From mboxrd@z Thu Jan 1 00:00:00 1970 From: Pablo Neira Ayuso Subject: Re: Xtables2 Netlink spec Date: Wed, 08 Dec 2010 12:36:16 +0100 Message-ID: <4CFF6DB0.2020703@netfilter.org> References: <4CEE4B94.8010307@netfilter.org> <4CEE70CE.60502@netfilter.org> <4CEF6F12.7080601@netfilter.org> <4CFDE6FB.4060103@netfilter.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: Netfilter Developer Mailing List To: Jan Engelhardt Return-path: Received: from mail.us.es ([193.147.175.20]:51861 "EHLO mail.us.es" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754758Ab0LHLgU (ORCPT ); Wed, 8 Dec 2010 06:36:20 -0500 In-Reply-To: Sender: netfilter-devel-owner@vger.kernel.org List-ID: On 07/12/10 14:30, Jan Engelhardt wrote: > > On Tuesday 2010-12-07 08:49, Pablo Neira Ayuso wrote: >> >>> I am skeptical about shrinkfitting something onto an older >>> interface (nfnetlink) when there is genetlink.. >> >> That's an empty argument. Tell me one feature that nfnetlink does not have have >> but genetlink does. > > I was thinking about the name resolution of netlink subsubsystems. In > plain netlink and nfnetlink, subsys pointers are kept in a static > array, which means the more subsystems are defined, the more memory > is used (for n -> infinity) even if only one subsys is used. > Would that be a reasonable concern? The number of nfnetlink subsystem is small and it will remain small along time I guess. I would not spend time on such optimization. Anyway, the static subsys ID number is a good thing IMO. The floating genetlink ID allows to have out-of-tree subsystems, which is something that I don't like that. Moreover, you have to send an initial message to resolve the ID number and subscribe to possible changes in it.