help with interrupt storm

Francisco Javier Cabello Torres fjcabello at visual-tools.com
Thu Jul 21 06:41:24 PDT 2011


Dear all,

I have an issue with an interrupt storm generated by 82945G graphics
controller.

I will try to give you as much as detail as I can.

System detail:
-Debian lenny
-Xorg version 1.7.7
-Intel driver 2.12.0
-Kernel 2.6.35.4

My application display video from 16 cameras. Each one in a Xv Port. The
video is gotten from saa7134 devices.

Due to my setup, i915 is sharing irq with several devices:

$ cat /proc/interrupts

# cat /proc/interrupts
           CPU0       CPU1
  0:         43          0   IO-APIC-edge      timer
  1:          8          0   IO-APIC-edge      i8042
  4:          1          0   IO-APIC-edge      serial
  6:          2          0   IO-APIC-edge      floppy
  9:          0          0   IO-APIC-fasteoi   acpi
 12:        105          0   IO-APIC-edge      i8042
 14:          0          0   IO-APIC-edge      ide0
 16:   23084793          0   IO-APIC-fasteoi   uhci_hcd:usb5, saa7130[0],
saa7130[2], saa7130[4], saa7130[6], saa7130[12], saa7130[13], saa7130[14],
saa7130[15], i915
 17:      34188          0   IO-APIC-fasteoi   Intel ICH7, saa7130[8],
saa7130[10]
 18:     144924          0   IO-APIC-fasteoi   uhci_hcd:usb4, eth0,
saa7130[1], saa7130[3], saa7130[5], saa7130[7]
 19:     208002          0   IO-APIC-fasteoi   ata_piix, uhci_hcd:usb3,
saa7130[9], saa7130[11]
 23:          0          0   IO-APIC-fasteoi   ehci_hcd:usb1, uhci_hcd:usb2
NMI:          0          0   Non-maskable interrupts
LOC:     390238     390215   Local timer interrupts
SPU:          0          0   Spurious interrupts
PMI:          0          0   Performance monitoring interrupts
PND:          0          0   Performance pending work
RES:     122117     183756   Rescheduling interrupts
CAL:        165        123   Function call interrupts
TLB:      52101      63521   TLB shootdowns
TRM:          0          0   Thermal event interrupts
THR:          0          0   Threshold APIC interrupts
MCE:          0          0   Machine check exceptions
MCP:          6          6   Machine check polls
ERR:          1
MIS:          0


I test IRQ rate checking /proc/interrupts every second and the rate for IRQ
16 is 175 irq/seconds. I have 8 video devices, each one generating 25
irq/sec (25 frames per second).
This is what I expect. The problem appears after some minutes. IRQ rate
rises to 30000 irq/seconds on IRQ16. It seems to be an interrupt storm.

I have tested with a newer Intel chipset which support MSI
(Message-Signalled-Interrupts) and the problem disappear.  The problem is
that MSI support is buggy on 945G chipset.

I have tried with newer Xorg version (1.9.3),  kernel  (2.6.38.6,
2.6.35.14), intel driver (2.15.0) and the behaviour is always the same.

I have googled a lot and I feel that my problem is similar to the people
claiming 'irq 16 nobody cared'. It happens to me also when I remove saa7134
driver. However I didn't find
any solution for 'irq 16 nobody cared problem'... :(

Do you have any clue about this problem?

Thanks in advance,

Paco
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.x.org/archives/xorg/attachments/20110721/f9fa8cf2/attachment.html>


More information about the xorg mailing list