Kernel hosed or Nvidia modules?

Kernel hosed or Nvidia modules?

[Matthias Andree]

Is that a genuine kernel bug (2.4.19-pre2-ac3 here) or Yet Another
Nvidia Driver Bug (would not be the first bug to bite me in their
driver)? The Nvidia driver is configured to use the kernel AGP rather
than the Nvidia AGP. (TNT board, FWIW).

This does not happen frequently. Does anyone have a contact address at
Nvidia so I can show them as well?

NVRM: AGPGART: freed 16 pages
NVRM: AGPGART: backend released
kernel BUG at page_alloc.c:131!
invalid operand: 0000
CPU:    0
EIP:    0010:[__free_pages_ok+325/788]    Tainted: P
EFLAGS: 00013286
eax: 00000020   ebx: c1216aa8   ecx: c02d6d00   edx: 00005917
esi: 00001000   edi: 00000000   ebp: 4042a000   esp: cbae7ef0
ds: 0018   es: 0018   ss: 0018
Process X (pid: 1064, stackpage=cbae7000)
Stack: c02782f0 00000083 c1216aa8 00001000 00000000 4042a000 00000001 c1216aa8
       00001000 c1216aa8 00001000 c012c3a9 c012c87f c1216aa8 c01200f9 c1216aa8
       cbd2e0a8 c012055b 0a483067 caa8c800 00001000 4002a000 00000000 cbaec404
Call Trace: [__free_pages+29/32] [free_page_and_swap_cache+51/56] [__free_pte+73/80] [zap_page_range+423/592] [do_munmap+443/568] [sys_munmap+53/84] [system_call+51/64]
kernel:
Code: 0f 0b 83 c4 08 8d b6 00 00 00 00 8b 43 18 24 eb 89 43 18 c6

ksymoops said "before first symbol".

-- 
Matthias Andree

Re: Kernel hosed or Nvidia modules?

[J Sloan]

To be fair, I've had far less trouble with
nvidia drivers than with some of the
open source dri ones -

But it's best if you can reproduce the bug
without the nvidia driver loaded, to remove
all doubt.

If 3D graphics support is needed on that
box, perhaps a  cheap voodoo3 would
get you going in the meantime.  :-)

Hopefully some nvidia folks lurk on the list,
in case it is somehow related...

Joe

Alan Cox wrote:

>>Is that a genuine kernel bug (2.4.19-pre2-ac3 here) or Yet Another
>>Nvidia Driver Bug (would not be the first bug to bite me in their
>>
>
>Only Nvidia can tell you that. Please don't bother this list with oopses
>when binary modules are loaded. Nobody can help you
>-
>To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>the body of a message to majordomo@vger.kernel.org
>More majordomo info at  http://vger.kernel.org/majordomo-info.html
>Please read the FAQ at  http://www.tux.org/lkml/
>

Re: Kernel hosed or Nvidia modules?

[Alan Cox]

> Is that a genuine kernel bug (2.4.19-pre2-ac3 here) or Yet Another
> Nvidia Driver Bug (would not be the first bug to bite me in their

Only Nvidia can tell you that. Please don't bother this list with oopses
when binary modules are loaded. Nobody can help you

Re: Kernel hosed or Nvidia modules?

[andrew may]

On Fri, Mar 29, 2002 at 05:51:51PM -0800, J Sloan wrote:
> Hopefully some nvidia folks lurk on the list,
> in case it is somehow related...

I really dobt they do, but they do have have forum at
nvnews.net and have even posted patches there.

Re: Kernel hosed or Nvidia modules?

[Zwane Mwaikambo]

On Sat, 30 Mar 2002, Matthias Andree wrote:

> Is that a genuine kernel bug (2.4.19-pre2-ac3 here) or Yet Another
> Nvidia Driver Bug (would not be the first bug to bite me in their
> driver)? The Nvidia driver is configured to use the kernel AGP rather
> than the Nvidia AGP. (TNT board, FWIW).
> 
> This does not happen frequently. Does anyone have a contact address at
> Nvidia so I can show them as well?
> 
> NVRM: AGPGART: freed 16 pages
> NVRM: AGPGART: backend released
> kernel BUG at page_alloc.c:131!

Using the kernel AGP driver with the TNT and 440BX seems to hose things 
pretty badly in my experience, either use the NvAGP or go sans AGP, 
ideally you don't want the kernel AGP loaded at all.

	Zwane

--
http://function.linuxpower.ca