ACPI screwing up USB ehci_hcd with kernel 2.6.16.18

ACPI screwing up USB ehci_hcd with kernel 2.6.16.18

[Paul Surgeon]

I had a problem with ACPI IRQ routing messing up the USB ehci module.
I have to disable the irq routing with apci=noirq in the bootloader config in 
order to use the ehci module.

Failing to do so results in :
usb 5-1: new high speed USB device using ehci_hcd and address 3
ehci_hcd 0000:00:10.4: Unlink after no-IRQ? Controller is probably using the 
wrong IRQ.

System :
Kernel 2.6.16.18
Chipset VIA KT400A

Regards
Paul

Re: ACPI screwing up USB ehci_hcd with kernel 2.6.16.18

[Sergio Monteiro Basto]

and before ?
can you send lspci -n ?


On Sun, 2006-05-28 at 21:50 +0200, Paul Surgeon wrote:
> I had a problem with ACPI IRQ routing messing up the USB ehci module.
> I have to disable the irq routing with apci=noirq in the bootloader config in 
> order to use the ehci module.
> 
> Failing to do so results in :
> usb 5-1: new high speed USB device using ehci_hcd and address 3
> ehci_hcd 0000:00:10.4: Unlink after no-IRQ? Controller is probably using the 
> wrong IRQ.
> 
> System :
> Kernel 2.6.16.18
> Chipset VIA KT400A
> 
> Regards
> Paul
> -
> To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: ACPI screwing up USB ehci_hcd with kernel 2.6.16.18

[Sergio Monteiro Basto]

[-- Attachment #1: Type: text/plain, Size: 2365 bytes --]

Hi,
Some of yours hardware have same IDs than mine 

you have as relevant ;
cat /usr/include/linux/pci_ids.h  |  grep 3189
#define PCI_DEVICE_ID_VIA_8377_0        0x3189
#define PCI_DEVICE_ID_VIA_82C586_1      0x0571
#define PCI_DEVICE_ID_VIA_82C586_2      0x3038
#define PCI_DEVICE_ID_VIA_8237          0x3227
#define PCI_DEVICE_ID_NVIDIA_GEFORCE4_TI_4800_8X    0x0281

I have 
PCI_DEVICE_ID_VIA_82C586_1 
PCI_DEVICE_ID_VIA_82C586_2
PCI_DEVICE_ID_VIA_8237_SATA 

On kernel 2.6.16.17 enter this patch which, if you don't use kernel-rc
version doesn't change nothing :D
http://www.kernel.org/diff/diffview.cgi?file=%2Fpub%2Flinux%2Fkernel%
2Fv2.6%2Fincr%2Fpatch-2.6.16.16-17.bz2;z=12

if you compile your kernel, you could try remove this 2 lines, 
DECLARE_PCI_FIXUP_ENABLE(PCI_VENDOR_ID_VIA, PCI_DEVICE_ID_VIA_82C586_1,
quirk_via_irq); 
DECLARE_PCI_FIXUP_ENABLE(PCI_VENDOR_ID_VIA, PCI_DEVICE_ID_VIA_82C586_2,
quirk_via_irq); 
to see if ACPI IRQ routing works better: 

Well, this can be done like this 
cd drives/pci
edit and delete the lines on quirks.c
cd ../..
make bzImage
(backup mv /boot/vmlinuz-2.6.16.18 /boot/vmlinuz-2.6.16.18.orig)
mv arch/x86_64/boot/bzImage /boot/vmlinuz-2.6.16.18)
and reboot 

My question which is not meaningfully, because I made a little
confusion, but is if you has used other version of kernel before ?  

Maybe the best is open o bug in bugzilla with a complete dmesg and one
lspci -n , etc.

I am little curious, why your interrupts are XT-PIC and not IO-APIC

Thanks for the report ,
On Tue, 2006-05-30 at 22:26 +0200, Paul Surgeon wrote:
> On Monday 29 May 2006 14:11, Sergio Monteiro Basto wrote:
> > and before ?
> > can you send lspci -n ?
> 
> I can send you a lot more than that.  :)
> I attached a zip file containing 10 log files which hopefully won't be an 
> issue with the mailing list. (a single uncompressed dmesg log is 15KB so 
> hopefully an 18KB zip file won't be a problem)
> 
> There are two versions of each log file. One before I added acpi=noirq to the 
> boot options and one after I added acpi=noirq to the boot options.
> 
> logs :
> dmesg (dmesg after bootup - USB storage device not plugged in)
> dmesg-usb-plugin  (dmesg output after plugging in USB storage device)
> interrupts
> lspci
> lspci-n
> 
> Hope that helps a bit
> Paul

[-- Attachment #2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 2166 bytes --]

Re: ACPI screwing up USB ehci_hcd with kernel 2.6.16.18

[Bjorn Helgaas]

On Sunday 28 May 2006 13:50, Paul Surgeon wrote:
> I had a problem with ACPI IRQ routing messing up the USB ehci module.
> I have to disable the irq routing with apci=noirq in the bootloader config in 
> order to use the ehci module.
> 
> Failing to do so results in :
> usb 5-1: new high speed USB device using ehci_hcd and address 3
> ehci_hcd 0000:00:10.4: Unlink after no-IRQ? Controller is probably using the 
> wrong IRQ.
> 
> System :
> Kernel 2.6.16.18
> Chipset VIA KT400A

Can you open a bug report at http://bugzilla.kernel.org in the ACPI
category?  Please attach the complete dmesg log both with and without
"acpi=noirq".

Thanks,
  Bjorn

Re: ACPI screwing up USB ehci_hcd with kernel 2.6.16.18

[Bjorn Helgaas]

On Monday 05 June 2006 10:45, Bjorn Helgaas wrote:
> On Sunday 28 May 2006 13:50, Paul Surgeon wrote:
> > I had a problem with ACPI IRQ routing messing up the USB ehci module.
> > I have to disable the irq routing with apci=noirq in the bootloader config in 
> > order to use the ehci module.
> > 
> > Failing to do so results in :
> > usb 5-1: new high speed USB device using ehci_hcd and address 3
> > ehci_hcd 0000:00:10.4: Unlink after no-IRQ? Controller is probably using the 
> > wrong IRQ.
> > 
> > System :
> > Kernel 2.6.16.18
> > Chipset VIA KT400A
> 
> Can you open a bug report at http://bugzilla.kernel.org in the ACPI
> category?  Please attach the complete dmesg log both with and without
> "acpi=noirq".

Did you resolve this problem?  If you can collect some information, we
can probably fix the problem.

Bjorn

Re: ACPI screwing up USB ehci_hcd with kernel 2.6.16.18

[Paul Surgeon]

On Friday 09 June 2006 23:34, Bjorn Helgaas wrote:
> On Monday 05 June 2006 10:45, Bjorn Helgaas wrote:
> > On Sunday 28 May 2006 13:50, Paul Surgeon wrote:
> > > I had a problem with ACPI IRQ routing messing up the USB ehci module.
> > > I have to disable the irq routing with apci=noirq in the bootloader
> > > config in order to use the ehci module.
> > >
> > > Failing to do so results in :
> > > usb 5-1: new high speed USB device using ehci_hcd and address 3
> > > ehci_hcd 0000:00:10.4: Unlink after no-IRQ? Controller is probably
> > > using the wrong IRQ.
> > >
> > > System :
> > > Kernel 2.6.16.18
> > > Chipset VIA KT400A
> >
> > Can you open a bug report at http://bugzilla.kernel.org in the ACPI
> > category?  Please attach the complete dmesg log both with and without
> > "acpi=noirq".
>
> Did you resolve this problem?  If you can collect some information, we
> can probably fix the problem.
>
> Bjorn

Well my resolution is to run with "acpi=noirq". I'm not sure if you'd call 
that a resolution but I can live with it. The main reason why I reported this 
problem was in case someone else runs into it in the future.

I attached three sets of dmesg and interrupt logs to the bug tracker :
http://bugzilla.kernel.org/show_bug.cgi?id=6654
If you need more info let me know.

Thanks
Paul