Hello,
I'm experimenting a 100% reproducable oops, ksymoops filtered
below, when trying to connect via ssh to the 2.4.2-ac2 machine.
(openssh 2.1.1p1)
ksymoops 2.3.7 on i686 2.4.2-ac2. Options used
-V (default)
-k /proc/ksyms (default)
-l /proc/modules (default)
-o /lib/modules/2.4.2-ac2/ (default)
-m /boot/System.map (specified)
Unable to handle kernel NULL pointer dereference at virtual address 00000000
cff761e0
*pde = 00000000
Oops: 0002
CPU: 0
EIP: 0010:[<cff761e0>]
Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00010246
eax: 00000000 ebx: 00000000 ecx: 00000000 edx: cff761e0
esi: 00000000 edi: 00000000 ebp: c140b2a0 esp: cd2b1f98
ds: 0018 es: 0018 ss: 0018
Process sshd (pid: 156, stackpage=cd2b1000)
Stack: 00000003 ce2fc680 c1407cc0 00000009 00000001 ce239000 cd2b0000 00000003
bfffe238 bffff310 c0108fc3 4016c03b bfffe238 40172ba4 00000003 bfffe238
bffff310 00000063 0000002b 0000002b 00000063 4011d691 00000023 00000246
Call Trace: [<c0108fc3>]
Code: c0 2b 4a c1 a0 b2 40 c1 60 61 f7 cf ec 60 f7 cf 50 f6 f7 cf
>>EIP; cff761e0 <_end+fd642b0/10612130> <=====
Trace; c0108fc3 <system_call+33/38>
Code; cff761e0 <_end+fd642b0/10612130>
00000000 <_EIP>:
Code; cff761e0 <_end+fd642b0/10612130> <=====
0: c0 2b 4a shrb $0x4a,(%ebx) <=====
Code; cff761e3 <_end+fd642b3/10612130>
3: c1 a0 b2 40 c1 60 61 shll $0x61,0x60c140b2(%eax)
Code; cff761ea <_end+fd642ba/10612130>
a: f7 (bad)
Code; cff761eb <_end+fd642bb/10612130>
b: cf iret
Code; cff761ec <_end+fd642bc/10612130>
c: ec in (%dx),%al
Code; cff761ed <_end+fd642bd/10612130>
d: 60 pusha
Code; cff761ee <_end+fd642be/10612130>
e: f7 (bad)
Code; cff761ef <_end+fd642bf/10612130>
f: cf iret
Code; cff761f0 <_end+fd642c0/10612130>
10: 50 push %eax
Code; cff761f1 <_end+fd642c1/10612130>
11: f6 f7 div %bh,%al
Code; cff761f3 <_end+fd642c3/10612130>
13: cf iret
Arnaud.
Le Fri, Feb 23, 2001 at 02:47:45PM +0100, Arnaud S . Launay a ?crit:
> I'm experimenting a 100% reproducable oops, ksymoops filtered
> below, when trying to connect via ssh to the 2.4.2-ac2 machine.
> (openssh 2.1.1p1)
uhuh -- just seen I'm using a gcc snapshot of today. this should
explain. (funny to see "3.1" when asking gcc --version).
Anyway, latest gcc seems to broke kernel once again.
Arnaud.