Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

"Disabling IRQ #9" - how to check for impact

1,983 views
Skip to first unread message

Christian Britz

unread,
Apr 26, 2022, 3:30:10 AM4/26/22
to
Hello Debianists,

some days ago I updated the BIOS of my Lenovo IdeaPad S145-15IIL (had to
boot a certain proprietary OS for this). I think since then there is a
new error in the kernel log. I never noticed it before.

[ 9.967601] irq 9: nobody cared (try booting with the "irqpoll" option)
[ 9.967607] CPU: 1 PID: 981 Comm: sddm-greeter Tainted: G
OE 5.10.0-13-amd64 #1 Debian 5.10.106-1
[ 9.967608] Hardware name: LENOVO 81W8/LNVNB161216, BIOS DKCN54WW
01/27/2022
[ 9.967608] Call Trace:
[ 9.967615] dump_stack+0x6b/0x83
[...]
[ 9.967636] handlers:
[ 9.967639] [<00000000872fa119>] acpi_irq
[ 9.967640] Disabling IRQ #9

This IRQ seems to be related to ACPI on this machine. I am unsure, what
exactly gets diabled and what might be the impact. So far I notice no
problems with performance or overheating, compiling a small tool went as
always I would say.

Do you have any hints for me about what I should/could check?

Regards,
Christian

--
http://www.cb-fraggle.de

IL Ka

unread,
Apr 27, 2022, 9:00:05 PM4/27/22
to
Hi.

Short answer: 

Long answer:
This message means some hardware generates an interrupt request, but the interrupt handler (kernel driver) failed to process it.

As we see from stacktrace, this handler is "acpi_irq" (you can also check it by reading /proc/interrupts):

Many hardware things in laptops use ACPI: Brightness buttons, FN buttons, volume buttons, lid etc.

It seems that your latest firmware (or BIOS as you called it) is incompatible with your kernel (either kernel or firmware should be fixed)

This situation is common with Lenovo hardware:
https://bugzilla.kernel.org/show_bug.cgi?id=207749 (I suggest you to write to this issue and to follow it)

If everything works as expected (you see no problem with lid, buttons etc) simply ignore it.
If no, try to install the latest kernel and file a bug to Debian (I think it should be forwarded upstream and handled by the kernel developer, probably duplicating 207749).

You can also add "irqpoll" kernel param which will ask the kernel to check all IRQ handlers to find the one which can process it, but I am 99% sure this wouldn't help since
we know it should be processed by ACPI.

Christian Britz

unread,
Apr 28, 2022, 4:40:06 AM4/28/22
to
Hello Ilya,

thank you for sharing so many interesting details!

On 2022-04-28 02:53 UTC+0200, IL Ka wrote:

> This is a known kernel
> bug: https://bugzilla.kernel.org/show_bug.cgi?id=207749
> <https://bugzilla.kernel.org/show_bug.cgi?id=207749> 

I was almost sure the messages first appeared after the firmware update,
but the kernel bug is much older. I will definitely follow this thread.

> As we see from stacktrace, this handler is "acpi_irq" (you can also
> check it by reading /proc/interrupts):

There is a high number on CPU1:

9: 0 819113116 0 0 0 0
0 0 IR-IO-APIC 9-fasteoi acpi

Should I be worried about that?

> Many hardware things in laptops use ACPI: Brightness buttons, FN
> buttons, volume buttons, lid etc.

They seem to work.
> If everything works as expected (you see no problem with lid, buttons
> etc) simply ignore it.

It seems so.

> If no, try to install the latest kernel and file a bug to Debian (I

Latest kernel from backports did not help.

> You can also add "irqpoll" kernel param which will ask the kernel to

I think I have read somewhere that this can make the machine very slow.

So far I notice no impact of the bug, luckily. I guess I will live with
it and hope for a fix in a kernel of a later Debian release.

Best Regards,
Christian

--
http://www.cb-fraggle.de

Christian Britz

unread,
May 3, 2022, 3:40:05 AM5/3/22
to
Problem seems to be gone with latest Debian stable kernel update! I
don't see the message anymore with 5.10.113.
0 new messages