2.4.4 Oops

2.4.4 Oops

[Frank van Maarseveen]

System is a dual PIII. Oops occurred while starting the automounter
(autofs).  starting it later on by hand again gave no oops anymore.

ksymoops 2.3.5 on i686 2.4.4-x40.  Options used
     -V (default)
     -k /proc/ksyms (default)
     -l /proc/modules (default)
     -o /lib/modules/2.4.4-x40/ (default)
     -m /boot/System.map-2.4.4-x40 (specified)

Apr 30 16:20:53 iapetus kernel: Unable to handle kernel paging request at virtual address 08058000 
Apr 30 16:20:53 iapetus kernel: c021b5f6 
Apr 30 16:20:53 iapetus kernel: *pde = 0894e067 
Apr 30 16:20:53 iapetus kernel: Oops: 0000 
Apr 30 16:20:53 iapetus kernel: CPU:    1 
Apr 30 16:20:53 iapetus kernel: EIP:    0010:[<c021b5f6>] 
Using defaults from ksymoops -t elf32-i386 -a i386
Apr 30 16:20:53 iapetus kernel: EFLAGS: 00010206 
Apr 30 16:20:53 iapetus kernel: eax: 00000000   ebx: 00001000   ecx: 00000276   edx: 00000000 
Apr 30 16:20:53 iapetus kernel: esi: 08058000   edi: c86e1628   ebp: 080579d8   esp: cbdd5f64 
Apr 30 16:20:53 iapetus kernel: ds: 0018   es: 0018   ss: 0018 
Apr 30 16:20:54 iapetus kernel: Process mount (pid: 416, stackpage=cbdd5000) 
Apr 30 16:20:54 iapetus kernel: Stack: 00001000 c86e1000 080579d8 00001000 c0145d99 cbdd4000 c0138c0d c86e1000  
Apr 30 16:20:54 iapetus kernel:        080579d8 00001000 cbdd4000 c0ed0000 08057a38 bffffa88 c0138f15 080579d8  
Apr 30 16:20:54 iapetus kernel:        cbdd5fbc cbdd4000 c0ed0000 08057a38 00000000 c0106e0c 00000000 c0106cdb  
Apr 30 16:20:54 iapetus kernel: Call Trace: [<c0145d99>] [<c0138c0d>] [<c0138f15>] [<c0106e0c>] [<c0106cdb>]  
Apr 30 16:20:54 iapetus kernel: Code: f3 a5 89 c1 f3 a4 89 cb eb 16 89 d9 c1 e9 02 31 c0 f3 ab f6  

>>EIP; c021b5f6 <__generic_copy_from_user+3a/64>   <=====
Trace; c0145d99 <dput+19/154>
Trace; c0138c0d <copy_mount_options+51/9c>
Trace; c0138f15 <sys_mount+15/120>
Trace; c0106e0c <error_code+34/3c>
Trace; c0106cdb <system_call+33/38>
Code;  c021b5f6 <__generic_copy_from_user+3a/64>
00000000 <_EIP>:
Code;  c021b5f6 <__generic_copy_from_user+3a/64>   <=====
   0:   f3 a5                     repz movsl %ds:(%esi),%es:(%edi)   <=====
Code;  c021b5f8 <__generic_copy_from_user+3c/64>
   2:   89 c1                     movl   %eax,%ecx
Code;  c021b5fa <__generic_copy_from_user+3e/64>
   4:   f3 a4                     repz movsb %ds:(%esi),%es:(%edi)
Code;  c021b5fc <__generic_copy_from_user+40/64>
   6:   89 cb                     movl   %ecx,%ebx
Code;  c021b5fe <__generic_copy_from_user+42/64>
   8:   eb 16                     jmp    20 <_EIP+0x20> c021b616 <__generic_copy_from_user+5a/64>
Code;  c021b600 <__generic_copy_from_user+44/64>
   a:   89 d9                     movl   %ebx,%ecx
Code;  c021b602 <__generic_copy_from_user+46/64>
   c:   c1 e9 02                  shrl   $0x2,%ecx
Code;  c021b605 <__generic_copy_from_user+49/64>
   f:   31 c0                     xorl   %eax,%eax
Code;  c021b607 <__generic_copy_from_user+4b/64>
  11:   f3 ab                     repz stosl %eax,%es:(%edi)
Code;  c021b609 <__generic_copy_from_user+4d/64>
  13:   f6 00 00                  testb  $0x0,(%eax)


-- 
Frank

Re: 2.4.4 Oops

[Brian Gerst]

Frank van Maarseveen wrote:
> 
> System is a dual PIII. Oops occurred while starting the automounter
> (autofs).  starting it later on by hand again gave no oops anymore.

[snip]

> Code;  c021b5f6 <__generic_copy_from_user+3a/64>   <=====
>    0:   f3 a5                     repz movsl %ds:(%esi),%es:(%edi)   <=====

There should be no way for this to cause an oops.  There should be an
exception handler here that will catch the page fault and deal with it. 
It appears that the exception table might be corrupted.  What compiler
are you using?

--

				Brian Gerst

Re: 2.4.4 Oops

[Frank van Maarseveen]

On Mon, Apr 30, 2001 at 11:07:21AM -0400, Brian Gerst wrote:
> > Code;  c021b5f6 <__generic_copy_from_user+3a/64>   <=====
> >    0:   f3 a5                     repz movsl %ds:(%esi),%es:(%edi)   <=====
> 
> There should be no way for this to cause an oops.  There should be an
> exception handler here that will catch the page fault and deal with it. 
> It appears that the exception table might be corrupted.  What compiler
> are you using?

iapetus ~ gcc -v
Reading specs from /usr/lib/gcc-lib/i386-redhat-linux/egcs-2.91.66/specs
gcc version egcs-2.91.66 19990314/Linux (egcs-1.1.2 release)

(from RH 6.1)

-- 
Frank