From mboxrd@z Thu Jan 1 00:00:00 1970 From: Amin Azez Subject: Re: [PATCH] ct-event API port to 2.6.11 Date: Thu, 07 Apr 2005 17:28:09 +0100 Message-ID: <42555F99.8010201@ufomechanic.net> References: <424747D5.8060500@eurodev.net> <42502B1B.7090309@trash.net> <42540839.8020405@ufomechanic.net> <4254265F.6080607@eurodev.net> <42554200.3020405@ufomechanic.net> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: Pablo Neira , Patrick McHardy Return-path: To: netfilter-devel@lists.netfilter.org In-Reply-To: <42554200.3020405@ufomechanic.net> 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 [Pablo, sorry, you got missed in the followups to the message I'm responding to, which is a shame as it was addressed to you, but I quoted the whole message here] Amin Azez wrote: > Pablo Neira wrote: > >> Amin Azez wrote: >> >>> It appears that pom-ng nfnetlink-ctnetlink-0.13 (ported by Pablo?) >>> requires ct-event (although this is not listed in the pom-ng >>> dependancies). >>> >>> Pablo, can you confirm this? >> >> >> I've just merged Harald's and my port. You can get it from: >> >> http://people.netfilter.org/~pablo/patches/nfnetlink-conntrack-0.50-2.6.11/ > > I had thought that your merge with Harald replaced > nfnetlink-ctnetlink-0.13 on pom-ng. > > After your new patch (which applies cleanly to 2.6.11) trying to compile > libctnetlink fails on not finding nfnetlink_conntrack.h which can't be > found in any kernel sources (patched or otherwise) but only be found in: > patch-o-matic-ng/nfnetlink-ctnetlink-0.13/linux/include/linux/nfnetlink_conntrack.h > > Do I still need to grab any files from nfnetlink-ctnetlink-0.13 in pom-ng? Another libctnetlink compile problem relates to linux/compiler.h not being included. I guess the best way to bring linux/compiler.h in is libctnetlink.c or some other new .h file whose purpose will be to bring in some of the global include files that are applied when the kernel is compiled. I've not done enough nf hacking to be able to make this judgement; what usually happens when kernel headers that applications want to use end up using magic kernel defines like "likely"? Do we try and generally fake that sort of environment for the whole project, or just per-file-that-needs-it? In file included from /lib/modules/2.6.11.6/build/include/asm/atomic.h:6, from /lib/modules/2.6.11.6/build/include/linux/netfilter_ipv4/ip_conntrack_icmp.h:4, from /lib/modules/2.6.11.6/build/include/linux/netfilter_ipv4/ip_conntrack.h:129, from libctnetlink.c:27: /lib/modules/2.6.11.6/build/include/asm/processor.h: In function `load_esp0': /lib/modules/2.6.11.6/build/include/asm/processor.h:474: warning: implicit declaration of function `unlikely' Sam