2.1.42: Oops when diald closes link

Henrik Storner (storner@osiris.ping.dk)
Sat, 31 May 1997 00:07:14 +0200 (MET DST)

After applying Linus' patch to the oops-printing routine, it is now
possible to see what goes wrong when diald closes a link. When the
link goes down, the default route through ppp0 is put back to
diald's sl0. This gives:

diald[225]: Closing down idle link.
diald[225]: Nonzero exit status (7) on command '/sbin/route add 172.25.99.202 metric 1 dev sl0'
diald[225]: Nonzero exit status (7) on command '/sbin/route add default metric 1 netmask 0.0.0.0 dev sl0'

which is what I usually get, but then:

Unable to handle kernel paging request at virtual address 21000407
current->tss.cr3 = 01724000, Lr3 = 01724000
*pde = 00000000
Oops: 0000
CPU: 0
EIP: 0010:[<c0120107>]
EFLAGS: 00010082
eax: c1b646dc ebx: c02711a0 ecx: 210003ff edx: c142efa0
esi: c1b645d0 edi: 00000202 ebp: bffffd54 esp: c171defc
ds: 0018 es: 0018 ss: 0018
Process diald (pid: 225, process nr: 25, stackpage=c171d000)
Stack: 00000001 c1b645d0 c1865488 bffffd54 0118d067 c1b646dc c1535000 c0148558
c1b645d0 c1b645d0 c1865410 c014839d c1b645d0 c1b645d0 c0163183 c1b645d0
c1865410 c148876c 00000000 c0150b35 c1865410 c1865410 c0163900 c1865410
Call Trace: [<c0148558>] [<c014839d>] [<c0163183>] [<c0150b35>] [<c0163900>] [<c014511f>] [<c01455c8>]
[<c0123c0f>] [<c0123c9a>] [<c0123d26>] [<c01095ca>]
Code: 8b 69 08 81 fd 2b 2f c3 a5 0f 85 fa 00 00 00 8b 69 0c 85 ed

Using `/System.map-2.1.42' to map addresses to symbols.

>>EIP: c0120107 <kfree+77/1e0>
Trace: c0148558 <kfree_skbmem+58/70>
Trace: c014839d <__kfree_skb+3d/50>
Trace: c0163183 <destroy_sock+93/250>
Trace: c0150b35 <packet_close+65/70>
Trace: c0163900 <inet_release+70/80>
Trace: c014511f <sock_release+1f/40>
Trace: c01455c8 <sock_close+38/40>
Trace: c0123c0f <__fput+1f/50>
Trace: c0123c9a <close_fp+5a/a0>
Trace: c0123d26 <sys_close+46/50>
Trace: c01095ca <system_call+3a/40>

Code: c0120107 <kfree+77/1e0>

This is on an AMD 5x86-133 (a fast 486) with 32 MB RAM,
NCR810 SCSI controller - the rest is pretty standard.
SLIP/slhc is compiled into kernel; ppp is a module autoloaded
by kerneld - it was still loaded after the Oops (auto-unloaded
a couple of minutes later).

-- 
Henrik Storner <storner@osiris.ping.dk>  http://eolicom.olicom.dk/~storner/ 
"The POP3 server service depends on the SMTP server service, which
 failed to start because of the following error:
 The operation completed successfully." -Windows NT Server v3.51