On 18/01/2011 15:34, Drew Weaver wrote:

Ah, it is a PCI card and there is only one PCI slot..

-Drew

-----Original Message----- From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of Giles Coochey Sent: Tuesday, January 18, 2011 9:24 AM To: centos@centos.org Subject: Re: [CentOS] Intel DH67BL + CentOS 5.5 IRQ #177 nobody cared

On 18/01/2011 15:22, Drew Weaver wrote:

...

Hello,

We have built a couple of CentOS 5.5 systems on the Intel DH67BL (Sandy Bridge) motherboard and overall they work pretty well.

I'm having one problem which keeps cropping up and that is an error message that says:

IRQ 177 nobody cared (try booting with the irqpoll option)

report bad irq, references CPU idle

then it references usb_hcd_irq and e1000_intr

then it disables the add-on PCI E1000 card and disconnects itself from the network.

I'm assuming what is happening here is the USB controller and the add-on E1000 controller we put in are having an old school IRQ conflict, the question is why and how can I avoid it?

I have tried disabling all of the extra stuff in the BIOS that I could, and this still happens fairly frequently.

Any advice would be great.

Have you tried putting one of the cards in a different slot?

Does the BIOS allow you to manually set a IRQ to the PCI line?

On 18/01/2011 16:12, Drew Weaver wrote:

Hi,

It doesn't appear to, it simply shows a little ASCII diagram of what is set for each slot.

thanks, -Drew

...

Have you tried putting one of the cards in a different slot?

Hi,

Sometimes that is only editable after you disable an option resembling "Plug and Play OS"

Why don't you use the mother board incorporated NIC? it does have one right? ------- Because the installer doesn't have drivers for the onboard and all of our installs are PXE and in general it removes a lot of confusion by just disabling the onboard NIC and having one single NIC for everything.

thanks, -Drew

________________________________ From: "Drew Weaver" drew.weaver@thenap.com To: "CentOS mailing list" centos@centos.org Sent: Tuesday, 18 January, 2011 2:22:04 PM Subject: [CentOS] Intel DH67BL + CentOS 5.5 IRQ #177 nobody cared

Hello,

We have built a couple of CentOS 5.5 systems on the Intel DH67BL (Sandy Bridge) motherboard and overall they work pretty well.

I'm having one problem which keeps cropping up and that is an error message that says:

IRQ 177 nobody cared (try booting with the irqpoll option) report bad irq, references CPU idle

then it references usb_hcd_irq and e1000_intr then it disables the add-on PCI E1000 card and disconnects itself from the network.

I'm assuming what is happening here is the USB controller and the add-on E1000 controller we put in are having an old school IRQ conflict, the question is why and how can I avoid it?

I have tried disabling all of the extra stuff in the BIOS that I could, and this still happens fairly frequently.

Any advice would be great.

thanks, -Drew

_______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos

Hi,

Just an update to the list.

This issue is still present, I even tried it on a different Intel socket 1155 motherboard (DH61WW) which is about as plain vanilla as it gets.

It must be something to do with the onboard video and PCI Express.

Has anyone been able to get around this?

-Drew

-----Original Message----- From: Kenni Lund [mailto:kenni@kelu.dk] Sent: Thursday, January 20, 2011 6:46 AM To: CentOS mailing list Cc: Drew Weaver Subject: Re: [CentOS] Intel DH67BL + CentOS 5.5 IRQ #177 nobody cared

2011/1/18 Drew Weaver drew.weaver@thenap.com:

Because the installer doesn't have drivers for the onboard and all of our installs are PXE and in general it removes a lot of confusion by just disabling the onboard NIC and having one single NIC for everything.

Drew, out of curiosity (I have a similar motherboard in backorder), does the latest CentOS kernel have drivers for the onboard NIC after the installation? If not, I'll probably cancel my order and try a Gigabyte board instead....thanks in advance!

Best regards Kenni

On 18/01/2011 16:43, Radu Radutiu wrote:

...

IRQ 177 nobody cared (try booting with the irqpoll option)

Have you tried what the error message suggests (add irqpool to the kernel line in grub.conf) ?

It may get it working, but I've heard it seriously affect performance...

http://www.lindevdoc.org/wiki/irqpoll_(kernel_parameter)

Hi,

There does not appear to be a 'Plug and Play OS' option in the BIOS like on previous Intel motherboards.

I have disabled all of the USB ports except for the rear panel ones and so far it has been running for awhile.

I will see what happens.

thanks, -Drew

From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of compdoc Sent: Tuesday, January 18, 2011 10:57 AM To: 'CentOS mailing list' Subject: Re: [CentOS] Intel DH67BL + CentOS 5.5 IRQ #177 nobody cared

IRQ 177 nobody cared (try booting with the irqpoll option) report bad irq, references CPU idle

I'm assuming what is happening here is the USB controller and the add-on E1000 controller we put in are having an old school IRQ conflict, the question is why and how can I avoid it?

IRQ177 means you are using APIC, which gives you over 250 interrupts - far more than there were in the olden days. There should be enough that nothing has to share IRQs.

But even before APIC came along, devices had gotten pretty good at sharing IRQs.

As someone suggested, disable the Plug and Play option in the bios. Try a different brand of Ethernet card?

Btw, this is the complete error.

Jan 16 07:03:23 server kernel: irq 177: nobody cared (try booting with the "irqpoll" option) Jan 16 07:03:23 server kernel: [<c044d886>] __report_bad_irq+0x2b/0x69 Jan 16 07:03:23 server kernel: [<c044da73>] note_interrupt+0x1af/0x1e8 Jan 16 07:03:23 server kernel: [<c044d081>] handle_IRQ_event+0x45/0x8c Jan 16 07:03:23 server kernel: [<c044d163>] __do_IRQ+0x9b/0xd6 Jan 16 07:03:23 server kernel: [<c044d0c8>] __do_IRQ+0x0/0xd6 Jan 16 07:03:23 server kernel: [<c040749e>] do_IRQ+0x99/0xc3 Jan 16 07:03:23 server kernel: [<c0405946>] common_interrupt+0x1a/0x20 Jan 16 07:03:23 server kernel: [<c05302c0>] acpi_processor_idle_simple+0x0/0x254 Jan 16 07:03:23 server kernel: [<c052fc0b>] acpi_safe_halt+0x14/0x20 Jan 16 07:03:23 server kernel: [<c0530367>] acpi_processor_idle_simple+0xa7/0x254 Jan 16 07:03:23 server kernel: [<c053007b>] acpi_processor_get_power_info+0x464/0x510 Jan 16 07:03:23 server kernel: [<c0403ca8>] cpu_idle+0x9f/0xb9 Jan 16 07:03:23 server kernel: ======================= Jan 16 07:03:23 server kernel: handlers: Jan 16 07:03:23 server kernel: [<c058afd1>] (usb_hcd_irq+0x0/0x50) Jan 16 07:03:23 server kernel: [<f8964b7f>] (e1000_intr+0x0/0x10b [e1000]) Jan 16 07:03:23 server kernel: Disabling IRQ #177

In case again anyone is interested.

I just tried updating the bios, if it freezes again I will boot with acpi=off and then retest.

-Drew

From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of Drew Weaver Sent: Tuesday, January 18, 2011 3:25 PM To: 'CentOS mailing list' Subject: Re: [CentOS] Intel DH67BL + CentOS 5.5 IRQ #177 nobody cared

Issue still persists, anyway I thought I'd let you guys know in case someone is thinking about getting one or more of these boards.

-Drew

From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of Drew Weaver Sent: Tuesday, January 18, 2011 11:42 AM To: 'CentOS mailing list' Subject: Re: [CentOS] Intel DH67BL + CentOS 5.5 IRQ #177 nobody cared

Hi,

There does not appear to be a 'Plug and Play OS' option in the BIOS like on previous Intel motherboards.

I have disabled all of the USB ports except for the rear panel ones and so far it has been running for awhile.

I will see what happens.

thanks, -Drew

From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of compdoc Sent: Tuesday, January 18, 2011 10:57 AM To: 'CentOS mailing list' Subject: Re: [CentOS] Intel DH67BL + CentOS 5.5 IRQ #177 nobody cared

IRQ 177 nobody cared (try booting with the irqpoll option) report bad irq, references CPU idle

I'm assuming what is happening here is the USB controller and the add-on E1000 controller we put in are having an old school IRQ conflict, the question is why and how can I avoid it?

IRQ177 means you are using APIC, which gives you over 250 interrupts - far more than there were in the olden days. There should be enough that nothing has to share IRQs.

But even before APIC came along, devices had gotten pretty good at sharing IRQs.

As someone suggested, disable the Plug and Play option in the bios. Try a different brand of Ethernet card?

The keyboard that I use to ctrl-alt-del it is USB, but there was no keyboard connected until after it locked up, so it's a chicken and egg thing.

-Drew

-----Original Message----- From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of m.roth@5-cent.us Sent: Tuesday, January 18, 2011 3:59 PM To: CentOS mailing list Subject: Re: [CentOS] Intel DH67BL + CentOS 5.5 IRQ #177 nobody cared

Drew Weaver wrote:

Btw, this is the complete error.

<snip> Hmmm... here's a thought: do you have anything plugged into a USB port? If so, try booting with it out; if not, try booting with one in.

mark

_______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos

Btw, this is the complete error.

Jan 16 07:03:23 server kernel: irq 177: nobody cared (try booting with the "irqpoll" option) Jan 16 07:03:23 server kernel:  [<c044d886>] __report_bad_irq+0x2b/0x69 Jan 16 07:03:23 server kernel:  [<c044da73>] note_interrupt+0x1af/0x1e8 Jan 16 07:03:23 server kernel:  [<c044d081>] handle_IRQ_event+0x45/0x8c Jan 16 07:03:23 server kernel:  [<c044d163>] __do_IRQ+0x9b/0xd6 Jan 16 07:03:23 server kernel:  [<c044d0c8>] __do_IRQ+0x0/0xd6 Jan 16 07:03:23 server kernel:  [<c040749e>] do_IRQ+0x99/0xc3 Jan 16 07:03:23 server kernel:  [<c0405946>] common_interrupt+0x1a/0x20 Jan 16 07:03:23 server kernel:  [<c05302c0>] acpi_processor_idle_simple+0x0/0x254 Jan 16 07:03:23 server kernel:  [<c052fc0b>] acpi_safe_halt+0x14/0x20 Jan 16 07:03:23 server kernel:  [<c0530367>] acpi_processor_idle_simple+0xa7/0x254 Jan 16 07:03:23 server kernel:  [<c053007b>] acpi_processor_get_power_info+0x464/0x510 Jan 16 07:03:23 server kernel:  [<c0403ca8>] cpu_idle+0x9f/0xb9            Jan 16 07:03:23 server kernel:  ======================= Jan 16 07:03:23 server kernel: handlers: Jan 16 07:03:23 server kernel: [<c058afd1>] (usb_hcd_irq+0x0/0x50) Jan 16 07:03:23 server kernel: [<f8964b7f>] (e1000_intr+0x0/0x10b [e1000]) Jan 16 07:03:23 server kernel: Disabling IRQ #177

In case again anyone is interested.

I just tried updating the bios, if it freezes again I will boot with acpi=off and then retest.

------------ acpi=off in the kernel options has allowed the machine to run for over 12 hours without crashing, we'll see what happens. doesn't seem like a hardware problem.

thanks, -Drew

Jan 16 07:03:23 server kernel: irq 177: nobody cared (try booting with the "irqpoll" option) I spoke too soon, the problem wasn't fixed but I found the cause of the issue. The above error occurs when you unplug the video cable from the onboard video. As our machines are mainly headless terminals, this will not do at all.

Does the BIOS allow you to disable the video hardware completely? Does that enable your kernel to boot?

If not, grab the kernel source and build a kernel without console support (which is what I had to do for our headless embedded systems) sshd is the only way to talk to my target machines.

Brian "Pronounces obscenity-laden curses on the Religionists who decided we don't need kernel source packages anymore"

******************************************************************* This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the system manager. This footnote also confirms that this email message has been swept for the presence of computer viruses. www.Hubbell.com - Hubbell Incorporated**

On Wednesday, January 19, 2011 12:35:18 pm Drew Weaver wrote:

The kernel boots fine, and everything works ok until you unplug the monitor from the DVI port on the motherboard.

When you unplug the monitor, that IRQ/ACPI message is displayed, and it screws up the USB and the e1000 card in the system.

These machines aren't always headless, sometimes we need to plug monitors into them.

Can you disable the video card's use of an IRQ? I've seen that before, where the video card had an IRQ whether the driver needed it or not.

If you use a DVI to VGA/analog adapter and unplug an analog monitor, does it still happen?