I'm going back to 2.0.14 with the irqtune utility, and without the irqtune
patch, that looked okay at first, but it definitely has some conflict with
other interrupts because its performance is directly related to mouse activity
on another port. And no, there is _no_ hardware conflicts, and I have disabled
all the APM stuff both in the kernel and in CMOS.
When these issues have been addressed, I will jump on the next release patch :-)
---russ