Messages in this thread | | | Date | Fri, 22 Jun 2012 17:12:07 -0400 | Subject | Re: Bug in net/ipv6/ip6_fib.c:fib6_dump_table() | From | Debabrata Banerjee <> |
| |
On Fri, Jun 22, 2012 at 2:13 PM, Eric Dumazet <eric.dumazet@gmail.com> wrote: > On Fri, 2012-06-22 at 08:44 -0500, Josh Hunt wrote: > >> Ahh. That makes sense and is what Alexey said before I just didn't put >> it all together. So we are OK reverting this patch? I cannot find a path >> where the walker's pointers are updated without the tb6_lock write_lock. >> > > There was a bug somewhere, not sure we want to NULL dereference again. >
As you identified, the tree seems to be protected by tb6_lock. I couldn't find a race by inspection either. If this is not the root of the problem, how would this patch fix it? So I think it does nothing. We are attempting to reproduce that crash to prove it, but like Gao feng I don't think we will see it.
My current favorite theory is that inet6_dump_fib was called with a NULL func in callback. This looks like the approximate area of the crash, but it's impossible to say without more information from Patrick McHardy.
| |