Re: 2.6.25 regression: powertop says 120K wakeups/sec

[Daniel Klaffenbach]

With the stable 2.6.25 release I still get a _massive_ amount of wakeups 
(around 50k). Also the power consumption is a lot higher than with 2.6.24.5. 
I am using a Mobile AMD Sempron (32bit) on an nx6125 laptop (Gentoo).
In 2.6.25 an "extra timer interrupt" appeared, which was not there in previous 
kernel versions.

Power usage:
2.6.24.5: ~12-15 W
2.6.25: ~16-22 W

Wakeups:
2.6.24.5: 10-300
2.6.25: 900-50000!!


I thought the patch posted went into the stable kernel, but it does not seem 
to do anything for me. I have also tried 2.6.25-git3, but no changes.

Here are some links to potentially useful files (idle, no WLAN):
dmesg-2.6.25: http://pastebin.ca/995545
powertop 2.6.25: http://pastebin.ca/995548
dmesg-2.6.24.5: http://pastebin.ca/995549
powertop 2.6.24.5: http://pastebin.ca/995550

If there is anything I can provide you with for debugging let me know.

- Daniel

Attachment: signature.asc
Description: This is a digitally signed message part.