Re: cfservd seg faulting

[Lance Albertson]

On Thu, 2004-02-26 at 09:13, Mark.Burgess@iu.hio.no wrote:
> I would certainly recommend 2.1.3
> 
> M

Finally tried using 2.1.3 on the troubled machine, and its still seg
faulting. Still looks there's a thread problem, but not with ssl
necessarily. Still seeing the the double free() call at the end. Below
is some detailed gdb output from the core.

(gdb) bt full
#0  0x08072627 in ExitCleanly ()
No symbol table info available.
#1  0x0807292e in ExitCleanly ()
No symbol table info available.
#2  0x08056ac8 in ExitCleanly ()
No symbol table info available.
#3  0x08050d68 in HandleConnection ()
No symbol table info available.
#4  0x29136060 in pthread_detach () from /lib/libpthread.so.0
No symbol table info available.

(gdb) disass $eip-0x20 $eip+0x20
Dump of assembler code from 0x8072607 to 0x8072647:
0x8072607 <ExitCleanly+133687>: add    $0x8d,%al
0x8072609 <ExitCleanly+133689>: subl   $0x63,(%eax)
0x807260c <ExitCleanly+133692>: (bad)  
0x807260d <ExitCleanly+133693>: decl   0xc6e82404(%ecx)
0x8072613 <ExitCleanly+133699>: sar    $0xff,%ebp
0x8072616 <ExitCleanly+133702>: mov    0x8(%ebp),%eax
0x8072619 <ExitCleanly+133705>: cmp    (%eax),%edi
0x807261b <ExitCleanly+133707>: je     0x8072660 <ExitCleanly+133776>
0x807261d <ExitCleanly+133709>: test   %edi,%edi
0x807261f <ExitCleanly+133711>: je     0x8072644 <ExitCleanly+133748>
0x8072621 <ExitCleanly+133713>: mov    0x18(%edi),%eax
0x8072624 <ExitCleanly+133716>: mov    0xffffff9c(%ebp),%edx
0x8072627 <ExitCleanly+133719>: mov    %eax,0x18(%edx)
0x807262a <ExitCleanly+133722>: mov    0x4(%edi),%eax
0x807262d <ExitCleanly+133725>: mov    %eax,(%esp,1)
0x8072630 <ExitCleanly+133728>: call   0x804e9ec <free>
0x8072635 <ExitCleanly+133733>: mov    0x8(%edi),%eax
0x8072638 <ExitCleanly+133736>: test   %eax,%eax
0x807263a <ExitCleanly+133738>: jne    0x8072656 <ExitCleanly+133766>
0x807263c <ExitCleanly+133740>: mov    %edi,(%esp,1)
0x807263f <ExitCleanly+133743>: call   0x804e9ec <free>
0x8072644 <ExitCleanly+133748>: mov    0x530(%ebx),%eax
End of assembler dump.

I can get you access to the core if you'd like. Its 5.2M :) Any idea
whats causing this?

Thanks!

-- 
Lance Albertson <ramereth@gentoo.org>
Gentoo Infrastructure

signature.asc
Description: This is a digitally signed message part