-cups-icon
14326 esj       18   0  3124  484  408 D  0.0  0.1   0:00.03 tail
  6842 root      16   0  137m   60   24 D  0.0  0.0 173:31.89 Xorg
  7612 esj       17   0 30184  816  760 D  0.0  0.2   0:07.73 nautilus
  7678 esj       16   0 17844  788  720 D  0.0  0.2   3:14.57 
gnome-netstatus
29833 esj       17   0  6440 1340 1236 D  0.0  0.3   1:16.61 sshd
29938 root      16   0 12512  828  764 D  0.0  0.2   0:29.92 emacs
  6331 hal       17   0  5060  168   84 D  0.0  0.0   0:37.39 hald

no process was hogging resources; memory or cpu.  Disk activity was 
through the roof and swap was rather high (kswapd at 4hrs run is a hint). 
kacpid had something like 47 hours of run time so that is a good 
candidate for the problem.


> Not sure if you did this already, but you may want to install ssh,
> then ssh into the box before before you close the lid.  Start up top
> and any other tools you want to use, then close the lid.  top has a
> way to write it's output to disk, though I'm not sure what it is right
> now

I was going to try that later when I could pay attention.

> You may want to pipe things through tee to get screen output as well
> as writing to disk.  That way if the box quits responding, you'll
> still have stuff written to disk for analysis after you reboot.

not a bad idea.

for grins, I closed the lid:

Nov  7 10:37:32 localhost kernel: [4302877.871000] read EC, IB not empty
Nov  7 10:37:32 localhost kernel: [4302877.921000] read EC, OB not full
Nov  7 10:37:32 localhost kernel: [4302877.921000]     ACPI-0423: *** 
Error: Handler for [EmbeddedControl] returned AE_TIME
Nov  7 10:37:32 localhost kernel: [4302877.921000]     ACPI-0508: *** 
Error: Method execution failed [\_GPE._L0B] (Node dffeed60), AE_TIME
Nov  7 10:37:32 localhost kernel: [4302877.921000]     ACPI-0563: *** 
Error: AE_TIME while evaluating method [_L0B] for GPE[ 0]

hmmm smells like a problem.

--- eric