calling cpufreq_set

calling cpufreq_set

[Stefan Buettcher]

Hi,

I am new to this list as well as cpufreq at all.

I have written a kernel module (2.4.19-something) on the
ipaq 5400. This module makes direct calls to cpufreq_set.
While the system does not break, the usb does. usbnet
suddenly stops working and will resist my efforts to work
afterwards.

When I do not do the calls from within the kernel but use
the proc file interface provided by cpufreq, the system is
stable and usb has no problems.

Has anyone experienced similar problems? Any idea why
this happens?

Stefan

Re: calling cpufreq_set

[Dominik Brodowski]

Hi Stefan,

On Wed, Nov 05, 2003 at 12:44:13AM +0100, Stefan Buettcher wrote:
> Hi,
> 
> I am new to this list as well as cpufreq at all.
> 
> I have written a kernel module (2.4.19-something) on the
> ipaq 5400. This module makes direct calls to cpufreq_set.
> While the system does not break, the usb does. usbnet
> suddenly stops working and will resist my efforts to work
> afterwards.
> 
> When I do not do the calls from within the kernel but use
> the proc file interface provided by cpufreq, the system is
> stable and usb has no problems.
> 
> Has anyone experienced similar problems? Any idea why
> this happens?

I'm not aware of the details of the ipaq cpufreq driver - it's neither in
the kernel sources, nor was it ever posted on this list... so it _might_ be
a driver- or platform-specific problem. Maybe some timing calculations are
wrong after calls to cpufreq_set, maybe there needs to be some delay between
subsequent calls to cpufreq_set, ...

Other than that, cpufreq_set() is deprecated in 2.6.

	Dominik