Hello list,
after upgrading to 2.6.32 kernel, I noticed my P3 500MHz laptop's fan
was working non-stop. Powertop showed more than 100K wakeups/s even though
cpu was idling almost 100% (no user nor system cpu usage). Further
investigation showed that this happened as soon as the "processor" module
was inserted, whence I also got a message like the following:
tsc unstable due to halts in idle
switching clocksource to acpi_pm
A workaround that currently works for me is to boot with "clocksource=pit"
parameter. FYI in 2.6.31 logs I had the same message about tsc
instability, and the final clocksource was acpi_pm too. Nevertheless
wakeups/s were much less and temperature was low.
I performed a bisection to find which commit to blame. Please look into
attached files which include dmesg output from good and bad case, and
bisection output.
What do you think?
Dimitris