oops on 2.4.21-pre3 plus csets to 20030124, smp, 1.9.9pre3

[tyme]

Unable to handle kernel NULL pointer dereference at virtual address 00000017
*pde = 00000000
Oops: 0000
CPU: 0
EIP: 0010:[<c02ed319>] Not tainted
Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00010206
eax: e9458b40 ebx: e9458aa0 ecx: 0000000f edx: d690a000
esi: 00000000 edi: f5dc8b40 ebp: f6962000 esp: f6963e40
ds: 0018 es: 0018 ss: 0018
Process named (pid: 13439, stackpage=f6963000)
Stack: 00000000 00000000 f6943cc0 d4dccac0 f6963e74 26954264 c02450b1 f5dc8b40
00000010 26954264 263289e4 00000010 0000000d 76e70002 07000001 f6963f28
f6963f44 f6963f28 00000020 00000000 00000000 00002611 000000ca f695def4
Call Trace: [<c02450b1>] [<c012513c>] [<c0105cac>] [<c0116429>] [<c0245d19>] [<c01077af>]
Code: 8b 41 08 8b 40 08 85 c0 74 de 80 b8 10 01 00 00 00 74 d5 c7


>>EIP; c02ed319 <gr_attach_curr_ip+e9/1a0> <=====

Trace; c02450b1 <sys_accept+101/150>
Trace; c012513c <kill_something_info+10c/1f0>
Trace; c0105cac <__switch_to+3c/e0>
Trace; c0116429 <schedule+2b9/510>
Trace; c0245d19 <sys_socketcall+d9/270>
Trace; c01077af <system_call+33/50>

Code; c02ed319 <gr_attach_curr_ip+e9/1a0> 00000000 <_EIP>:
Code; c02ed319 <gr_attach_curr_ip+e9/1a0> <=====
0: 8b 41 08 mov 0x8(%ecx),%eax <=====
Code; c02ed31c <gr_attach_curr_ip+ec/1a0>
3: 8b 40 08 mov 0x8(%eax),%eax
Code; c02ed31f <gr_attach_curr_ip+ef/1a0>
6: 85 c0 test %eax,%eax
Code; c02ed321 <gr_attach_curr_ip+f1/1a0>
8: 74 de je ffffffe8 <_EIP+0xffffffe8>
Code; c02ed323 <gr_attach_curr_ip+f3/1a0>
a: 80 b8 10 01 00 00 00 cmpb $0x0,0x110(%eax)
Code; c02ed32a <gr_attach_curr_ip+fa/1a0>
11: 74 d5 je ffffffe8 <_EIP+0xffffffe8>
Code; c02ed32c <gr_attach_curr_ip+fc/1a0>
13: c7 00 00 00 00 00 movl $0x0,(%eax)

[spender]

does the oops occur on a 2.4.20 machine? If it doesn't, then you must have patched something incorrectly.

-Brad

[tyme]

The patch applies with only trivial rejects... but yeah it could be some patch interaction issue. That's the first oops in about a week and only on one of two machines though, so it'd be fairly difficult to be sure that it's a 2.4.21-pre change that caused it.