[sheepdog] [PATCH v2 08/11] work: don't call get_node_info in the worker thread

[MORITA Kazutaka]

At Fri, 19 Apr 2013 16:35:46 +0800,
Liu Yuan wrote:
> 
> On 04/19/2013 04:28 PM, Liu Yuan wrote:
> > core dump's bt:
> > (gdb) bt
> > #0  0x00007ffada064b7b in raise (sig=<optimized out>) at ../nptl/sysdeps/unix/sysv/linux/pt-raise.c:42
> > #1  0x000000000042530f in reraise_crash_signal (signo=11, status=1) at util.c:414
> > #2  <signal handler called>
> > #3  bs_thread_request_done (fd=<optimized out>, events=<optimized out>, data=<optimized out>) at work.c:160
> > #4  0x0000000000421899 in do_event_loop (sort_with_prio=false, timeout=<optimized out>) at event.c:211
> > #5  0x0000000000404ebf in main (argc=<optimized out>, argv=<optimized out>) at sheep.c:691
> 
> From this NULL dereference example, I don't think install a handler for
> SEGV is a good idea. It just obscures where the NULL def happens.

I don't get it.  It's really clear that you got SEGV at the stack
frame #3 from this info.  It really helps us to include the backtrace
in sheep.log with the handler when we cannot reproduce the problem.
Actually, I couldn't solve this problem without your log.

Thanks,

Kazutaka