Disabling IRQ #16

English Hardware
1

Problem: kernel:[10275.557720] Disabling IRQ #16
Symptom: High ping using the added Lan Card
I have an Opensuse 11.3 Server with two NIC. 1st onboard no problem, 2nd additional Lan card experienced high ping after this message kernel:[10275.557720] Disabling IRQ #16. is someone experienced this?

I am a newbie in Linux I hope someone can help me

output
Mar 25 12:57:36 Vicidial kernel: [87665.570010] irq 16: nobody cared (try booting with the “irqpoll” option)
Mar 25 12:57:36 Vicidial kernel: [87665.570015] Pid: 0, comm: swapper Not tainted 2.6.34.10-0.6-pae #1
Mar 25 12:57:36 Vicidial kernel: [87665.570017] Call Trace:
Mar 25 12:57:36 Vicidial kernel: [87665.570036] <c0206343>] try_stack_unwind+0x173/0x190
Mar 25 12:57:36 Vicidial kernel: [87665.570039] <c020508f>] dump_trace+0x3f/0xe0
Mar 25 12:57:36 Vicidial kernel: [87665.570042] <c02063ab>] show_trace_log_lvl+0x4b/0x60
Mar 25 12:57:36 Vicidial kernel: [87665.570043] <c02063d8>] show_trace+0x18/0x20
Mar 25 12:57:36 Vicidial kernel: [87665.570046] <c05d964c>] dump_stack+0x6d/0x72
Mar 25 12:57:36 Vicidial kernel: [87665.570049] <c029da34>] __report_bad_irq+0x24/0x90
Mar 25 12:57:36 Vicidial kernel: [87665.570051] <c029dc06>] note_interrupt+0x166/0x1a0
Mar 25 12:57:36 Vicidial kernel: [87665.570053] <c029e75a>] handle_fasteoi_irq+0xaa/0xd0
Mar 25 12:57:36 Vicidial kernel: [87665.570055] <c0205045>] handle_irq+0x15/0x20
Mar 25 12:57:36 Vicidial kernel: [87665.570057] <c0204d67>] do_IRQ+0x47/0xc0
Mar 25 12:57:36 Vicidial kernel: [87665.570059] <c0203729>] common_interrupt+0x29/0x30
Mar 25 12:57:36 Vicidial kernel: [87665.570077] <f7a7cfc0>] acpi_idle_enter_bm+0x210/0x239 [processor]
Mar 25 12:57:36 Vicidial kernel: [87665.570082] <c04ebf1a>] cpuidle_idle_call+0x7a/0x100
Mar 25 12:57:36 Vicidial kernel: [87665.570084] <c0201b4f>] cpu_idle+0x3f/0x80
Mar 25 12:57:36 Vicidial kernel: [87665.570086] handlers:
Mar 25 12:57:36 Vicidial kernel: [87665.570086] <f7e64c60>] (nouveau_irq_handler+0x0/0x190 [nouveau])
Mar 25 12:57:36 Vicidial kernel: [87665.570092] <fa9b9750>] (rtl8139_interrupt+0x0/0x230 [8139too])
Mar 25 12:57:36 Vicidial kernel: [87665.570095] Disabling IRQ #16

thanks in advance

2

So here are some thoughts on this situation.

  1. This seldom seems to cause any problems with a stable running system.
  2. Its often thought to be associated with multiple devices using the same IRQ.
  3. Kernel developers will often blame this on a proprietary video driver or Wireless code said to taint the kernel.
  4. Adding in the kernel load option: pci=routeirq is said to help.
  5. Sometimes doing a complete BIOS reset back to defaults might help get rid of the problem.

In general, when I have seen this, nothing bad seemed to happen and I do load the nVIDIA proprietary video driver. It can be ignored by anyone running a stable Linux system. If you load a propritary video driver, you will have trouble finding someone to help find the problem.

Thank You,

3

On 03/27/2012 05:36 PM, jdmcdaniel3 wrote:
> 1. This seldom seems to cause any problems with a stable running
> system.
> 2. Its often thought to be associated with multiple devices using the
> same IRQ.
> 3. Kernel developers will often blame this on a proprietary video
> driver or Wireless code said to taint the kernel.
> 4. Adding in the kernel load option: pci=routeirq is said to help.
> 5. Sometimes doing a complete BIOS reset back to defaults might help
> get rid of the problem.
>
> In general, when I have seen this, nothing bad seemed to happen and I
> do load the nVIDIA proprietary video driver. It can be ignored by
> anyone running a stable Linux system. If you load a propritary video
> driver, you will have trouble finding someone to help find the problem.

If you load ANY non-GPL driver, developers will not look at the problem.

This type of problem will cause problems if you have a needed device sharing
that interrupt.

You can see what devices want to use that interrupt by running the command

cat /proc/interrupts

4

Thanks for the info Larry. When run on my system, it finds irq 16 is indeed being used by nVIDIA and I have see this error pop up before in terminal.

16:    5360516          0          0          0          0          0          0          0   IO-APIC-fasteoi   nvidia

Thank You,

5

So here are some thoughts on this situation.

  1. This seldom seems to cause any problems with a stable running system.
  2. Its often thought to be associated with multiple devices using the same IRQ.
  3. Kernel developers will often blame this on a proprietary video driver or Wireless code said to taint the kernel.
  4. Adding in the kernel load option: pci=routeirq is said to help.
  5. Sometimes doing a complete BIOS reset back to defaults might help get rid of the problem.

In general, when I have seen this, nothing bad seemed to happen and I do load the nVIDIA proprietary video driver. It can be ignored by anyone running a stable Linux system. If you load a propritary video driver, you will have trouble finding someone to help find the problem.

Thank You,

  1. Adding in the kernel load option: pci=routeirq is said to help.
    Care to explain this in detail on how to do this on my system? this is my first time I will be editing the kernel so I hope I do not mess up

thanks guys for the replies. really appreciate this

6

also, I forgot yes I have two hardware sharing for interrupt 16. the Video card and Lan card. If say I will be blocking the Video card from using interrupt 16 how should I do that so Lan card will only be use.

thanks :wink:

7

This an option you can try to see if it gets rid of the error message. Here is the definition of this kernel option:

pci=routeirq

Do IRQ routing for all PCI devices. This is normally done in pci_enable_device(), so this option is a temporary workaround for broken drivers that don't call it.

The following link shows many of the kernel load options that exist: Kernel boot command-line parameter reference: Chapter 9 - Linux Kernel in a Nutshell - O’Reilly Media

Thank You,