* Re: IRQ routing / assignment issue
[not found] <fa.Ayf504otegg3VVbRsozI1UFd/6Y@ifi.uio.no>
@ 2008-09-01 17:21 ` Robert Hancock
2008-09-02 10:23 ` Radu Rendec
0 siblings, 1 reply; 4+ messages in thread
From: Robert Hancock @ 2008-09-01 17:21 UTC (permalink / raw)
To: Radu Rendec; +Cc: linux-kernel
Radu Rendec wrote:
> Hi,
>
> I'm trying to make the kernel assign different irqs for two NICs
> connected to _separate_ pci buses. Messing around with pirq= kenrel
> parameter didn't seem to change anything.
The kernel doesn't really have any way to change the physical
device/slot to IRQ mapping. In some cases there are BIOS settings which
might influence this, but in most cases it's impossible to change (i.e.
devices physically wired to the same IRQ line).
>
> The machine is a smp (dual-xeon). Relevant lspci output is:
> 01:00.0 Ethernet controller: Broadcom Corporation NetXtreme BCM5721
> Gigabit Ethernet PCI Express (rev 11)
> 02:00.0 Ethernet controller: Broadcom Corporation NetXtreme BCM5721
> Gigabit Ethernet PCI Express (rev 11)
>
> IRQ assignment dmesg output is:
> ACPI: PCI Interrupt 0000:01:00.0[A] -> GSI 16 (level, low) -> IRQ 16
> ACPI: PCI Interrupt 0000:02:00.0[A] -> GSI 16 (level, low) -> IRQ 16
>
> At the moment I'm almost clueless about how irq routing / assignment
> works (either in general or in linux), so any suggestion would help -
> even "rtfm" would be helpful, as long as you specify what is the
> appropriate manual ;)
>
> If you are so kind to reply, please include my address in cc (I'm not
> subscribed).
>
> Thanks,
>
> Radu Rendec
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: IRQ routing / assignment issue
2008-09-01 17:21 ` IRQ routing / assignment issue Robert Hancock
@ 2008-09-02 10:23 ` Radu Rendec
2008-09-02 14:47 ` Robert Hancock
0 siblings, 1 reply; 4+ messages in thread
From: Radu Rendec @ 2008-09-02 10:23 UTC (permalink / raw)
To: Robert Hancock; +Cc: linux-kernel
On Mon, 2008-09-01 at 11:21 -0600, Robert Hancock wrote:
> The kernel doesn't really have any way to change the physical
> device/slot to IRQ mapping. In some cases there are BIOS settings which
> might influence this, but in most cases it's impossible to change (i.e.
> devices physically wired to the same IRQ line).
Thanks for your quick reply. I know there are 4 physical interrupt
lines, interleaved for the many devices connected to a bus, so that, if
most of the devices issue INTA, the interrupts are actually asserted on
different pins of the apic (an almost uniform distribution of the
interrupts, but still, two devices can share the same physical line). I
also know that it's possible to change the (logical) irq levels assigned
to physical INTA-INTD (perhaps using the pirq= parameter).
As I have limited knowledge of hardware implementations, I'm asking you
if it's usual/possible for a dual-processor system to have the same
physical irq pin connected to the same irq line of _different_ pci
buses. To put it in other words, is it possible for two devices on
_separate_ pci buses to physically share the same irq pin?
Thanks,
Radu Rendec
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: IRQ routing / assignment issue
2008-09-02 10:23 ` Radu Rendec
@ 2008-09-02 14:47 ` Robert Hancock
0 siblings, 0 replies; 4+ messages in thread
From: Robert Hancock @ 2008-09-02 14:47 UTC (permalink / raw)
To: Radu Rendec; +Cc: linux-kernel
Radu Rendec wrote:
> On Mon, 2008-09-01 at 11:21 -0600, Robert Hancock wrote:
>> The kernel doesn't really have any way to change the physical
>> device/slot to IRQ mapping. In some cases there are BIOS settings which
>> might influence this, but in most cases it's impossible to change (i.e.
>> devices physically wired to the same IRQ line).
>
> Thanks for your quick reply. I know there are 4 physical interrupt
> lines, interleaved for the many devices connected to a bus, so that, if
> most of the devices issue INTA, the interrupts are actually asserted on
> different pins of the apic (an almost uniform distribution of the
> interrupts, but still, two devices can share the same physical line). I
> also know that it's possible to change the (logical) irq levels assigned
> to physical INTA-INTD (perhaps using the pirq= parameter).
>
> As I have limited knowledge of hardware implementations, I'm asking you
> if it's usual/possible for a dual-processor system to have the same
> physical irq pin connected to the same irq line of _different_ pci
> buses. To put it in other words, is it possible for two devices on
> _separate_ pci buses to physically share the same irq pin?
Most of the systems I've seen with multiple PCI buses (servers, mostly)
have normally had separate IRQ pins for each slot for at least INTA, in
fact. However, I don't think there's any reason why you couldn't have
the same IRQ pin routed to slots on different buses..
^ permalink raw reply [flat|nested] 4+ messages in thread
* IRQ routing / assignment issue
@ 2008-09-01 12:59 Radu Rendec
0 siblings, 0 replies; 4+ messages in thread
From: Radu Rendec @ 2008-09-01 12:59 UTC (permalink / raw)
To: linux-kernel
Hi,
I'm trying to make the kernel assign different irqs for two NICs
connected to _separate_ pci buses. Messing around with pirq= kenrel
parameter didn't seem to change anything.
The machine is a smp (dual-xeon). Relevant lspci output is:
01:00.0 Ethernet controller: Broadcom Corporation NetXtreme BCM5721
Gigabit Ethernet PCI Express (rev 11)
02:00.0 Ethernet controller: Broadcom Corporation NetXtreme BCM5721
Gigabit Ethernet PCI Express (rev 11)
IRQ assignment dmesg output is:
ACPI: PCI Interrupt 0000:01:00.0[A] -> GSI 16 (level, low) -> IRQ 16
ACPI: PCI Interrupt 0000:02:00.0[A] -> GSI 16 (level, low) -> IRQ 16
At the moment I'm almost clueless about how irq routing / assignment
works (either in general or in linux), so any suggestion would help -
even "rtfm" would be helpful, as long as you specify what is the
appropriate manual ;)
If you are so kind to reply, please include my address in cc (I'm not
subscribed).
Thanks,
Radu Rendec
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2008-09-02 14:47 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
[not found] <fa.Ayf504otegg3VVbRsozI1UFd/6Y@ifi.uio.no>
2008-09-01 17:21 ` IRQ routing / assignment issue Robert Hancock
2008-09-02 10:23 ` Radu Rendec
2008-09-02 14:47 ` Robert Hancock
2008-09-01 12:59 Radu Rendec
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox