From: Harald Welte Subject: Re: [PATCH] x_tables, take 5 (Final Review) Date: Thu, 12 Jan 2006 17:58:38 +0100 Message-ID: <20060112165838.GL4430@sunbeam.de.gnumonks.org> References: <20060108212619.GE24266@sunbeam.de.gnumonks.org> <200601090852.k098qjeO004027@toshiba.co.jp> <20060109103504.GK24266@sunbeam.de.gnumonks.org> <200601100418.k0A4IxhL008706@toshiba.co.jp> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="dQ2UWBmeJhPxzxhA" Cc: netfilter-devel@lists.netfilter.org, kaber@trash.net Return-path: To: Yasuyuki KOZAKAI Content-Disposition: inline In-Reply-To: <200601100418.k0A4IxhL008706@toshiba.co.jp> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: netfilter-devel-bounces@lists.netfilter.org Errors-To: netfilter-devel-bounces@lists.netfilter.org List-Id: netfilter-devel.vger.kernel.org --dQ2UWBmeJhPxzxhA Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Tue, Jan 10, 2006 at 01:18:59PM +0900, Yasuyuki KOZAKAI wrote: > From: Harald Welte > Date: Mon, 9 Jan 2006 11:35:04 +0100 >=20 > > > I have one more comment about need_conntrack(). The kernel with x_tab= les patch > > > doesn't load nf_connrack_ipv4 when user types "iptables -A INPUT -m s= tate...". > > > Then no packets might be untracked after that. One solution is to try= load > > > nf_conntrack_{ipv4,ipv6} and to ignore error on load, but I'm not sur= e that > > > this is the best way. > >=20 > > Mh. Can we put this into checkentry()? Probably there are some issues > > with regard to locking, but it would be the logical place to do it. We > > [can] know whether it's an iptables or an ip6tables rule, and therefore > > we know which tracking we want to request. > >=20 > > we should have something like a > > "need_conntrack(int pf)" function provided by the conntrack core. Then > > anyone in the kernel can request something like > > "need_conntrack(AF_INET)". >=20 > sounds good idea. I'll implement this later on. Today I want to finally submit x_tables, even with this minor issue left for an incremental patch. > > Also, my Kconfig magic (originally suggested by Rusty) doesn't seem to > > work. It does not automatically select the XT matches/targets based on > > the previous IPT selections :( >=20 > How about this ? > - make {IP,IP6,ARP}_NF_IPTABLES select NETFILTER_XTABLES > - make IP_NF_{MATCH,TARGET}_* select NETFILTER_XT_{MATCH,TARGET}_* > - hide the menus for NETFILTER_XTABLES and > NETFILTER_XT_{MATCH,TARGET}_* > - show the menus for {IP,IP6,ARP}_NF_{MATCH,TARGET}_* no, I don't like that idea. It would mean that we'd hide the existance of x_tables in the long term from the user. Rather the opposite direction is the case: less and less {ip,arp,ip6}_tables, and more functionality moving to the core. --=20 - Harald Welte http://netfilter.org/ =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D "Fragmentation is like classful addressing -- an interesting early architectural error that shows how much experimentation was going on while IP was being designed." -- Paul Vixie --dQ2UWBmeJhPxzxhA Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) iD8DBQFDxoq+XaXGVTD0i/8RAlJYAKCtJDcarrADnFMfl4DJmPccYnGhygCgnMkx cU4HJCDJArUWpxJywVAs/Ko= =MhyQ -----END PGP SIGNATURE----- --dQ2UWBmeJhPxzxhA--