From mboxrd@z Thu Jan 1 00:00:00 1970 From: Julian Anastasov Subject: Re: unregister_netdevice: waiting for DEV to become free (2) Date: Wed, 22 Aug 2018 07:11:41 +0300 (EEST) Message-ID: References: <0000000000007d22100573d66078@google.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Cc: syzbot+30209ea299c09d8785c9@syzkaller.appspotmail.com, ddstreet@ieee.org, Dmitry Vyukov , LKML , Linux Kernel Network Developers , syzkaller-bugs@googlegroups.com To: Cong Wang Return-path: In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org List-Id: netdev.vger.kernel.org Hello, On Mon, 20 Aug 2018, Cong Wang wrote: > For the one I look into, dst_cache doesn't matter, because the xmit > path doesn't even use tunnel dst_cache at all, and it is ip6tnl0 FB > device, unlike this one which is tun device. Ops, of course, it is dev tun and not ip tun... > For the one I look into, it seems some fib6_info is not released for > some reason. It seems to be the one created by addrconf_prefix_route(), > which is supposed to be released by fib6_clean_tree() I think, but it > never happens. May be, it is not direct reference to dev but one that is moved to loopback, like from dst, route... The repro.c creates permanent neighbours and addresses. Regards -- Julian Anastasov