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

Paranoid · Posted: Thu Jul 01, 2004 6:41 pm Post subject:

Check this out:

https://bugs.gentoo.org/show_bug.cgi?id=54413
_________________
A paranoid is someone who knows a little of what's going on.
William S. Burroughs

Isaiah · Guru Joined: 25 Feb 2003 Posts: 359

martinbishop · n00b Joined: 16 Dec 2003 Posts: 66

I also have this problem with my nforce2 motherboard. It is a problem with ehci itself, and there are numerous patches on the kernels bugzilla, not sure if any really work. Supposidly it is fixed in the latest bk sources, but I'm not brave enough to try it.

I get that same nobody cared error, and forcedeth doesn't load (because ehci and ethernet share the same irq)
_________________
"I'm too busy doing nothing to care."

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

Where are the patches??? I'll try them!!!

martinbishop · n00b Joined: 16 Dec 2003 Posts: 66

You can try some of http://linux.bkbits.net:8080/linux-2.6/search/?expr=ehci&search=ChangeSet+comments these, or http://www.ussg.iu.edu/hypermail/linux/kernel/0406.1/1289.html that, a few others if you look
_________________
"I'm too busy doing nothing to care."

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

Yoho!!

I'll try tomorrow or past tomorrow when i have time, lots of thanks!!!

radulucian · Posted: Sun Jul 04, 2004 12:30 pm Post subject: watching this thread

i was preparing to update to gentoo-dev-sources 2.6.7-r8 and good thing i checked the forums as i also have an nforce2 mobo and if i would have lost the netcard it would have been BAD as it's a remotely hosted server and it's sunday
if someone figures 100% what causes the problem (is it USB support?!) and a solution please post it asap, because the 2.6.5 kernel is subject to several recent vulnerabilities

i am watching the thread.

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

I have not tried it, but by not loading on kernel ehci (putting n in menuconfig) should go fine when using the integrated lan, at least this is my case, ehci and forcedth share the same irq and when loading the kernel shuts down the irq.
Have you seen or know if in the r8 is that issue fixed? i use r6.

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

I compiled and booted a 2.6.7-gentoo-r8 kernel last night and observed the same behavior as seen with other 2.6.7-gentoo kernels I tested, using the same configuration as with earlier attempts (copied the .config file and ran make oldconfig).
_________________
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

Let's see if someday i am able to test the patches that are said above.

If you overwrite the .config and then make oldconfig , isn't it the same as a normal make && make modules_install with the .config overwritten?

pdr · l33t Joined: 20 Mar 2004 Posts: 618

Just to add to the list:
Did make oldconfig using my 2.6.5 gentoo-dev-sources config file building 2.6.7-r8. Mobo is intel D875PBZ and took default answers (usually N) for new options.

With ACPI enabled I get a panic right after printk "ACPI: Subsystem revision 20040326" that it is unable to handle kernel NULL pointer.

Without ACPI (APM instead) then right after Reiserfs has recognized my /dev/hdg boot drive (SATA) I get a unhandled IRQ 0x18, then it goes into this weird print some stuff, get another unrecognized 0x18, print some stuff, get another unrecognized 0x18, ...

Am sticking with 2.6.5 until madness is over

argasek · Posted: Mon Jul 05, 2004 10:22 pm Post subject:

I've got the same problem. It appered when I connected my CDROM and CD burner on the second cable on my 2nd controller (HightPoint 370).

Here's my log:

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

On my system the error specifically mentions USB. I'll skip most of the log and just give the important tidbits:

carsten90 · n00b Joined: 30 Jun 2004 Posts: 21

It is a USB problem!
Try it: dont load the module (compile it as a module if neccessary) and everything will go well - if we're talking about the same bug. I already said: look at /proc/pci, what IRQ USB uses. If its the same IRQ thats used by a device, thats not working and USB 2.0 dont work it is propably this bug.

The mm-sources fixed this bug, but it has some new ones:

- Java doesnt work with this kernel for me, and some others:
https://forums.gentoo.org/viewtopic.php?t=191188
- accessing some servers doesnt work:
https://forums.gentoo.org/viewtopic.php?t=193214

argasek · Posted: Fri Jul 09, 2004 12:49 pm Post subject:

USB problem? Don't think so. Look at my dump. It has nothing related to USB. I have also tried the latest, -mm6 patch (which, as I have read on gentoo bugzilla has the USB problem fixed). The message still appears... So it looks like it's a more general problem than just USB.
_________________
RLU #137109 | http://soundcloud.com/argasek

Isaiah · Guru Joined: 25 Feb 2003 Posts: 359

2.6.7-love7 working fine here with CONFIG_4KSTACKS disabled

Paranoid · Posted: Fri Jul 09, 2004 4:33 pm Post subject:

It's a double problem with both ACPI & ehci.
_________________
A paranoid is someone who knows a little of what's going on.
William S. Burroughs

rapsure · Posted: Sat Jul 10, 2004 3:43 am Post subject: USB, but don't know where

I don't think the entire problem is wrapped into the ehci_hcd, and ACPI. Everyone here is using the ohci_hcd also, and that seems to also be broken at the time in the 2.6.7 kernel. Maybe I will have to try the latest development snapshot. In the 2.6.7 kernel I was able to get the IRQ 10: don't care! message if I ran my scanner for enough time, and I wasn't running the other devices that are also using that irq. The emu10k1, eth0, and usb are on irq 10 for me. I think the problem was already there, but it showed it's teeth in the 2.6.7 kernel. I have a huntch that it is also in the 2.6.6 kernel. I say try that, and if it is there it is most likely being caused by some erroneous dynamically allocated memory. Supposedly by compiling usb as modules, and then haveing them loaded by autoload.d the usb ports will begin working. There are patches in the latest development for the ehci_hcd and acpi, I doubt that will fix it. Try the latest devlopment snapshot of the 2.6.7 kernel. :?:

_________________

Hindi ko naintindihan, pakiulit. Sometimes my code works.

Paranoid · Posted: Sat Jul 10, 2004 11:29 am Post subject:

If you read the bug report you will find that mm-sources >=mm4 work fine.
_________________
A paranoid is someone who knows a little of what's going on.
William S. Burroughs

argasek · Posted: Sun Jul 11, 2004 10:22 am Post subject:

gronau · Posted: Sun Aug 01, 2004 11:40 pm Post subject:

I was running in the same problems with gentoo-dev-sources-2.6.7-r11.
Activating APIC and IOAPIC solved all my problems. (ACPI is enabled, too).

Greetz
gronau

PS.: Since kernel 2.6.6/2.6.7 there is no danger activating APIC on nforce2 boards.

Souperman · Posted: Mon Aug 02, 2004 8:56 am Post subject:

I have ACPI, APIC and IOAPIC enabled already.

_________________
moo

pdr · l33t Joined: 20 Mar 2004 Posts: 618

I still cannot use ACPI, even after recompiling the DSDT - the interpreter is blowing up or something. I don't think it affected the "don't care" problem, but to do above I did apply the patch to allow replacing the DSDT (but .config is not using CONFIG_ACPI).

However, finally got 2.6.7-r11 running. I had to enable SCSI and use the PIIX SATA driver (and change my drives from hde/g to sda/b in grub and fstab). This is on the intel 875 chipset... Under 2.6.5 I notice that my SATA was using IRQ 0x18 (the one that wasn't caring)

(2.6.7 has alway worked fine on my T42p laptop)

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

Hey! I'm so happy!!!

I've upgraded the kernel to gentoo-dev-sources 2.6.7-r12 and activated IOAPIC (ACPI and APIC were activated in the last compilation :p ) and also installed the new Nvidia 61.11 drivers, and now i don't have any problem of noboy cared, so my Nforce2 integrated LAN now works perfectly

Perhaps it's a patch of the kernel, perhaps it's because using IOAPIC? the best thing is that for me now it runs perfect (at the moment :p)

Paranoid · Posted: Sun Aug 08, 2004 4:21 pm Post subject:

Still the same old shit here with ACPI, APIC & IOAPIC enabled. At least 2.6.5-r1 works....
_________________
A paranoid is someone who knows a little of what's going on.
William S. Burroughs