GP with 2.0.31pre9

Dan Hollis (goemon@sasami.anime.net)
Tue, 16 Sep 1997 19:09:22 -0700 (PDT)


Got the following with 2.0.31pre9.

Running a threaded webserver called Roxen. Apparently cgi's are making the
kernel GP. I don't think this is supposed to happen ;-)

-Dan

general protection: 0000
CPU: 0
EIP: 0010:[<0012d448>]
EFLAGS: 00010286
eax: f000ef6f ebx: 00b54810 ecx: 00000000 edx: 00000076
esi: 00000000 edi: f000ef6f ebp: 00fb6810 esp: 01091f7c
ds: 0018 es: 0018 fs: 002b gs: 002b ss: 0018
Process cgi (pid: 2835, process nr: 63, stackpage=01091000)
Stack: 00000009 00122413 00b54810 00000000 00000009 000000b2 00000006 0011678a
00000000 00b54810 ffffffff fffffffc 00000000 00116872 00000000 0010a5f5
00000000 00000000 00001590 ffffffff fffffffc 00000000 ffffffda 0000002b
Call Trace: [<00122413>] [<0011678a>] [<00116872>] [<0010a5f5>]
Code: 8b 50 50 85 d2 74 22 f6 42 20 01 74 0f 53 83 c0 50 50 e8 15
Using `../System.map' to map addresses to symbols.

>>EIP: 12d448 <locks_remove_locks+c/38>
Trace: 122413 <close_fp+37/5c>
Trace: 11678a <do_exit+112/1ec>
Trace: 116872 <sys_exit+e/10>
Trace: 10a5f5 <system_call+55/80>

Code: 12d448 <locks_remove_locks+c/38> movl 0x50(%eax),%edx
Code: 12d44b <locks_remove_locks+f/38> testl %edx,%edx
Code: 12d44d <locks_remove_locks+11/38> je 12d471 <locks_remove_locks+35/38>
Code: 12d44f <locks_remove_locks+13/38> testb $0x1,0x20(%edx)
Code: 12d453 <locks_remove_locks+17/38> je 12d464 <locks_remove_locks+28/38>
Code: 12d455 <locks_remove_locks+19/38> pushl %ebx
Code: 12d456 <locks_remove_locks+1a/38> addl $0x50,%eax
Code: 12d459 <locks_remove_locks+1d/38> pushl %eax
Code: 12d45a <locks_remove_locks+1e/38> call 9090002c <_EIP+9090002c>
Code: 12d45f <locks_remove_locks+23/38> nop