/proc/interrupt oddities [was Re: APIC error on CPU0: 40(40)]

[D. Hugh Redelmeier]

| From: D. Hugh Redelmeier <hugh at mimosa.com>

| | > I'm getting a lot of these messages showing up in dmesg output.

| 	APIC error on CPU0: 40(40)

| | > Description of system:
| | > - HP Pavilion a1350n desktop computer
| | > - AMD Athlon(tm) 64 X2 Dual Core Processor 3800+
| | > - ATI chipset
| | > - nVidia 7800GTX PCIe video card (and 450w Antec power supply to
| | >   support it)
| | > - Fedora Core 4 for x86_64 with updates as of last week
| | > - kernel-smp-2.6.15-1.1831_FC4 for x86_64

While investigating my APIC errors, I examined /proc/interrupts:

           CPU0       CPU1       
  0:  115072355    1332193    IO-APIC-edge  timer
  1:      66716          0    IO-APIC-edge  i8042
  7:    1332193   56869707    IO-APIC-edge  parport0
  8:          0          0    IO-APIC-edge  rtc
 12:     812608          0    IO-APIC-edge  i8042
 14:    8299026          0    IO-APIC-edge  ide0
169:          2          0   IO-APIC-level  acpi, ohci1394
201:     302988          0   IO-APIC-level  libata
209:    1124567          0   IO-APIC-level  eth0
217:         12          0   IO-APIC-level  ATI IXP
225:   51656405          0   IO-APIC-level  ehci_hcd:usb1, ohci_hcd:usb2, ohci_hcd:usb3
NMI:       6739       2960 
LOC:   58204857   58204845 
ERR:        558
MIS:          0

(uptime is 5 days)

ERR is the same as the count of APIC errors, as far as I can tell.  It
should be.  See the section on /proc/interrupts in
  http://www.linux.com/guides/Linux-Filesystem-Hierarchy/proc.shtml

The number of parport0 interrupts, on both CPUs is astonishing.  I
don't have anything hooked up to the parallel port.  1,223,193 on CPU0
and 56,869,707 on CPU1.  About 120 per second on CPU1.

Why would it be getting these interrupts?  How are they distributed
between CPUs?  Why so unevenly?