Re: occurs segmentation fault, when I make a file system(ext3 or reiserfs)

[Munjun Kang]

Dear Eric.

Thanks for your reply.
Follows are "ksymoops" decoding message.

Regards,
malas

----------------- ksymoops decoding message -----------------

c022ade5
*pde = 00017063
Oops: 0002
CPU:    0
EIP:    0010:[<c022ade5>]    Not tainted
EFLAGS: 00010246
eax: 00000000   ebx: c58bc000   ecx: 00001000   edx: 00000000
esi: 0806acb8   edi: c58bc000   ebp: c109f2c8   esp: c1089f34
ds: 0018   es: 0018   ss: 0018
Process mkreiserfs (pid: 127, stackpage=c1089000)
Stack: 00000000 00001000 00001000 c124f054 00000000 00001000 00000000 00000000 
       011de000 00000000 c109f220 00000000 00000001 0806acb8 c109f28c c109f220 
       c13bbaa0 c01290df c13bbaa0 0806acb8 00001000 c13bbac0 0806acb8 c13bbaa0 
Call Trace:    [<c01290df>] [<c0133c65>] [<c0108d43>]
Code: f3 aa 58 59 e9 9f df ef ff b8 f2 ff ff ff 30 d2 e9 b7 14 f0
> > EIP; c022ade5 <rwsem_down_failed_common+1175/2d00>   <=====

Trace; c01290df <generic_file_write+6f/b0>
Trace; c0133c65 <sys_write+85/100>
Trace; c0108d43 <system_call+33/40>

Code;  c022ade5 <rwsem_down_failed_common+1175/2d00>
00000000 <_EIP>:
Code;  c022ade5 <rwsem_down_failed_common+1175/2d00>   <=====
   0:   f3 aa                     repz stos %al,%es:(%edi)   <=====
Code;  c022ade7 <rwsem_down_failed_common+1177/2d00>
   2:   58                        pop    %eax
Code;  c022ade8 <rwsem_down_failed_common+1178/2d00>
   3:   59                        pop    %ecx
Code;  c022ade9 <rwsem_down_failed_common+1179/2d00>
   4:   e9 9f df ef ff            jmp    ffefdfa8 <_EIP+0xffefdfa8> c0128d8d <do_generic_file_write+3fd
/6e0>
Code;  c022adee <rwsem_down_failed_common+117e/2d00>
   9:   b8 f2 ff ff ff            mov    $0xfffffff2,%eax
Code;  c022adf3 <rwsem_down_failed_common+1183/2d00>
   e:   30 d2                     xor    %dl,%dl
Code;  c022adf5 <rwsem_down_failed_common+1185/2d00>
  10:   e9 b7 14 f0 00            jmp    f014cc <_EIP+0xf014cc> c112c2b1 <END_OF_CODE+e4ce6d/????

1 warning issued.  Results may not be reliable.

----- Original Message ----- 
From: "Eric W. Biederman" 
To: "Munjun Kang" 
Cc: 
Sent: Saturday, November 23, 2002 4:02 PM
Subject: Re: occurs segmentation fault, when I make a file system(ext3 or reiserfs)
> "Munjun Kang"  writes:
> > Dear linuxbios.
> >
> > Hello.
> >
> > I've worked for development of Embedded linux project in my corp.
> > Following is decription of my projcet.
> > CPU : VIA C3 800Mhz (EBGA)
> > Reference Board : VT6015
> > Northbridge : VT8606 (include S3 twister video core)
> > Southbridge : VT8231 (include VIA rhine II LAN core)
> > Bootbios : Linuxbios
> > OS : linux 2.4.19 ( ramdisk include glibc 2.3.1, busybox 0.60.5 and tinylogin
> > 1.2 )
> >
> > Now, It works well except hard disk.
> > I try to boot the kernel in two way.
> > First, it loads compressed kernel and ramdisk image via network using tftp
> > protocol.
> >
> > Second, it loads elfImage, generated by 'mkelfImage' via network.
> > All of them work great. (It means that I can see the prompt '\ '.)
> >
> > But, when I try to make a file system the harddisk, 'segmentation fault' occurs.
> >
> > mke2fs and mkreiserfs have same problem.
> >
> > For find out the cause, I try to make a file system in various method.
> > Buildin Bios + Redhat 8.0 installed harddisk + test harddisk
> > Linuxbios + Redhat 8.0 installed harddisk + test harddisk
> > Linuxbios + Kernel(newly compiled), own ramdisk + test harddisk
> >
> > After boot, I try to make a file system on test harddisk.
> >
> > In case 1, It works well. Of course, it must be.
> > In case 2 and 3, I meet the segmentation fault.
> > So, I think that this problem caused in Linuxbios.
> >
> > Maybe, this problem caused by harddisk initialization in linuxbios.
> > But, there are no harddisk initialization code in linuxbios.
> >
> > Cound you help me?
> >
> > Thanks for reading.
> > Regards,
> > malas
> >
> > PS. In addition, memtest86 v3.0 works well without error.
>
> My hunch is that it is a DMA problem, or a similar hardware timing
> problem.  But you missed the most important part of the kernel crash.
> The symbolic output of ksymoops.
>
> Without knowing where the code crashed it is hard to guess where the
> code crashed in the kernel, and thus have a good idea what the real
> problem is.
>
> Eric
>
> Linuxbios mailing list
> [e-mail]
> http://www.clustermatic.org/mailman/listinfo/linuxbios