Server, interrupted.
dsr at tao.merseine.nu
dsr at tao.merseine.nu
Wed Jan 26 15:08:28 EST 2005
On Wed, Jan 26, 2005 at 02:28:21PM -0500, David Kramer wrote:
> On Wed, 26 Jan 2005, Gregory Boyce wrote:
> > Are you using IO-APIC on the machine? An easy way to check
> > /proc/interrupts and look for either "XT-PIC" or "IO-APIC". Using the
> > APIC can sometimes help with IRQ issues I believe.
> >
>
> [david at uni ~]$ cat /proc/interrupts
> CPU0
> 0: 42051885 XT-PIC timer
> 1: 6834 XT-PIC keyboard
> 2: 0 XT-PIC cascade
> 8: 2 XT-PIC rtc
> 9: 95591 XT-PIC acpi, DE450-TA (eth0), ehci_hcd,
> usb-uhci, usb-uhci, usb-uhci
> 10: 1665 XT-PIC eth1, ohci1394
> 11: 3526734 XT-PIC EMU10K1, radeon at PCI:1:0:0
> 12: 56078 XT-PIC PS/2 Mouse
> 14: 531629 XT-PIC ide0
> NMI: 0
> LOC: 0
> ERR: 115
> MIS: 0
>
> Never heard those terms before. The limited STFW I could do at work
> indicates that those terms are associated with multiprocessor boxen, which
> mine isn't.
CPU0 CPU1
0: 481084576 487725791 IO-APIC-edge timer
1: 11705 10119 IO-APIC-edge keyboard
2: 0 0 XT-PIC cascade
4: 25679 17876 IO-APIC-edge serial
8: 4 0 IO-APIC-edge rtc
9: 0 0 IO-APIC-level acpi
14: 24118901 23759014 IO-APIC-edge ide0
15: 3887540 3916558 IO-APIC-edge ide1
16: 0 0 IO-APIC-level usb-uhci
18: 0 0 IO-APIC-level usb-uhci
19: 0 0 IO-APIC-level usb-uhci
24: 21904909 21553414 IO-APIC-level ide2, ide3
32: 17856807 17659950 IO-APIC-level eth0
NMI: 0 0
LOC: 968811046 968811044
ERR: 0
MIS: 7958
>From a recent uniproc:
CPU0
0: 27120934 IO-APIC-edge timer
1: 8 IO-APIC-edge i8042
4: 7 IO-APIC-edge serial
7: 2 IO-APIC-edge parport0
8: 4 IO-APIC-edge rtc
9: 0 IO-APIC-level acpi
14: 91315875 IO-APIC-edge ide0
15: 15 IO-APIC-edge ide1
169: 0 IO-APIC-level libata
177: 0 IO-APIC-level ehci_hcd, uhci_hcd, uhci_hcd,
uhci_hcd, uhci_hcd
185: 0 IO-APIC-level via82cxxx
193: 333098450 IO-APIC-level eth0
NMI: 0
LOC: 27418477
ERR: 0
MIS: 0
Notice that IO-APIC gives you both level and edge-triggered
IRQs, and a whole lot more of them as a result.
-dsr-
--
Nothing to sig here, move along.
More information about the Discuss
mailing list