[linux-audio-user] APIC is bad?

tim hall tech at glastonburymusic.org.uk
Sat Jul 17 12:59:40 EDT 2004


Last Saturday 17 July 2004 12:33, Malcolm Baldridge was like:
> > I have my soundcard (an onboard i8x0) sharing an interrupt (IRQ 11) with
> > eth0 and usb-uhci. Would this be likely to give rise to xruns in a
> > similar way?
>
> I think motherboard audio bugs/latency issues will be more of a problem
> than shared IRQs.  Turning off the PnP Setting in the BIOS and then "Reset
> Configuration Data" may cause a re-assignment to occur during the next time
> you go through BIOS POST.

It's worth a try.

> Moving the ethernet card should get you another IRQ, though keep in mind,
> it's a bit stranger with PCI than ISA.  To make it even spicier, some PCI
> slots are not capable of bus-mastering DMA.  But the answer to your
> question is yes: moving the card will get you a different IRQ.

Great, thanks.

> As for the onboard-USB, well, that might be harder to "move".  The problem
> is that the IRQs are "mapped" to PCI INT-levels, and it seems that many
> system hardware designers get very lazy and slopping with how they use
> them.

Last Saturday 17 July 2004 13:26, Jan Depner was like:
>         In addition to the time to determine interrupt source there is
> another problem with shared interrupts - don't use any USB devices or run
> your network while trying to record.

My suspicions were correct then, I don't use USB at the moment, so I'll see if 
I can turn it off in the BIOS.

Back to Malcolm:
> Shared IRQs have been with us for a few years now, and I doubt it's the
> source of most xruns people see on their systems these days.  We are
> talking about microseconds of additional time to determine the interrupt
> source here.  If your xruns are in the hundreds of milliseconds, this is
> not your problem.  If you're on the borderlines, THEN it might be something
> worth looking into.

Actually, I think I'm doing fairly well compared to similar systems, I _can_ 
run JAMin (just, but with good results), so I'm probably being a bit picky.

> This brings up a point about technical diagnostics and troubleshooting: try
> to assess the magnitude of your xruns to give you valuable hints as to
> where to look for what might be wrong.  If they're in the hundreds of
> milliseconds, your problem is most likely a
> big-kernel-lock/preemptible-kernel related fault, or just really bad
> hardware.  File system choices affect this as well.

I don't think I have anything majorly wrong, I'm just checking if there is 
anything I've overlooked as I still think I could improve on my system's 
current performance. Thanks for the advice, I think this could be quite a 
worthwhile exercise.

tim hall




More information about the Linux-audio-user mailing list