[sldev] Weird crash

[Dale Glass]

В сообщении от 21 января 2007 20:42 Jan-Benedict Glaw написал(a):
> On Sun, 2007-01-21 19:58:04 +0100, Dale Glass <dale at daleglass.net> wrote:
> > Program received signal SIGSEGV, Segmentation fault.
> > [Switching to Thread 4136236752 (LWP 12862)]
> > 0xf757f736 in free () from /lib32/libc.so.6
> > (gdb) bt
> > #0  0xf757f736 in free () from /lib32/libc.so.6
> > #1  0xf75815ad in malloc () from /lib32/libc.so.6
> > #2  0xf77193b9 in operator new ()
> > from /usr/lib/gcc/x86_64-pc-linux-gnu/4.1.1/32/libstdc++.so.6
>
> Looks like a crash after the internal malloc data structures were
> fscked up.  Try to start with an allocation debugger.
>
> MfG, JBG

No, it's not that. Look at the length of the stack trace, there are 40256 
lines in it!

What happens is that it's looping.

lltreenode.h:118 -> lloctree.h:397 -> lltreenode.h:118 -> lloctree.h:391 -> 
lltreenode.h:118 and the loop completes.

So it goes on like that until it reaches the stack size limit, at which point 
it crashes in malloc, but malloc doesn't really have anything to do with it, 
it's just the place where it runs out of stack space.

The right place to debug is not the place of the crash, but figuring out what 
makes the loop start.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://lists.secondlife.com/pipermail/sldev/attachments/20070121/be61bf93/attachment.pgp