apic vs. ASUS P5N-E (nforce 650)

John Baldwin jhb at freebsd.org
Tue Nov 6 13:38:31 PST 2007

On Thursday 01 November 2007 02:40:12 pm Max Laier wrote:
> Hi,
> I'm having big trouble with this board.  It works fine once I set 
> hint.apic.0.disabled=1, but unfortunately without using the second core.  
> When the apic is used, the nforce IDE/ATA controller and the nforce NIC 
> are not working at all.  No disks or CDROMs are recognized, the NIC 
> doesn't send or receive any packets.
> Attached are verbose dmesgs (captured via serial):
>   boot_v.apic   - ACPI enabled  APIC enabled
>   boot_v.noacpi - ACPI disabled APIC enabled
>   boot_v.noapic - ACPI enabled  APIC disabled (this works)
> FWIW, linux (knoppix CD) does boot and seems to use the apic (dmesg 
> attached).

FreeBSD + ACPI uses the same IRQs for the nvidia ata0 and ata1 (legacy IRQs 14 
and 15) and for nfe0 (GSI 23 which in FreeBSD is IRQ 23 via an ACPI PCI link 
device) as Linux.

> I suspect that there is something wrong with the ACPI interrupt overrides.  
> IRQ9 is listed as (level, high) and IRQ0->2 seems only to make sense if 
> the HEPT is being used, but neither ACPI_Q_MADT_IRQ0 nor forcing (level, 
> low) for IRQ9 in madt.c doesn't seem to do the trick.

No, those are quite normal.  The 0 => 2 has nothing to do with HPET but is 
because normally pin 0 on the first I/O APIC is an ExtINT pin connected to 
the master 8259A and IRQ0 is hooked up to pin 2 instead.  Also, I've seen 
lots of odd settings for IRQ 9 (normally used for ACPI SCI), so that is 
probably fine.

John Baldwin

More information about the freebsd-current mailing list