I sporadically get this crash while running "arp" under both 2.5.46
and 2.5.47.
ksymoops 2.4.6 on i586 2.5.47. Options used
-v /usr/src/linux/vmlinux (specified)
-k /proc/ksyms (default)
-l /proc/modules (default)
-o /lib/modules/2.5.47/ (default)
-m /usr/src/linux/System.map (default)
kernel BUG at net/core/skbuff.c:178!
invalid operand: 0000
CPU: 0
EIP: 0060:[<c01d1be7>] Not tainted
Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00010286
eax: 0000003a ebx: c3e54ce0 ecx: 00000000 edx: c024c69c
esi: 000001d0 edi: 00000000 ebp: 00000018 esp: c055de94
ds: 0068 es: 0068 ss: 0068
Stack: c023df60 c01d1044 c3e54ce0 c3e548c0 c01d1044 00000001 000001d0 00000000
c0214497 c3e54ce0 00000001 00000000 000001d0 7fffffff c3e54ce0 fffffff4
c3743620 c3743620 c055df00 0000006e bfffeda8 c01cfa46 c3743620 c055df00
Call Trace: [<c01d1044>] [<c01d1044>] [<c0214497>] [<c01cfa46>] [<c01894d2>] [<c01d02fb>] [<c0108d67>]
Code: 0f 0b b2 00 c7 d2 23 c0 5b 58 e9 6c fe ff ff 8d 76 00 8d bc
>>EIP; c01d1be7 <alloc_skb+1b7/1d0> <=====
>>ebx; c3e54ce0 <_end+3b7772c/6524aac>
>>edx; c024c69c <log_wait+0/8>
>>esp; c055de94 <_end+2808e0/6524aac>
Trace; c01d1044 <sock_wmalloc+24/50>
Trace; c01d1044 <sock_wmalloc+24/50>
Trace; c0214497 <unix_stream_connect+97/340>
Trace; c01cfa46 <sys_connect+56/80>
Trace; c01894d2 <copy_from_user+32/3c>
Trace; c01d02fb <sys_socketcall+8b/1b0>
Trace; c0108d67 <syscall_call+7/b>
Code; c01d1be7 <alloc_skb+1b7/1d0>
00000000 <_EIP>:
Code; c01d1be7 <alloc_skb+1b7/1d0> <=====
0: 0f 0b ud2a <=====
Code; c01d1be9 <alloc_skb+1b9/1d0>
2: b2 00 mov $0x0,%dl
Code; c01d1beb <alloc_skb+1bb/1d0>
4: c7 d2 23 c0 5b 58 mov $0x585bc023,%edx
Code; c01d1bf1 <alloc_skb+1c1/1d0>
a: e9 6c fe ff ff jmp fffffe7b <_EIP+0xfffffe7b> c01d1a62 <alloc_skb+32/1d0>
Code; c01d1bf6 <alloc_skb+1c6/1d0>
f: 8d 76 00 lea 0x0(%esi),%esi
Code; c01d1bf9 <alloc_skb+1c9/1d0>
12: 8d bc 00 00 00 00 00 lea 0x0(%eax,%eax,1),%edi
<0>Kernel Panic: Aiee, killing interrupt handler!
On 13 Nov 02 at 14:15, Maciej Babinski wrote:
> I sporadically get this crash while running "arp" under both 2.5.46
> and 2.5.47.
Try 2.5.47-ac2, it contains fix for arp problems. It should
fix also your problem, as this call trace was generated by
BUG() in alloc_skb triggered by in_interrupt() != 0.
Petr Vandrovec
> >>EIP; c01d1be7 <alloc_skb+1b7/1d0> <=====