[OOPS] 100% repeatable OOPS, 2.5.61-66, NFS and reiserfs

[OOPS] 100% repeatable OOPS, 2.5.61-66, NFS and reiserfs

[David Ford]

1. Power loss this morning
2. Fixed filesystems (reiserfstools is fscking useless on root filesystems)
3. Now server OOPSes when nfs client tries to stat/read files/dirs

Unable to handle kernel NULL pointer dereference at virtual address 00000000
 printing eip:
00000000
*pde = 00000000
Oops: 0000
CPU:    0
EIP:    0060:[<00000000>]    Not tainted
EFLAGS: 00010202
EIP is at 0x0
eax: 00000000   ebx: 00000006   ecx: c03efb28   edx: c827be94
esi: c83ad014   edi: dfd16e00   ebp: c827beac   esp: c827be70
ds: 007b   es: 007b   ss: 0068
Process nfsd (pid: 1188, threadinfo=c827a000 task=df3c2740)
Stack: c01ce6b8 dfd16e00 c827bea0 c827be94 c01ad070 dfff04f4 00000006 
c83ad004
       11270000 00000e07 000008dd 0000004d 0002f4f2 00000e07 01214b98 
c827bef4
       c01ad4c7 dfd16e00 c83ad014 00000006 00000006 c01ad070 dfff04f4 
c83ad004
Call Trace:
 [<c01ce6b8>] reiserfs_decode_fh+0xb4/0xc0
 [<c01ad070>] nfsd_acceptable+0x0/0x114
 [<c01ad4c7>] fh_verify+0x343/0x4f8
 [<c01ad070>] nfsd_acceptable+0x0/0x114
 [<c01ae694>] nfsd_access+0x28/0xf4
 [<c01b4ec9>] nfsd3_proc_access+0xc1/0xd0
 [<c01ab4a3>] nfsd_dispatch+0xc3/0x18f
 [<c033b5cd>] svc_process+0x3ed/0x67a
 [<c01ab200>] nfsd+0x23c/0x41c
 [<c01aafc4>] nfsd+0x0/0x41c
 [<c0107145>] kernel_thread_helper+0x5/0xc


Code:  Bad EIP value.
 <1>Unable to handle kernel NULL pointer dereference at virtual address 
00000000
 printing eip:
00000000
*pde = 00000000
Oops: 0000
CPU:    1
EIP:    0060:[<00000000>]    Not tainted
EFLAGS: 00010202
EIP is at 0x0
eax: 00000000   ebx: 00000006   ecx: c03efb28   edx: c8269e94
esi: c83a0014   edi: dfd16e00   ebp: c8269eac   esp: c8269e70
ds: 007b   es: 007b   ss: 0068
Process nfsd (pid: 1189, threadinfo=c8268000 task=df5ba720)
Stack: c01ce6b8 dfd16e00 c8269ea0 c8269e94 c01ad070 dfff04f4 00000006 
c83a0004
       11270000 00000e07 000008dd 0000004d 0002f4f2 00000e07 01214b98 
c8269ef4
       c01ad4c7 dfd16e00 c83a0014 00000006 00000006 c01ad070 dfff04f4 
c83a0004
Call Trace:
 [<c01ce6b8>] reiserfs_decode_fh+0xb4/0xc0
 [<c01ad070>] nfsd_acceptable+0x0/0x114
 [<c01ad4c7>] fh_verify+0x343/0x4f8
 [<c01ad070>] nfsd_acceptable+0x0/0x114
 [<c01ae694>] nfsd_access+0x28/0xf4
 [<c01b4ec9>] nfsd3_proc_access+0xc1/0xd0
 [<c01ab4a3>] nfsd_dispatch+0xc3/0x18f
 [<c033b5cd>] svc_process+0x3ed/0x67a
 [<c01ab200>] nfsd+0x23c/0x41c
 [<c01aafc4>] nfsd+0x0/0x41c
 [<c0107145>] kernel_thread_helper+0x5/0xc

(repeats several times)

Server is now useless and needs to be rebooted.  I've shutdown my NFS 
clients in order to run the server minimally, it's a mail/web server.

Assistance would be appreciated.

Re: [OOPS] 100% repeatable OOPS, 2.5.61-66, NFS and reiserfs

[Neil Brown]

On Tuesday April 8, david+powerix@blue-labs.org wrote:
> 1. Power loss this morning
> 2. Fixed filesystems (reiserfstools is fscking useless on root filesystems)
> 3. Now server OOPSes when nfs client tries to stat/read files/dirs
> 
> Unable to handle kernel NULL pointer dereference at virtual address 00000000
>  printing eip:
> 00000000
> *pde = 00000000
> Oops: 0000
> CPU:    0
> EIP:    0060:[<00000000>]    Not tainted

Development kernels are like that....

 This is a bug in the kernel which is triggered by using nfs-utils
 1.0.3

 Either upgrade to the latest 2.5 kernel, or downgrade nfs-utils until
 you can upgrade the kernel.

NeilBrown

Re: [OOPS] 100% repeatable OOPS, 2.5.61-66, NFS and reiserfs

[David Ford]

The latest 2.5 kernel being .67?

Yes, dev kernels are like that...  but I'm not going to get into my 
angst over 'release' kernels and NFS ;)

david

Neil Brown wrote:

>On Tuesday April 8, david+powerix@blue-labs.org wrote:
>  
>
>>1. Power loss this morning
>>2. Fixed filesystems (reiserfstools is fscking useless on root filesystems)
>>3. Now server OOPSes when nfs client tries to stat/read files/dirs
>>
>>Unable to handle kernel NULL pointer dereference at virtual address 00000000
>> printing eip:
>>00000000
>>*pde = 00000000
>>Oops: 0000
>>CPU:    0
>>EIP:    0060:[<00000000>]    Not tainted
>>    
>>
>
>Development kernels are like that....
>
> This is a bug in the kernel which is triggered by using nfs-utils
> 1.0.3
>
> Either upgrade to the latest 2.5 kernel, or downgrade nfs-utils until
> you can upgrade the kernel.
>
>NeilBrown
>  
>

Re: [OOPS] 100% repeatable OOPS, 2.5.61-66, NFS and reiserfs

[Neil Brown]

On Tuesday April 8, david+powerix@blue-labs.org wrote:
> The latest 2.5 kernel being .67?

Yep, .67 is recent enough.

NeilBrown

Re: [OOPS] 100% repeatable OOPS, 2.5.61-66, NFS and reiserfs

[David Ford]

Yes, it's fixed.

Thank you,
david

Neil Brown wrote:

>On Tuesday April 8, david+powerix@blue-labs.org wrote:
>  
>
>>The latest 2.5 kernel being .67?
>>    
>>
>
>Yep, .67 is recent enough.
>
>NeilBrown
>