From mboxrd@z Thu Jan 1 00:00:00 1970 From: Phil Oester Subject: ip6_finish_output2 change broke netfilter xt_TEE target Date: Tue, 17 Sep 2013 12:54:34 -0700 Message-ID: <20130917195434.GA8424@linuxace.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: netdev@vger.kernel.org, netfilter-devel@vger.kernel.org To: yoshfuji@linux-ipv6.org Return-path: Received: from mail-pd0-f179.google.com ([209.85.192.179]:34341 "EHLO mail-pd0-f179.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753416Ab3IQTyN (ORCPT ); Tue, 17 Sep 2013 15:54:13 -0400 Received: by mail-pd0-f179.google.com with SMTP id v10so6095055pde.10 for ; Tue, 17 Sep 2013 12:54:13 -0700 (PDT) Content-Disposition: inline Sender: netfilter-devel-owner@vger.kernel.org List-ID: The change made in commit 6fd6ce20 (ipv6: Do not depend on rt->n in ip6_finish_output2) broke the xt_TEE target for IPv6 packets. Instead of using the nexthop provided in the --gateway option, ip6_finish_output2 is now performing neighbor solicitation for the original daddr in the copied skb. Similar breakage occurred in IPv4, and was fixed (in 2ad5b9e4) by using the flag FLOWI_FLAG_KNOWN_NH. I can find no easy way to make use of that flag here. Reverting 6fd6ce20 makes TEE work again, but I am not clear on what problem that commit was attempting to solve. Yoshifuji? Phil