During the full load when the system does not have opportunity
to sleep the wakeup counter counts too much (mostly more than 1000).
This issue originates from RH bugzilla:
https://bugzilla.redhat.com/show_bug.cgi?id=528945
The problem is that if no event occurs the timerstats are used, but
PowerTOP exits upon receiving SIGWINCH which is probably not correct
behaviour - it removes the possibility to resize the PowerTOP's window
if running inside the XTerm or such.
This issue originates from RH bugzilla:
https://bugzilla.redhat.com/show_bug.cgi?id=644800
The proposed patch is
On Tue, 14 Dec 2010 03:51:40 -0500 (EST)
Jaroslav Skarvada jskar...@redhat.com wrote:
PowerTOP should not recommend ondemand when p4-clockmod is used.
AFAIK the p4-clockmod is autoloaded for thermal purposes, but the
latency is too high for ondemand. There is RH bugzilla on it:
kernel/cpu/cpufreq/Kconfig says that p4-clockmod shouldn't be used at
all, SpeedStep should be used instead, which would be the
acpi-cpufreq
driver, right?
So maybe powertop should suggest using that (if CPU supports it)
instead of p4-clockmod?
Maybe, if the suggestion would be written
