I found a very interesting oops, from the "random pointer corruption"
series, on a 486 running 2.0.32. The machine is a (rather busy) lpd server
and is also spooling print jobs through samba; since some printers can
only do appletalk, the machines is also running netatalk.
Here it is.. Notice the interesting value in EAX ("??==").
Unable to handle kernel paging request at virtual address ff3f3d65
current->tss.cr3 = 01ff2000, ^Xr3 =
01ff2000
*pde = 00000000
Oops: 0000
CPU: 0
EIP: 0010:[__release_sock+15/144]
EFLAGS: 00010202
eax: 3f3f3d3d ebx: 0137a018 ecx: 0137a204 edx: 000000f5
esi: 0137a0d8 edi: 00000000 ebp: 00000001 esp: 0102bf2c
ds: 0018 es: 0018 fs: 002b gs: 002b ss: 0018
Process lpd (pid: 25200, process nr: 29, stackpage=0102b000)
Stack: 0137a018 0014a983 0137a018 0137a018 01d69b90 01d69b90 00155437 0137a018
00000000 01d69b00 00000000 00139f0c 01d69b90 00000000 01d69b00 01d69b00
01d69b00 bfffeea8 0013a185 01d69b90 00b18f24 00123dbc 01d69b00 00b18f24
Call Trace: [tcp_close+531/560] [inet_release+103/112] [sock_release+92/160]
[sock_close+37/48] [__fput+28/64] [close_fp+88/112]
[sys_close+68/80] [system_call+85/124]
Code: 83 78 28 00 74 71 ff 05 14 39 19 00 90 8d 83 8c 00 00 00 8b
Code: 00000000 <_EIP>:
Code: 0 <_EIP+0>: 83 78 28 00 cmpl $0x0,0x28(%eax)
Code: 4 <_EIP+4>: 74 71 je 77 <_EIP+0x77>
Code: 6 <_EIP+6>: ff 05 14 39 19 incl 0x193914
Code: 11 <_EIP+11>: 00
Code: 12 <_EIP+12>: 90 nop
Code: 13 <_EIP+13>: 8d 83 8c 00 00 leal 0x8c(%ebx),%eax
Code: 18 <_EIP+18>: 00
Code: 19 <_EIP+19>: 8b 00 movl (%eax),%eax
Kernel version:
Linux version 2.0.32 (root@) (gcc version 2.7.2.1) #3 Wed Nov 19 11:58:37 EST 1997
Hardware: 486, 32 MB RAM, 2 IDE drives on a Promise VLB IDE controller,
3c509 ISA NIC.
The kernel is custom compiled for this machine, with 3c509 and appletalk
as modules, everything else statically linked into the kernel. In the
networking section I have enabled multicasting, IP accounting and large
windows. I don't have any accounting rules enabled, though.
As far as debugging goes, I am willing to apply Alan's "highwinds" TCP
bug detection patch. Just let me know..
Thanks,
Ionut
--
It is better to keep your mouth shut and be thought a fool,
than to open it and remove all doubt.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu