2.4.19pre6+preempt problem...

2.4.19pre6+preempt problem...

[Andrea Aime]

Hi everybody,
while exiting from X windows yesterday I got the following error (the system 
didn't crash, thought):

waiting for X server to shut down kernel BUG at page_alloc.c:108!
invalid operand: 0000
CPU:    0
EIP:    0010:[<c0131c90>]    Tainted: P
EFLAGS: 00013282
eax: 00000000   ebx: c10c77a8   ecx: c10c77a8   edx: c528a9b0
esi: 0016f000   edi: 00000000   ebp: 08048000   esp: c387feb8
ds: 0018   es: 0018   ss: 0018
Process X (pid: 15554, stackpage=c387f000)
Stack: c6418b40 c68f12c0 c01286bf c6418b40 c68f12c0 0805fd5c 00000000 00041000
       0016f000 c4e14224 08048000 c012737e c10c77a8 00000029 00000000 081b7000
       c3f30080 00000000 081b7000 c3f30080 c0115fb3 c6418b40 c68f12c0 0805fd5c
Call Trace: [<c01286bf>] [<c012737e>] [<c0115fb3>] [<c0129df8>] [<c011833a>]
   [<c011ce16>] [<c0107f00>] [<c0108a9b>]

Code: 0f 0b 6c 00 5a 49 28 c0 89 d8 2b 05 10 02 33 c0 69 c0 a3 8b
 <3>X[15554] exited with preempt_count 1

To be sure I restarted the system, but I haven't notice any other problem so 
far... (btw, I'm using NVidia latest drivers, compiled the kernel on a 
Mandrake 8.2 with gcc 2.96 compiler).
Best regards
Andrea Aime

Re: 2.4.19pre6+preempt problem...

[Christian Bornträger]

Andrea Aime wrote:
>while exiting from X windows yesterday I got the following error (the system 
[...]
> EIP:    0010:[<c0131c90>]    Tainted: P
[...]
> so far... (btw, I'm using NVidia latest drivers, compiled the kernel on a

And NVIDIA is the reason why propably nobody here can really help you. We 
don't have the source we don't know what the NVIDIA-module is doing.

Pipe your BUG through ksymoops and send the report to NVIDIA.

greetings

Christian

Re: 2.4.19pre6+preempt problem...

[Robert Love]

On Thu, 2002-04-18 at 08:26, Andrea Aime wrote:
>  <3>X[15554] exited with preempt_count 1
> 
> To be sure I restarted the system, but I haven't notice any other problem so 
> far... (btw, I'm using NVidia latest drivers, compiled the kernel on a

Two things:

(a) ask nvidia.  binary modules => no help here

(b) it is not a preemption problem.  that preemption error message
    is because X died while holding a lock and exiting with a nonzero
    lock count prints a warning.

	Robert Love

Re: 2.4.19pre6+preempt problem...

[Denis Vlasenko]

On 18 April 2002 10:26, Andrea Aime wrote:
> while exiting from X windows yesterday I got the following error (the
> system didn't crash, thought):
>
> waiting for X server to shut down kernel BUG at page_alloc.c:108!
> invalid operand: 0000
> CPU:    0
> EIP:    0010:[<c0131c90>]    Tainted: P
> EFLAGS: 00013282
> eax: 00000000   ebx: c10c77a8   ecx: c10c77a8   edx: c528a9b0
> esi: 0016f000   edi: 00000000   ebp: 08048000   esp: c387feb8
> ds: 0018   es: 0018   ss: 0018
> Process X (pid: 15554, stackpage=c387f000)
> Stack: c6418b40 c68f12c0 c01286bf c6418b40 c68f12c0 0805fd5c 00000000
> 00041000 0016f000 c4e14224 08048000 c012737e c10c77a8 00000029 00000000
> 081b7000 c3f30080 00000000 081b7000 c3f30080 c0115fb3 c6418b40 c68f12c0
> 0805fd5c Call Trace: [<c01286bf>] [<c012737e>] [<c0115fb3>] [<c0129df8>]
> [<c011833a>] [<c011ce16>] [<c0107f00>] [<c0108a9b>]
>
> Code: 0f 0b 6c 00 5a 49 28 c0 89 d8 2b 05 10 02 33 c0 69 c0 a3 8b
>  <3>X[15554] exited with preempt_count 1

Please decode it with ksymoops.
It not that important for BUGs but for real dirty oopses
ksumoops is a must.
--
vda