I hope this is the appropriate mailing list...

The top three wakeup sources on my X61 Tablet account for 85% of the 
wakeups! And the two top ones could be avoided IMHO (but I'm not a 
kernel hacker, so what would I know...).

~40%              X : fbcon_add_cursor_timer (cursor_timer_handler)
~30%  <kernel core> : usb_hcd_poll_rh_status (rh_timer_func)
~15%        xfsbufd : schedule_timeout (process_timeout)

While powertop was running, there was no external usb device attached to 
the laptop and X wasn't running. That's why I think that the top two 
sources could be avoided. As you can see I'm using XFS, and while it may 
not be the most power-effective filesystem, 15% is a whole lot if the 
system is completely idle (swap was disabled).
The other wakeup sources contribute each <5%, but contain for example 
the eth0 interrupt even though no cable is plugged in (at 4%), the e1000 
watchdog (at 4%, too, does the watchdog need to kick in that often?), 
and with <1% each: some other <kernel core/IPI> stuff (page writeback etc).

Any suggestions who to contact in order to improve the affected modules 
to be more power-friendly?

thanks
tom

_______________________________________________
Discuss mailing list
[email protected]
http://mail.lesswatts.org/mailman/listinfo/discuss

Reply via email to