"irq 10:nobody cared!" - gentoo-dev-sources-2.6.7

Souperman

Just upgraded from 2.6.5-gentoo-r1 to 2.6.7-gentoo and I get this in my dmesg:

NeddySeagoon · Posted: Mon Jun 21, 2004 6:42 pm Post subject:

Souperman,

What USB 2.0 devices do you have?
From your irq post of 2.6.5, I suspect none, since there were never any requests on irq 10.

ehci_hcd is for USB 2.0 devices, it is not needed for USB 1.0 or 1.1. Thats what ohci_hcd is for. If you are not using ehci_hcd, don't load it.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.

KpR2000 · n00b Joined: 18 Aug 2003 Posts: 55

Hi,

I got the same problem with irq12.

Here are some cuts from dmesg:

RAPUL · Posted: Wed Jun 23, 2004 8:58 am Post subject: Hmmm..

That it is not a problem.
If you don't have any device using irq10 then it is disabled. All the crap you see between the irq10 message and the disabling message is because of you compiled kernel with some level of debugging.
_________________
Entropy rulz world.
Redundancy sux.
World is full of redundancy.
World sux.

Souperman · Posted: Wed Jun 23, 2004 9:07 am Post subject:

@NedSeagoon: I only have a MS Optical Wheel Mouse which is probably not USB2.

Last night I tried to watch TV using tvtime and it segfaulted. This never happened with 2.6.5 and dmesg said "irq7: nobody cared!". My tv tuner was using irq7. I thought of unloading and then reloading the saa7134 module but as soon as I did 'rmmod saa7134' the system hardlocked. I think I'll go back to 2.6.5.

_________________
moo

phex · n00b Joined: 08 Jun 2002 Posts: 32

it could be a problem with 2.6.7 on nforce/nforce2 motherboards, because i have a A7N8X DLX-E, and receive the same error but on IRQ 5. If i load the ehci_hcd module more than twice on my system, it will halt, and force me to do a hard reboot. I never came across this problem when using 2.6.6, so I might be switching back to 2.6.6 whenever I get around to rebooting this machine.

gkmac · Posted: Wed Jun 23, 2004 4:52 pm Post subject:

I have an NForce 2 based motherboard and kernel 2.6.7 gives me that "irq 10: nobody cared" error along with a permanently black display when X starts (with my GeForce 5700 card).

Everyone who's complained about the nvidia video card black/corrupt X seems to be using an NForce 2 chipset (still to get definite confirmation of this). Could a so-called "improvement" in 2.6.7 regarding the NForce 2 chipset be causing these bugs?

PiniHadad · n00b Joined: 16 Jun 2003 Posts: 69

I'm having a similar issue also with an nForce2 (not sure which IRQ it was- I'm not at my machine now). I saw some mention of this on the LKML but don't know if there was any resolution.

NeddySeagoon · Posted: Wed Jun 23, 2004 6:40 pm Post subject:

gkmac,

I get the irq 10 nobody cared when I load ehci_hcd becuse I don't have any USB devices. Every else works normally. I have an nForce2 chip set with an AGP slot nVida card.

If several things share irq 10 (or any irq for that matter) and the kernel disables it, no other devices on that irq will work either.
The "nobody cared" happens when when an interrupt is raised but no service routine is installed.

Do not load the unused module(s) that are causing the nobody cared problem.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.

Souperman · Posted: Thu Jun 24, 2004 9:37 am Post subject:

I hear what you are saying but I don't believe that is the solution. Fact is that I could load ehci module just fine on 2.6.5. Whether I have any USB2 devices or not is irrelevant. I might get one tomorrow and then I must load the module.
_________________
moo

bengi · Posted: Thu Jun 24, 2004 1:35 pm Post subject:

Just a stab in the dark, but when you compiled your kernel did you compile it with sleep-inside-spinlock checking enabled which is in the section called kernel hacking. As I had a similar message, and when I recompiled my kernel without sleep-inside-spinlock checking the message went away.

Tankred · n00b Joined: 20 Aug 2003 Posts: 36 Location: NRW (Germany)

Souperman · Posted: Thu Jun 24, 2004 2:51 pm Post subject:

I re-used my 2.6.5 .config for 2.6.7 and I'm pretty sure sleep inside spinlock checking was already there in 2.6.5, so if it's enabled now, it would've been enabled in 2.6.5, which brings me back to why is it happening now and never under 2.6.5?

Nonetheless, I will check that out tomorrow, perhaps it'll make my tv tuner behave again.
_________________
moo

Paranoid · Posted: Thu Jun 24, 2004 5:37 pm Post subject:

I'm also getting the irq disabled message along with the error/debugging messages. I can say for a fact I have no kernel debugging options enabled. nforce2 board also-started after upgrading to gen-dev-sources-2.6.7, 2.6.5 worked fine.
_________________
A paranoid is someone who knows a little of what's going on.
William S. Burroughs

Isaiah · Guru Joined: 25 Feb 2003 Posts: 359

I have had this same problem with various sources also - I don't have this problem with love-sources-2.6.7_rc3-r1 however :?:

Isaiah · Guru Joined: 25 Feb 2003 Posts: 359

Souperman · Posted: Fri Jun 25, 2004 5:42 pm Post subject:

OK, I disabled sleep-inside-spinlock and it still happens. Back to 2.6.5-gentoo-r1 for now. Works like a charm.

_________________
moo

Joebel · Posted: Sat Jun 26, 2004 4:15 am Post subject:

WarpFlyght · n00b Joined: 04 May 2003 Posts: 38 Location: Atlanta, GA

Just for the sake of adding to the numbers, I'm experiencing the same issue here with an ASUS A7N8X Deluxe (nForce2), dealing with IRQ #5 being disabled. 2.6.7-gentoo-r6 has issues, but 2.6.5-gentoo-r1 works like a charm. I used make oldconfig to generate my 2.6.7 configuration based on the 2.6.5 config file. Unfortunately on my system I see a hard hang before the system is able to reach a login prompt (verified to not be caused by a particular process loading on boot).

Thanks for the link Joebel. It's good to know about an available workaround, even if it's not the most elegant.

Does anyone know if this issue has already been reported as a bug (to Gentoo's Bugzilla or otherwise)?
_________________
Help save the world. Adopt an unanswered post.
·····························
Gentoo Linux 1.4 no more. SGI Indy R4400 on 2.6.13-mipscvs

Souperman · Posted: Tue Jun 29, 2004 8:22 am Post subject:

Personally I don't usually bother with Bugzilla. It's just a horrible horrible mess and, looking at the GWN, every week there are more new bugs than closed bugs. In this case, the factors influencing my apathy are 1) the fact that 2.6.5 works just fine for me and 2), the fact that 2.6.7 doesn't offer anything over 2.6.5, so I'm not missing anything.

[insert obligatory "YMMV" stuff here]
_________________
moo

WarpFlyght · n00b Joined: 04 May 2003 Posts: 38 Location: Atlanta, GA

Yeah... Well, I'd look into reporting the issue upstream, but we're supposed to report bugs to the Gentoo devs and they'll send them upstream if necessary, etc etc etc... This issue has surely been encountered elsewhere, so we're probably in the clear, but it does seem like a good idea to get it reported to the Gentoo developers anyway, just in case. (After all, who knows... It could be the result of a Gentoo patch to the kernel.)

PiniHadad said there was discussion about it on the LKML, so that's probably a good sign that it's being taken care of. I suppose if I had more time I'd try to test some other kernel branches and see if I could work out what code is causing the problem... I don't, so hopefully the kernel maintaners do! ;)
_________________
Help save the world. Adopt an unanswered post.
·····························
Gentoo Linux 1.4 no more. SGI Indy R4400 on 2.6.13-mipscvs

Salemixu · n00b Joined: 01 Jul 2004 Posts: 72 Location: Spain

Hi all, i think i have a similar problem, when i power on my computer it happens the same, irg 10 (or 11): nobody cared, perhaps it can be of help what i'll say, i hope so :p

I tried disabling acpi and didn't work, changing some things from kernel as somw threads said, and didn't work, and i tried some more things without success. One day (not much as i have 4 or 5 days gentoo installed on this machine, i have since february my little gentoo server

but with kernel 2.4 )
it started working but the only thing i noticed is that the computer clock was reset, it was set at 1 january of 2002 :/ then i started trying some things and arrived at the conclusion that if i power on the machine and go to linux the irq problem happens, but if i go into windows and then soft reboot the machine and go to linux the problem doesn't happen and the integrated lan of nforce2 goes ok with fordeth drivers.
It's a very strange thing but it is the only way i managed to make it work, i don't want to change to a lower version of kernel :p

I saw in a post that there was a report in bugzilla that its name is: "gentoo-dev-sources 2.6.7 breaks forcedeth", perhaps is the same thing, but anyway, here you have the link to it:

https://bugs.gentoo.org/show_bug.cgi?id=54639

Any help to fix it is welcome

bye!!!

Isaiah · Guru Joined: 25 Feb 2003 Posts: 359

Cerement · Guru Joined: 14 Jun 2003 Posts: 404

This problem has come up several times on #gentoo. So far as everyone can tell, its being caused by some funky code in the ehci (USB 2.0) driver. When the ehci_hcd module gets initialized (built in or modprobe), it scans the bus and, for some reason, determines some IRQ is not in use and promptly disables it (in my case, it was the IRQ for my AGP slot -- no 3D accel). The only solution anyone's been able to find so far is to disable ehci completely (ie. set to 'N' in kernel config -- if you set to 'M' and have hotplug running, hotplug will try to load it anyways) and resorting to the default uhci (USB 1.0) or the legacy ohci drivers ...

carsten90 · n00b Joined: 30 Jun 2004 Posts: 21

Oh, I'm so happy, that I'm not the only one having this problem... I'm not as stupid as I thought

ehci disables IRQ 11 in my case. Thats probably the problem, why I cant get alsa working -> my AC97 controller is at IRQ 11.

I cannot use a USB 2.0 device, it causes a
usb 1-2: control timeout on ep0out
usb 1-2: device not accepting address 14, error -110

Well, but it notices the device....

Greetings

Carsten