Bug#503029: problem recognized
On Sat, Oct 25, 2008 at 03:29:38PM +0200, Radek Warowny wrote:
> I have recognized the problem I think. The issue occurs when you have
> cpu frequency scaling function activated. It appears when CPU was set
> to different frequency while inserting analog driver and different
> while reading joystick data (for example running a game).
> I think the problem is located in analog kernel driver and is related
> to usage of rdtscl() function to read time (in GET_TIME macro). This
> function returns number of clock ticks, not real time. So when cpu
> frequency changes the same difference (delta) of two reads is not the
> same difference of time (one clock lasts different time period). When
> the analog driver is loaded it computes some timing data that is used
> further while reading joystick state. When CPU frequency change
> occures it compares that timing data values with new timing reads
> which are differently scaled.
> A workaround (until the issue will be fixed) is to unload analog
> driver, set you CPU do fixed frequency, load the driver again an run
> you game.
There have been no related changes to the analog joystick driver up
to 2.6.28. Could you report this at http://bugzilla.kernel.org and
send the bug number to this bug?