Interrupt latency problems

[Andrew Gallatin]

Paul Richards writes:
 > 
 > As a further datapoint to my last email:
 > 
 >    20  ??  WL     5:16.38  (irq9: fxp0 acpi0)
 > 
 > I'll see if I can shift the interrupts around in the bios but I don't
 > think I can since fxp is onboard.

I think you might be able to reserve IRQ 9 for some legacy device or
something in the BIOS.

 > Perhaps acpi doesn't do anything to check whether the interrupt belongs
 > to it or not and so when an fxp interrupt arrives the acpi handler gets
 > called.

That's a good theory.  I think you might have something there.


Drew