kernel:Disabling IRQ - why, and.. so what...?

[morgano]

Occasionally i recieve in all terminals the message

Code: Select all

Message from syslogd@svarten at Nov 29 15:44:16 ...
 kernel:Disabling IRQ #17

Trying to understand, i used

Code: Select all

$ cat /proc/interrupts
            CPU0       CPU1       CPU2       CPU3       CPU4       CPU5       CPU6       CPU7                               
-snipped by me-
  17:     100001          0          0          0          0          0          0          0   IO-APIC-fasteoi   firewire_ohci
-snipped by me-


I searched the net some, and found users having some device shut off when its interrupt got disabled.
But i do not have anything connected to firewire at all...

I use latest bios. And manufacturer made it impossible to downgrade so i can not test other versions

Fully updated mga1 x86_64, kernel 2.6.38.8-desktop-8.mga

[doktor5000]

You should look at the dmesg output about that, there's not enough context information currently.
Could you upload your dmesg output as an attachment here, please?

[morgano]

Thanks doktor, Iĺl try to catch it next time. (above post was from old capture i saved to investigate "later")

[morgano]

From my systemlog the minute it threw the message today:

Code: Select all

Dec 15 09:45:52 svarten kernel: irq 17: nobody cared (try booting with the "irqpoll" option)
Dec 15 09:45:52 svarten kernel: Pid: 0, comm: swapper Tainted: P            2.6.38.8-desktop-8.mga #1
Dec 15 09:45:52 svarten kernel: Call Trace:
Dec 15 09:45:52 svarten kernel: <IRQ>  [<ffffffff810b923b>] ? __report_bad_irq.clone.2+0x2b/0xa0
Dec 15 09:45:52 svarten kernel: [<ffffffff810b94ce>] ? note_interrupt+0x18e/0x1d0
Dec 15 09:45:52 svarten kernel: [<ffffffff810ba2f5>] ? handle_fasteoi_irq+0xb5/0xe0
Dec 15 09:45:52 svarten kernel: [<ffffffff810057c2>] ? handle_irq+0x22/0x40
Dec 15 09:45:52 svarten kernel: [<ffffffff8100544d>] ? do_IRQ+0x5d/0xe0
Dec 15 09:45:52 svarten kernel: [<ffffffff813ef8d3>] ? ret_from_intr+0x0/0x15
Dec 15 09:45:52 svarten kernel: <EOI>  [<ffffffff8131c6a2>] ? poll_idle+0x42/0x80
Dec 15 09:45:52 svarten kernel: [<ffffffff8131c673>] ? poll_idle+0x13/0x80
Dec 15 09:45:52 svarten kernel: [<ffffffff8131c787>] ? cpuidle_idle_call+0xa7/0x250
Dec 15 09:45:52 svarten kernel: [<ffffffff81001266>] ? cpu_idle+0xa6/0xf0
Dec 15 09:45:52 svarten kernel: [<ffffffff813d7f05>] ? rest_init+0x75/0x80
Dec 15 09:45:52 svarten kernel: [<ffffffff81745cae>] ? start_kernel+0x3fa/0x405
Dec 15 09:45:52 svarten kernel: [<ffffffff8174532b>] ? x86_64_start_reservations+0x132/0x136
Dec 15 09:45:52 svarten kernel: [<ffffffff81745435>] ? x86_64_start_kernel+0x106/0x115
Dec 15 09:45:52 svarten kernel: handlers:
Dec 15 09:45:52 svarten kernel: [<ffffffffa02c3100>] (irq_handler+0x0/0x540 [firewire_ohci])
Dec 15 09:45:52 svarten kernel: Disabling IRQ #17

And attached is my /var/log/dmesg ... got "The extension is not allowed" so i added ".txt"

[doktor5000]

Suspected something like that, maybe you can find some hints into debugging/fixing this at the following links:
https://bugzilla.redhat.com/show_bug.cgi?id=457648
The "irqpoll" boot option mentioned above, did you try that?
FWIW, from looking at your dmesg output, irq17 seems to be in some relation to a device driven by xhci_hcd,
which is a USB3.0 host controller driver. Do you have/use USB3.0 devices? Does it work without those?
Strangely later on firewire_ohci is shown to have that same relation, also HDA intel ...
By any chance, if you switch away from fglrx driver to radeon or vesa, do you still get those messages?

Nevertheless, you should report this as a bug against the kernel.

[morgano]

I have USB3 ports, but nothing ever attached there.
And i do not have any USB3 device, but now when i plugged in a USB stick there it could mount OK.
(- but OTOH maybe it is routed to a USB2 chip anyway?)

How is firewire_ohci related?
I have a firewire port,
(EDIT) and dvgrab could capture a video on that port (/Edit)

Tried vesa and radeon drivers att install tests but IIRC the current is the only that worked. (first install, default settings ended in black screen)

I have now added irqpoll to standard boot line in grub
Saw in logs it have been at least two days between a couple occasions, but i will report back if i see this again.

[doktor5000]

I have USB3 ports, but nothing ever attached there.
And i do not have any USB3 device, but now when i plugged in a USB stick there it could mount OK.
(- but OTOH maybe it is routed to a USB2 chip anyway?)

How is firewire_ohci related?

Sorry, don't know, maybe they just share an IRQ, that's what can be read in your dmesg output:

xhci_hcd 0000:03:00.0: PCI INT A -> GSI 17 (level, low) -> IRQ 17
xhci_hcd 0000:03:00.0: setting latency timer to 64
xhci_hcd 0000:03:00.0: xHCI Host Controller
xhci_hcd 0000:03:00.0: new USB bus registered, assigned bus number 2
[...]
usb usb2: No SuperSpeed endpoint companion for config 1 interface 0 altsetting 0 ep 129: using minimum values
usb usb2: New USB device found, idVendor=1d6b, idProduct=0003
usb usb2: New USB device strings: Mfr=3, Product=2, SerialNumber=1
usb usb2: Product: xHCI Host Controller
usb usb2: Manufacturer: Linux 2.6.38.8-desktop-8.mga xhci_hcd
usb usb2: SerialNumber: 0000:03:00.0
xHCI xhci_add_endpoint called for root hub
xHCI xhci_check_bandwidth called for root hub
hub 2-0:1.0: USB hub found
hub 2-0:1.0: 4 ports detected
xhci_hcd 0000:06:00.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
[...]
firewire_ohci 0000:08:01.0: PCI INT A -> GSI 17 (level, low) -> IRQ 17
firewire_ohci: Added fw-ohci device 0000:08:01.0, OHCI v1.10, 4 IR + 8 IT contexts, quirks 0x11
fglrx: module license 'Proprietary. (C) 2002 - ATI Technologies, Starnberg, GERMANY' taints kernel.

[morgano]

I believe this is not related to any problem i actually experience, so i let it rest.
Have not seen it again yet.
Maybe a later kernel do not show this effect.

[doktor5000]

Well, in general this error message in search results goes back to like kernel-2.6.20something, so i don't think so.
Please remember, if you don't report such bugs (either at Mageia or directly at kernel bugzilla) they won't solve themselves.

[morgano]

I just do not know what to report yet...

[morgano]

Still no problem since last post - still having irqpoll in kernel boot line.
Now i removed it to see if the problem reappears...

[morgano]

Interesting - it seem to be related to hangup
viewtopic.php?f=8&t=1602

This morning it did not respond to keyboard nor mouse (This machine is almost always on (contributing to BOINC))
- until i gave it Ctrl-Alt-F12, Ctrl-Alt-F7, and could login from screen lock
Then after some minutes i wanted to Ctrl-Alt-F12, but then it never updated screen again - i had to pull the power.

[morgano]

Entered https://bugs.mageia.org/show_bug.cgi?id=4160