From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jean Tourrilhes Subject: Re: [PATCH 1/1] eal: Don't fail secondary if primary is missing tailqs Date: Wed, 5 Oct 2016 10:34:55 -0700 Message-ID: <20161005173455.GB12182@labs.hpe.com> References: <20160922204637.GA3166@labs.hpe.com> <20161005164906.GB11912@labs.hpe.com> <7491622.GqnA43pcBO@xps13> Reply-To: jean.tourrilhes@hpe.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: dev@dpdk.org, David Marchand , Sergio Gonzalez Monroy , olivier.matz@6wind.com, David Hunt To: Thomas Monjalon Return-path: Received: from g2t2353.austin.hpe.com (g2t2353.austin.hpe.com [15.233.44.26]) by dpdk.org (Postfix) with ESMTP id 8C5852A1A for ; Wed, 5 Oct 2016 19:34:57 +0200 (CEST) Content-Disposition: inline In-Reply-To: <7491622.GqnA43pcBO@xps13> List-Id: patches and discussions about DPDK List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" On Wed, Oct 05, 2016 at 07:09:14PM +0200, Thomas Monjalon wrote: > > Probably that you would have some aligned builds if Snort was using > a pkg-config approach to link DPDK. I seriously doubt it, but maybe there is some deep linker magic that would pick the appropriate set of constructor. > > For tailq, I agree. For mempool constructors, order do matter. > > I don't know why such a complex function (rte_mempool_register_ops) is > called inside a constructor. Maybe that's the main problem. No. The problem is that the list of constructors linked by the linker in each binary is different, whereas DPDK expect them to be the same. Regards, Jean