Florin Andrei wrote:
I rebuilt the latest Fedora 2 kernel update to enable
preempt and
IO-APIC.
I'm not yet sure about preempt, but IO-APIC has been acting weird.
It gave me more interrupts (up to 21 instead of 15), but the devices
were distributed suboptimal.
Without APIC, the nvidia module was alone on its own interrupt, the
EMU10K1 was alone, the ide and eth modules were on separate interrupts,
etc. Quite ok.
With IO-APIC, nvidia, EMU10K1 and bttv were on the same interrupt, ide2,
ide3 and eth0 were on the same interrupt.
Instead of messing with the kernel again, i just rebooted with the
"noapic" parameter and now the interrupts are looking good again.
BTW, anyone has any measurements on how bad it is to put essential
devices on the same interrupt? (in terms of xruns)
Please note - The 'optimal' numerical assignment of interrupts with
using the APIC model has nothing to do with the older, non-APIC, order.
Please do not confuse the idea that 'interrupt #9 is best' with the
numbers assigned on an APIC system. These are completely different models.
Personally I have done no optimization work on APIC systems and cannot
tell you what would be best. However, TTBOMK, there is no simple way
today to optimize APIC interrupts in a Linux system anyway, so as far as
I can tell you basically take what you get. If you get too many xruns
then I guess you go back to non-APIC mode.