kernel BUG at page_alloc.c:82

kernel BUG at page_alloc.c:82

[Fabian Svara]

Hi!
Recently, the battery of my wireless mouse went empty - not realizing
the problem was realted to the battery, i switched to a text console and
saw the following BUG report from the kernel:

kernel BUG at page_alloc.c:82
invalid operand: 0000
CPU:	0
EIP:	0010:[<c0125183>]    Tainted: P
EFLAGS:	00013286
eax: 0000001f	ebx: c14ce240	ecx: c01f4a00	edx: 007d1d3a
esi: c14ce240	edi: 0011a900	ebp: 00000000	esp: d350fe48
ds: 0018   es: 0018   ss: 0018
Process XFree86 (pid: 6240, stackpage=d350f000)
Stack: c01c90c1 00000052 c14ce240 c14ce240 0011a900 08800000 0011a900 c01261e5
       c023cee0 c023cee0 c0124351 c01259c8 0011a900 c0125da3 0011a900 c023cee0
       c14ce240 c01263e9 c14ce240 00400000 00043000 cb5db10c c011c01c 0011a900
Call Trace: [<c01261e5>] [<c0124351>] [<c01259c8>] [<c0125da3>] [<c01263e9>]
   [<c011c01c>] [<c011e2f1>] [<c010f445>] [<c011330d>] [<c01069d8>] [<c0147624>]

   [<c012ab67>] [<c0106032>] [<c0106c1c>] [<c0106b64>]

Code: 0f 0b 83 c4 08 8b 46 18 a8 40 74 11 6a 54 68 c1 90 1c c0 e8

Which, using ksymoops, resolves to

>>EIP; c0125183 <__free_pages_ok+93/29c>   <=====

>>ebx; c14ce240 <_end+12755cc/145e538c>
>>ecx; c01f4a00 <console_sem+0/14>
>>edx; 007d1d3a Before first symbol
>>esi; c14ce240 <_end+12755cc/145e538c>
>>edi; 0011a900 Before first symbol
>>esp; d350fe48 <_end+132b71d4/145e538c>

Trace; c01261e5 <swap_free+25/2c>
Trace; c0124351 <lru_cache_del+5/14>
Trace; c01259c8 <page_cache_release+2c/30>
Trace; c0125da3 <delete_from_swap_cache+47/50>
Trace; c01263e9 <free_swap_and_cache+69/88>
Trace; c011c01c <zap_page_range+1a4/244>
Trace; c011e2f1 <exit_mmap+b5/12c>
Trace; c010f445 <mmput+39/58>
Trace; c011330d <do_exit+95/1ec>
Trace; c01069d8 <do_signal+220/288>
Trace; c0147624 <ext3_release_file+14/1c>
Trace; c012ab67 <fput+af/d0>
Trace; c0106032 <sys_sigreturn+b6/e4>
Trace; c0106c1c <error_code+34/3c>
Trace; c0106b64 <signal_return+14/18>

Code;  c0125183 <__free_pages_ok+93/29c>
00000000 <_EIP>:
Code;  c0125183 <__free_pages_ok+93/29c>   <=====
   0:   0f 0b                     ud2a      <=====
Code;  c0125185 <__free_pages_ok+95/29c>
   2:   83 c4 08                  add    $0x8,%esp
Code;  c0125188 <__free_pages_ok+98/29c>
   5:   8b 46 18                  mov    0x18(%esi),%eax
Code;  c012518b <__free_pages_ok+9b/29c>
   8:   a8 40                     test   $0x40,%al
Code;  c012518d <__free_pages_ok+9d/29c>
   a:   74 11                     je     1d <_EIP+0x1d> c01251a0 <__free_pages_ok+b0/29c>
Code;  c012518f <__free_pages_ok+9f/29c>
   c:   6a 54                     push   $0x54
Code;  c0125191 <__free_pages_ok+a1/29c>
   e:   68 c1 90 1c c0            push   $0xc01c90c1
Code;  c0125196 <__free_pages_ok+a6/29c>
  13:   e8 00 00 00 00            call   18 <_EIP+0x18> c012519b <__free_pages_ok+ab/29c>

I don't know what triggered it, as it could've been displayed on that terminal for a long time already, and I noticed it only now...

I am running Linux 2.4.17 on a Pentium 3.

-Fabian Svara

Re: kernel BUG at page_alloc.c:82

[Christian Bornträger]

Fabian Svara wrote:
> EIP:	0010:[<c0125183>]    Tainted: P

You have the Binary-NVIDIA Driver loaded, haven't you?
If yes, please report this bug to Nvidia, since nobody here can help you 
because the main part of the driver is closed source and nobody knows, what 
the driver is doing with the memory.

Christian

Re: kernel BUG at page_alloc.c:82

[Rogier Wolff]

Christian [Borntr_ger] wrote:
> Fabian Svara wrote:
> > EIP:	0010:[<c0125183>]    Tainted: P
> 
> You have the Binary-NVIDIA Driver loaded, haven't you?

Would it be an idea to print the name of the module that (first)
tainted the kernel here? That would eliminate this "guessing".

				Roger. 

-- 
** R.E.Wolff@BitWizard.nl ** http://www.BitWizard.nl/ ** +31-15-2137555 **
*-- BitWizard writes Linux device drivers for any device you may have! --*
* There are old pilots, and there are bold pilots. 
* There are also old, bald pilots. 

Re: kernel BUG at page_alloc.c:82

[arjan]

In article <200205051706.TAA08782@cave.bitwizard.nl> you wrote:
> Christian [Borntr_ger] wrote:
>> Fabian Svara wrote:
>> > EIP:        0010:[<c0125183>]    Tainted: P
>> 
>> You have the Binary-NVIDIA Driver loaded, haven't you?
> 
> Would it be an idea to print the name of the module that (first)
> tainted the kernel here? That would eliminate this "guessing".

That would mean storing it in kernel memory and then people would object
to using kernel memory for that....

Anyway in this case there's no need.... this is a classic nvidia report;
just the same as all the other dozens.... I wonder what bug the latest
driver has to cause the exact same oops all over ;)

Re: kernel BUG at page_alloc.c:82

[Fabian Svara]

On Sun, 5 May 2002 18:53:38 +0200
Christian Borntr  <linux-kernel@borntraeger.net> wrote:

> Fabian Svara wrote:
> > EIP:	0010:[<c0125183>]    Tainted: P
> 
> You have the Binary-NVIDIA Driver loaded, haven't you?
> If yes, please report this bug to Nvidia, since nobody here can help
> you because the main part of the driver is closed source and nobody
> knows, what the driver is doing with the memory.
> 
> Christian

Yes, indeed. I will send the report to nvidia. I thought it wasn't
nvidia-realted as there was no mention of it in the ksymoops-resolved
report... But in fact, all problems involving there kernel I've had here
were somehow related to those drivers.

-Fabian Svara