Strange behaviour

[Felix Oxley]

 My test is now as follwos:

1. reboot (on AC power)
2. do not log in to X
3. On VT1 do the following:
4. check powertop = >5000 wakeups
5. init 3
6. check powertop = >5000 wakeups
7. init 1
8. check powertop = >5000 wakeups
9. start killing processes and removing kernel modules
10. kill all wifi modules
11. kill all bluetooth modules
12. kill firewire modules
13. check powertop = >5000 wakeups
14. kill all snd modules

powertop says:
wakeups: 6075
90% (8.0) <kernel module> : usb_hcd_poll_rh_status (rh_timer_func)

As far as I can tell this is to do with the iwl4965 module.
If I boot with it disabled then I never get to 5000 wakeups.
However once it has been enabled _and_ connected to an AP then no matter
what I do the wakeups do not reduce.

Reminder: System is Fedora 9, kernel 2.6.25.

Any suggestions?

Also, what do the numbers in brackets in powertop (after the %ages)
indicate?

Please CC me   :-)

Thanks,
Felix


2008/6/1 Felix Oxley <felix at oxley.org>:

> Hi Magnus, thanks for your reply.
>
> The behaviour of my machine has changed since yesterday!  :-(
> I now get this high wake up load without starting FireFox.
> Immediatley after boot the machine will show <150 wakeups per second.
> Then it will start to show much more.
>
> My problem is that the output of powertop is not giving me any information
> about where these wakeups are coming from.
> The output of powertop -d is below.
> Can any one tell me which process is causing 5000 wakeups?
>
> [root at fedora ~]# powertop -d
> PowerTOP 1.9    (C) 2007 Intel Corporation
>
> Collecting data for 15 seconds
> Cn              Avg residency
> C0 (cpu running)        (19.2%)
> C1          0.0ms ( 0.0%)
> C2          0.0ms ( 0.0%)
> C3          0.1ms (80.8%)
> P-states (frequencies)
>   2.51 Ghz     3.9%
>   2.50 Ghz     0.0%
>   2.00 Ghz     0.0%
>    800 Mhz    96.1%
> Wakeups-from-idle per second : 5860.5    interval: 15.0s
> Power usage (ACPI estimate): 18.3W (2.3 hours)
> Top causes for wakeups:
>   60.3% ( 50.5)       <interrupt> : extra timer interrupt
>   19.4% ( 16.3)      <kernel IPI> : Rescheduling interrupts
>   15.3% ( 12.8)       <interrupt> : ata_piix, eth0, iwl4965
>    1.4% (  1.2)   <kernel module> : usb_hcd_poll_rh_status (rh_timer_func)
>    0.7% (  0.6)            iscsid : schedule_timeout (process_timeout)
>    0.4% (  0.3)   multiload-apple : schedule_timeout (process_timeout)
>    0.4% (  0.3)    gnome-terminal : schedule_timeout (process_timeout)
>    0.2% (  0.2)    im-info-daemon : do_nanosleep (hrtimer_wakeup)
>    0.2% (  0.1)    NetworkManager : tg3_open (tg3_timer)
>    0.2% (  0.1)   netspeed_applet : schedule_timeout (process_timeout)
>    0.2% (  0.1)            iscsid : do_nanosleep (hrtimer_wakeup)
>    0.2% (  0.1)   setroubleshootd : schedule_timeout (process_timeout)
>    0.2% (  0.1)    NetworkManager : schedule_timeout (process_timeout)
>    0.2% (  0.1)              Xorg : schedule_timeout (process_timeout)
>    0.1% (  0.1)       <interrupt> : PS/2 keyboard/mouse/touchpad
>    0.1% (  0.1)        kerneloops : schedule_timeout (process_timeout)
>    0.1% (  0.1)      clock-applet : schedule_timeout (process_timeout)
>    0.1% (  0.1)             mount : start_this_handle (commit_timeout)
>    0.1% (  0.1)     <kernel core> : neigh_table_init_no_netlink
> (neigh_periodic_timer)
>    0.1% (  0.1)        pulseaudio : schedule_timeout (process_timeout)
>    0.1% (  0.1)           iwl4965 : ieee80211_sta_work
> (ieee80211_sta_timer)
>    0.1% (  0.1)       gnome-panel : schedule_timeout (process_timeout)
>    0.1% (  0.1)   <kernel module> : neigh_table_init_no_netlink
> (neigh_periodic_timer)
>    0.1% (  0.1)          libvirtd : br_stp_enable_bridge
> (br_hello_timer_expired)
>    0.1% (  0.1)   <kernel module> : sta_info_start (sta_info_cleanup)
>    0.1% (  0.1)               top : schedule_timeout (process_timeout)
> [root at fedora ~]#
>
> Thanks
> Felix
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.bughost.org/pipermail/power/attachments/20080601/f3b06e73/attachment.html