[BUG?] nvidia oops 2.6.8-rc3-mm2

[BUG?] nvidia oops 2.6.8-rc3-mm2

[Kevin O'Shea]

I hope its ok to post about this here, but I thought it might be a
kernel bug not nvidia.

This is the oops with the new 6111 driver (it worked fine on mm1).

Thanks,
Kevin


Unable to handle kernel NULL pointer dereference at virtual address 000000a9
 printing eip:
c050ed00
*pde = 00000000
Oops: 0002 [#1]
PREEMPT 
Modules linked in: nvidia
CPU:    0
EIP:    0060:[<c050ed00>]    Tainted: P   VLI
EFLAGS: 00010212   (2.6.8-rc3-mm2) 
EIP is at add_pin_to_irq+0x0/0x70
eax: 000000a9   ebx: 00000010   ecx: 00000080   edx: 00000020
esi: 00000010   edi: 00000000   ebp: 00000001   esp: c4667d98
ds: 007b   es: 007b   ss: 0068
Process modprobe (pid: 26835, threadinfo=c4667000 task=de4d7930)
Stack: c0114375 00000010 00000000 00000010 00000093 c4667000 00000292 dfe3f540 
       00000001 0001a900 01000000 00000010 00000010 00000000 00010000 c011246b 
       00000000 00000010 00000010 00000001 00000001 00000001 00000001 00000010 
Call Trace:
 [<c0114375>] io_apic_set_pci_routing+0x1e5/0x210
 [<c011246b>] mp_register_gsi+0xeb/0x180
 [<c01101a1>] acpi_register_gsi+0x61/0xc0
 [<c025bc41>] acpi_pci_irq_enable+0x100/0x15f
 [<c03774f8>] pcibios_enable_device+0x28/0x30
 [<c023bdab>] pci_enable_device_bars+0x2b/0x40
 [<c023bdde>] pci_enable_device+0x1e/0x30
 [<e17863a8>] nv_kern_probe+0x227/0x33f [nvidia]
 [<c018e003>] create_dir+0x153/0x1c0
 [<c023d7d2>] pci_device_probe_static+0x52/0x70
 [<c023d82b>] __pci_device_probe+0x3b/0x50
 [<c023d86c>] pci_device_probe+0x2c/0x50
 [<c028d72f>] bus_match+0x3f/0x70
 [<c028d859>] driver_attach+0x59/0x90
 [<c028dd55>] bus_add_driver+0xa5/0xd0
 [<c028e31f>] driver_register+0x2f/0x40
 [<c023daec>] pci_register_driver+0x5c/0x90
 [<e1a4305f>] nvidia_init_module+0x5f/0x3a0 [nvidia]
 [<c013157e>] sys_init_module+0x12e/0x280
 [<c0104135>] sysenter_past_esp+0x52/0x71
Code: 00 00 00 00 00 00 00 00 00 00 00 87 db 02 00 06 de 02 00 00 00
00 00 00 00 00 00 03 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 <00>
00 00 00 04 ed 50 c0 04 ed 50 c0 00 00 00 00 89 0e 00 00 e0

Re: [BUG?] nvidia oops 2.6.8-rc3-mm2

[Jesse Stockall]

On Sun, 2004-08-08 at 22:36, Kevin O'Shea wrote:
> I hope its ok to post about this here, but I thought it might be a
> kernel bug not nvidia.
> 

If the problem persists without the Nvidia driver, post the results
here, otherwise report them to Nvidia.

-- 
Jesse Stockall <stockall@magma.ca>

Re: nvidia oops 2.6.8-rc3-mm2

[Con Kolivas]

Kevin O'Shea writes:

> I hope its ok to post about this here, but I thought it might be a
> kernel bug not nvidia.
> 
> This is the oops with the new 6111 driver (it worked fine on mm1).
> 
> Thanks,
> Kevin

> Call Trace:
>  [<c0114375>] io_apic_set_pci_routing+0x1e5/0x210

I quote from akpm's announcement:

- If some devices mysteriously stop working, try booting with pci=routeirq.
  If that fixes it, please send a report, Cc'ing bjorn.helgaas@hp.com.  See
  remove-unconditional-pci-acpi-irq-routing.patch

That looks like it might help your issue too since the oops talks about pci 
routing.

Cheers,
Con

Re: nvidia oops 2.6.8-rc3-mm2

[Kevin O'Shea]

That fixed it.

I now have a new error in my dmesg however (thought I'd add it here
incase its related):

kobject_register failed for sata_via (-17)
[<c0232bdb>] kobject_register+0x5b/0x60
[<c028dd00>] bus_add_driver+0x50/0xd0
[<c028e31f>] driver_register+0x2f/0x40
[<c023daec>] pci_register_driver+0x5c/0x90
[<e10bb00f>] svia_init+0xf/0x1d [sata_via]
[<c013157e>] sys_init_module+0x12e/0x280
[<c0104135>] sysenter_past_esp+0x52/0x71


Thanks for the help so far!

Kevin O'Shea
MyUID.com

On Mon, 09 Aug 2004 13:28:13 +1000, Con Kolivas <kernel@kolivas.org> wrote:
> Kevin O'Shea writes:
> 
> > I hope its ok to post about this here, but I thought it might be a
> > kernel bug not nvidia.
> >
> > This is the oops with the new 6111 driver (it worked fine on mm1).
> >
> > Thanks,
> > Kevin
> 
> > Call Trace:
> >  [<c0114375>] io_apic_set_pci_routing+0x1e5/0x210
> 
> I quote from akpm's announcement:
> 
> - If some devices mysteriously stop working, try booting with pci=routeirq.
>   If that fixes it, please send a report, Cc'ing bjorn.helgaas@hp.com.  See
>   remove-unconditional-pci-acpi-irq-routing.patch
> 
> That looks like it might help your issue too since the oops talks about pci
> routing.
> 
> Cheers,
> Con
> 
>