Re: interrupts across PCI bridge(s) not handled

[Zwane Mwaikambo]

On Mon, 3 Nov 2003, Linus Torvalds wrote:

> 
> On Mon, 3 Nov 2003, Charles Martin wrote:
> >
> > I have a pci backplane extender, with 4 cards 
> > (named piraq) in it. The cards are detected by 
> > the PCI system, and irqs 92-95 are assigned, 
> > as shown in /var/log/messages:
> > 
> > kernel: PCI->APIC IRQ transform: (B6,I4,P0) -> 93 
> > kernel: PCI->APIC IRQ transform: (B6,I6,P0) -> 95 
> > kernel: PCI->APIC IRQ transform: (B6,I7,P0) -> 92
> > kernel: PCI->APIC IRQ transform: (B6,I9,P0) -> 94
> 
> Can you enable APIC_DEBUG debugging in "include/asm-i386/apic.h", and make 
> sure that you build the kernel with a big printk buffer. Then, in case 
> your distribution comes with a broken "dmesg" binary that doesn't show 
> more than about 20kB of data, compile this trivial program and run it 
> after bootup, and send the whole log out..
> 
> It would be a pity to have to boot with "noapic", since this is exactly 
> the kind of situation where you _want_ the extra interrupts.

I wonder how the request_irq worked there, doesn't 'none' denote no 
irq_desc[i].handler == no_irq_type?
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/