Subject: Re: IRQ 19 with the IO-APIC? (also IPv6)

Trever Adams (arabian@onramp.net)
Thu, 05 Feb 1998 21:17:22 +0000

> On Thu, 5 Feb 1998, Bernd Wolf wrote:
>
> > [...] If giving the "pirq=0" the systems seems to be very GOOOOD! BUT:
> > my NE2000-clone gets the irq=18 and the scsi says something of irq=19 ?!
> > is this ok? [...]
>
> yes, this is indeed intended. With this we have more interrupts, thus more
> devices might be installed, which could not share interrupts previously.
> Newer motherboards have all kinds of kludges in them like mandatory USB
> interrupts, and in some cases it might prove useful to have those low IRQs
> freed up a bit.
>
> this is how 'high irqs' show up normally:
>
> CPU0 CPU1
> 0: 11816 0 XT PIC timer
> 1: 88 55 IO-APIC keyboard
> 2: 0 0 XT PIC cascade
> 8: 0 0 IO-APIC parport0
> 12: 0 0 IO-APIC PS/2 Mouse
> 13: 4 0 XT PIC fpu
> 16: 21 21 IO-APIC aic7xxx
> 18: 334 333 IO-APIC 3c905 Boomerang 100baseTx
> 19: 3476 3479 IO-APIC aic7xxx
>
> IRQ16, IRQ18, IRQ19 were 'low irqs' previously, now we map them to 'high
> vectors' with the IO-APIC. There are 8 new vectors, IRQ16 being the
> lowest, IRQ23 the highest.
>
> - -- mingo

I have been trying to understand PCI and the IO-APIC code. Is IO-APIC available on non-smp machines, or on a SMP motherboard with
only one chip in it? It seems like a newer/faster/better way of doing interrupts that breaks PCI away from ISA concepts, is this
accurate? If so these are true does 2.1.8x support the IO-APIC on non-smp?

Also, I saw someone say IPv6 doesn't work. Did it ever in 2.1.x? IF so can someone point me in the direction of the projects
page and I will try to at least help clean some of it up. (Again, I am new to kernel hacking but I really wish to add
something/anything positive back to Linux and its community).

Trever Adams
arabian@onramp.net

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu