IRQ Conflict

Tue Jan 6 00:43:54 UTC 2009

I found that using the noirqdebug kernel option seems to be the catch-all solution to this problem.  The description of this kernel option is: "By default, the kernel attempts to detect and disable unhandled
	interrupt sources because they can cause problems with the
	responsiveness of the rest of the kernel if left unchecked. This
	option will disable this logic."  Basically this no longer displays the IRQ Disabled message and does not disable the IRQ at all.

So, it seems like there is an unhandled interrupt which in causing the kernel to disable the IRQ, causing the devices to hang up.  Others have fixed the problem by disabling/re-configuring the device causing the unhandled interrupt.  My question is why is the kernel not handling the interrupt?

I did mention the hardware involved in my first message, but specifically I have a stock Dell SC1420 with a Highpoint RAID 2322 add on card.  My USB device, NIC, and RAID card are sharing an IRQ and after some time the kernel disables that IRQ and the system becomes unresponsive.  For other people the symtpoms they've seen depend on which devices were involved in the diabled IRQ.

I don't have the ability to change any IRQ assignments in my BIOS.

- Curtis George

> Date: Mon, 5 Jan 2009 14:32:35 +0900
> From: debian at herakles.homelinux.org
> To: fedora-test-list at redhat.com
> Subject: Re: IRQ Conflict
> 
> Curtis George wrote:
> > Thanks for the info, I have discovered some interesting things recently.  I went ahead and reverted to Fedora 9 since I knew that one worked for me and found that the IRQ is being shared by the same 3 devices.  I then updated the kernel on Fedora 9 to 2.6.27 (which is also what Fedora 10 starts out with) and got the same IRQ disabled messages and sure enough my computer locked up within a few minutes of booting that kernel.  Luckily I still had the 2.6.25 kernel installed, so I booted into that one instead.  
> > 
> > Does this mean there is a possible problem with the 2.6.27 kernel?  What is the actual cause of this problem?
> > 
> > If you do a google search on Linux IRQ Disabled you will find that this problem has existed in various forms without identifying the actual cause.  I should mention that I cannot move around my devices (they are either builtin or using the only compatible slot).  I also do not have a disable PnP OS option in BIOS.  I also saw a suggestion to change the SATA mode to AHCI in BIOS instead of IDE, but I don't have that option either (for me only RAID or non-RAID).  This worked from some, but I think in their case the ata device was involved in the interrupt that was being disabled.
> 
> Likely, if there's no more serious problem than the message folk don't 
> notice it and/or simply ignore it.
> 
> Do you have an option to change IRQ assignments in the BIOS?
> 
> 
> I don't recall that you've mentioned anything about the hardware. There 
> are some pointers in my sig to "asking good questions."  Many problems 
> arise on a small set of hardware or in rare circumstances. You are our 
> eyes, and if you don't give us the information we're blind, and a lot of 
> people will ignore you.
> 
> > 
> > - Curtis George
> > 
> >> Date: Sun, 4 Jan 2009 21:03:14 +0900
> >> From: debian at herakles.homelinux.org
> >> To: fedora-test-list at redhat.com
> >> Subject: Re: IRQ Conflict
> >>
> >> Mick M. wrote:
> >>>> I'm guessing that when I installed Fedora 9 with all
> >>>> the USB unplugged that is chose differet IRQs so there was
> >>>> no confilict (I'm not sure though).  I can't
> >>>> manually set the IRQ on any of these devices, it seems to be
> >>>> up to the OS to choose.  Any ideas on what is causing this
> >>>> or how to fix it?
> >>>>
> >>>> - Curtis George
> >>>>
> >>> Hi;
> >>>   try moving cards around to different slots.
> >>> Some motherboards assign irq's to specific slots.
> >>> Then some cards only work with certain irq's.
> >>>
> >>> On bootup you may see what irq goes where.
> >>> You may be able to play in the BIOS to help out.
> >> PCI interrupts are supposed to be sharable. There are four interrupts, 
> >> INT#A through INT#D which can be mapped to ISA interrupts any way the 
> >> BIOS or OS likes.
> >>
> >> If the PCI interrupts on your system are not sharable, then the system's 
> >> broken.
> >>
> >> If your BIOS has asks about a PnP OS, change it and try that. Generally, 
> >> I say I have one, have said so since 2.4 kernels.
> >>
> >> If that does not work, say "no" and try assigning some IRQs in the BIOS.
> >>
> >>
> >> Shuffling cards might help, but most of my systems' PCI slots are empty.
> >>
> >>
> >>
> >>
> >>
> >> -- 
> >>
> >> Cheers
> >> John
> >>
> >> -- spambait
> >> 1aaaaaaa at coco.merseine.nu  Z1aaaaaaa at coco.merseine.nu
> >> -- Advice
> >> http://webfoot.com/advice/email.top.php
> >> http://www.catb.org/~esr/faqs/smart-questions.html
> >> http://support.microsoft.com/kb/555375
> >>
> >> You cannot reply off-list:-)
> >>
> >> -- 
> >> fedora-test-list mailing list
> >> fedora-test-list at redhat.com
> >> To unsubscribe: 
> >> https://www.redhat.com/mailman/listinfo/fedora-test-list
> > 
> > _________________________________________________________________
> > Send e-mail anywhere. No map, no compass.
> > http://windowslive.com/oneline/hotmail?ocid=TXT_TAGLM_WL_hotmail_acq_anywhere_122008
> > 
> 
> 
> -- 
> 
> Cheers
> John
> 
> -- spambait
> 1aaaaaaa at coco.merseine.nu  Z1aaaaaaa at coco.merseine.nu
> -- Advice
> http://webfoot.com/advice/email.top.php
> http://www.catb.org/~esr/faqs/smart-questions.html
> http://support.microsoft.com/kb/555375
> 
> You cannot reply off-list:-)
> 
> -- 
> fedora-test-list mailing list
> fedora-test-list at redhat.com
> To unsubscribe: 
> https://www.redhat.com/mailman/listinfo/fedora-test-list

_________________________________________________________________
It’s the same Hotmail®. If by “same” you mean up to 70% faster.
http://windowslive.com/online/hotmail?ocid=TXT_TAGLM_WL_hotmail_acq_broad1_122008
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/fedora-test-list/attachments/20090105/e825448a/attachment.htm>