From: Vincent Bernat Subject: Re: [ANNOUNCE] Release of iptables-1.4.3.2 Date: Thu, 09 Apr 2009 21:27:18 +0200 Message-ID: References: <49D9E9A6.7010303@netfilter.org> <49DA0F49.4090802@conversis.de> <49DAD5E0.9020303@caf.com.tr> <49DDB24A.80400@chello.at> <49DDF7C2.80008@gmail.com> Mime-Version: 1.0 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: <49DDF7C2.80008@gmail.com> (Eray Aslan's message of "Thu, 09 Apr 2009 16:27:30 +0300") Sender: netfilter-owner@vger.kernel.org List-ID: Content-Type: text/plain; charset="utf-8" To: netfilter@vger.kernel.org OoO En ce d=C3=A9but d'apr=C3=A8s-midi ensoleill=C3=A9 du jeudi 09 a= vril 2009, vers 15:27, Eray Aslan disait=C2=A0: >> At least it violates the concept of filtering in the filter table. > Yes. Correct place to filter in in the filter table. But if you ins= ist > on doing the "wrong" thing, who is to interfere? Well, it depends what "wrong" is. When dropping in filter table, yo= u leave an entry in the conntrack. If you try to drop to avoid some kin= d of DOS, this is an unwanted effect. When you drop in PREROUTING in th= e nat table, there is no entry in the conntrack. This is really convenien= t with matches like "hashlimit". This loss of functionality should be better explained. Maybe this is th= e prelude of some simplification in Netfilter? --=20 Choose a data representation that makes the program simple. - The Elements of Programming Style (Kernighan & Plauger)