From mboxrd@z Thu Jan 1 00:00:00 1970 From: Pablo Neira Ayuso Subject: Re: [PATCH nf-next 2/4] netfilter: add BUILD_BUG_ON asserts for hook array sizes Date: Wed, 22 Nov 2017 13:54:39 +0100 Message-ID: <20171122125439.GA563@salvia> References: <20171113164107.11259-1-fw@strlen.de> <20171113164107.11259-3-fw@strlen.de> <20171122122025.GA29456@salvia> <20171122124410.GI24866@breakpoint.cc> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit Cc: netfilter-devel@vger.kernel.org To: Florian Westphal Return-path: Received: from mail.us.es ([193.147.175.20]:40098 "EHLO mail.us.es" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751216AbdKVM56 (ORCPT ); Wed, 22 Nov 2017 07:57:58 -0500 Received: from antivirus1-rhel7.int (unknown [192.168.2.11]) by mail.us.es (Postfix) with ESMTP id 089461BFA85 for ; Wed, 22 Nov 2017 13:57:57 +0100 (CET) Received: from antivirus1-rhel7.int (localhost [127.0.0.1]) by antivirus1-rhel7.int (Postfix) with ESMTP id EC1DFDA793 for ; Wed, 22 Nov 2017 13:57:56 +0100 (CET) Content-Disposition: inline In-Reply-To: <20171122124410.GI24866@breakpoint.cc> Sender: netfilter-devel-owner@vger.kernel.org List-ID: On Wed, Nov 22, 2017 at 01:44:10PM +0100, Florian Westphal wrote: > Pablo Neira Ayuso wrote: > > Hi Florian, > > > > On Mon, Nov 13, 2017 at 05:41:05PM +0100, Florian Westphal wrote: > > > Check that the array hooks are not accessed out-of-bounds. > > > Next patch will then reduce their sizes to reflect the number > > > of hooks implemented for each family. > > > > I'm hitting this here. > > > > In file included from ./include/linux/kernel.h:10:0, > > from ./include/linux/uio.h:12, > > from ./include/linux/socket.h:8, > > from net/decnet/dn_route.c:60: > > In function ‘nf_hook.constprop’, > > inlined from ‘NF_HOOK.constprop’ at > > ./include/linux/netfilter.h:279:6: > > ./include/linux/compiler.h:319:38: error: call to > > ‘__compiletime_assert_221’ declared with attribute error: BUILD_BUG_ON > > failed: hook >= ARRAY_SIZE(net->nf.hooks_decnet) > > _compiletime_assert(condition, msg, __compiletime_assert_, __LINE__) > > Thanks for the report, I guess gcc doesn't follow inline trail > or argument isn't known at compile time. > > I'll add a contant_p() test as well in v2. Side note: It's only happening with decnet.