nfsd & xfsdump strike again

[Matteo Centonza <matteo@sif.it>]

Hi,

this oops come from 2.4.14-xfs-1 (XFS 1.0.2) and xfsdump-1.1.14-0.

Feb  5 20:07:20 embeh kernel: Unable to handle kernel NULL pointer dereference at virtual a
Feb  5 20:07:20 embeh kernel: 00000000
Feb  5 20:07:20 embeh kernel: *pde = 00000000
Feb  5 20:07:20 embeh kernel: Oops: 0000
Feb  5 20:07:20 embeh kernel: CPU:    0
Feb  5 20:07:20 embeh kernel: EIP:    0010:[agp_frontend_cleanup+0/96]    Not tainted
Feb  5 20:07:20 embeh kernel: EIP:    0010:[<00000000>]    Not tainted Using defaults from ksymoops -t elf32-i386 -a i386
Feb  5 20:07:20 embeh kernel: EFLAGS: 00210282
Feb  5 20:07:20 embeh kernel: eax: 00000000   ebx: df47d140   ecx: df47d59c   edx: c04a79c0
Feb  5 20:07:20 embeh kernel: esi: df47d540   edi: df47d140   ebp: df47d140   esp: ddf6be50
Feb  5 20:07:20 embeh kernel: ds: 0018   es: 0018   ss: 0018
Feb  5 20:07:20 embeh kernel: Process nfsd (pid: 721, stackpage=ddf6b000)
Feb  5 20:07:20 embeh kernel: Stack: c01887a4 de5d27c0 df47d540 ddf84e04 dee58000 c0188bf6 df47d140 ddf84e04 
Feb  5 20:07:20 embeh kernel:        00000002 ddfc3800 11270000 db33f300 dee581f0 00000002 c0188f19 dee58000 
Feb  5 20:07:20 embeh kernel:        ddf84e14 00000002 00000001 00000001 00000008 00000008 de1f9ccc ddf84e04 
Feb  5 20:07:20 embeh kernel: Call Trace: [nfsd_findparent+52/224] [find_fh_dentry+502/784]
Feb  5 20:07:20 embeh kernel: Call Trace: [<c01887a4>] [<c0188bf6>] [<c0188f19>] [<c03127d9
Feb  5 20:07:20 embeh kernel:    [<c0110e42>] [<c018f7ae>] [<c01915e0>] [<c0186d83>] [<c0323888>] [<c0186b21>] 
Feb  5 20:07:20 embeh kernel:    [<c0105523>] 
Feb  5 20:07:20 embeh kernel: Code:  Bad EIP value.

>>EIP; 00000000 Before first symbol
Trace; c01887a4 <nfsd_findparent+34/e0>
Trace; c0188bf6 <find_fh_dentry+1f6/310>
Trace; c0188f19 <fh_verify+209/3f0>
Trace; c0110e42 <schedule+2a2/3e0>
Trace; c018f7ae <nfsd3_proc_lookup+14e/160>
Trace; c01915e0 <nfs3svc_decode_diropargs+a0/110>
Trace; c0186d83 <nfsd_dispatch+d3/1a0>


xfsdump was doing incrementals.

Probably the last one from the ``conflicting'' interaction between nfsd 
and xfsdump ;-).

As a side note, searching with google i've noticed that similar Oops have 
a reference to agp_frontend_cleanup. Is that a coincidence?

Thanks for any insight.

Ciao,

-m