Re: OOPS on reboot with 2.1.20
Tue, 7 Jan 1997 03:00:38 -0500

I get an almost identical oops with a 486dx2-50 when I C-A-D with 2.1.20:
Unable to handle kernel NULL pointer dereference at virtual address
current->tss.cr3 = 004fd000, %cr3 = 04fd000
*pde = 00000000
Oops: 0002
CPU: 0
EIP: 0010: [<c0109875>]
EFLAGS: 00010246
eax: c0250000 ebx: 00000000 ecx: 28121969 edx: 01234567
esi: ffffffff edi: 08048e50 ebp: bffffda4 esp: c07b2fa4
ds: 0018 es: 0018 ss: 0018
Process reboot (pid: 207, process nr: 5, stackpage = c07b2000)
Stack: 00000000 ffffffff 08048e50 bffffda4 00000000 c0141a1 c025ec0c
fea1dead 28121969 01234567 ffffffff 08048e50 bffffda4 00000058
0000002b 00000058 4007642c 00000023 00000282 bffffda0 0000002b
Call trace: [<c01141a1>] [<c010a868>]
Code: 66 a3 72 04 00 00 bd bc a7 00 00 8b 15 18 56 1d c0 89 e8 f7
/etc/rc.d/rc.6: line 89: 207 Segmentation fault $command -f
INIT: no more processes left in this runlevel

BTW I am using the modutils-970104 with kernel patch

Opcodes are the same as reported by Mark A. Breuer, but addresses and pid
understandably differ.

After recording this oops, I powered-down and restarted with 2.0.27 with no

I later tried 2.1.20, with similar results - the oops showed the same codes,
but some
different addresses and different pid

I am also using a Sanyo cdrom - had to re-arrange some code in isp16.c and
to move module dependant code into #ifdef MODULE / #endif, since I compile
cdrom driver into the kernel, not as a module

Hope this helps

Darrell A Escola
Clovis, California USA