Just executing lsof gives a segmentation fault.
This is 2.5.73-mm3 and reproducable on dual p3 and single p3-800.
Will try 2.5.74-mm1 now.

Thanks,

Jan

<1>Unable to handle kernel NULL pointer dereference at virtual address 
00000000
  printing eip:
c01431bd
*pde = 00000000
Oops: 0000 [#4]
PREEMPT SMP
CPU:    0
EIP:    0060:[<c01431bd>]    Not tainted VLI
EFLAGS: 00010046
EIP is at kfree+0x31/0x60
eax: 00000000   ebx: d5a8a740   ecx: 00000000   edx: c1000000
esi: 00000100   edi: 00000206   ebp: d7fd8f00   esp: c4c0ff40
ds: 007b   es: 007b   ss: 0068
Process lsof (pid: 7808, threadinfo=c4c0e000 task=d7cfd310)
Stack: d5a8a740 d29bf300 d745b210 c017528f 00000100 d29bf300 d29bf300 
d745b210
        c0158207 d745b210 d29bf300 d29bf300 00000000 d1451100 d29bf300 
d6fd0080
        c01581cc c0156d00 d29bf300 d1451100 d29bf300 d1451100 d1451100 
c4c0e000
Call Trace:
  [<c017528f>] seq_release_private+0x17/0x2c
  [<c0158207>] __fput+0x37/0xf4
  [<c01581cc>] fput+0x14/0x18
  [<c0156d00>] filp_close+0x50/0x5c
  [<c0156d87>] sys_close+0x7b/0xb0
  [<c0109073>] syscall_call+0x7/0xb

Code: 10 85 f6 74 4f 9c 5f fa 8b 15 58 62 3d c0 8d 86 00 00 00 40 c1 e8 
0c 8d 04 80 8b 4c c2 08 b8 00 e0 ff ff 21 e0 8b 40 10 c1 e0 02 <8b> 1c 
08 8b 13 3b 53 04 73 09 8d 43 10 89 34 90 eb 13 90 53 51