From mboxrd@z Thu Jan 1 00:00:00 1970 From: Roland Dreier Subject: Re: [PATCH] net/9p: limit RDMA to INET Date: Mon, 03 Nov 2008 12:52:50 -0800 Message-ID: References: <20081103111731.d12dbb84.randy.dunlap@oracle.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: In-Reply-To: <20081103111731.d12dbb84.randy.dunlap@oracle.com> (Randy Dunlap's message of "Mon, 3 Nov 2008 11:17:31 -0800") Sender: netdev-owner@vger.kernel.org To: Randy Dunlap Cc: lkml , linux-next@vger.kernel.org, sfr , rolandd@cisco.com, ericvh@gmail.com, akpm , davem@davemloft.net, netdev List-Id: linux-next.vger.kernel.org > Make 9p's RDMA option depend on INET since it uses Infiniband rdma_* > functions and that code depends on INET. Otherwise 9p can try to > use symbols which don't exist. > > ERROR: "rdma_destroy_id" [net/9p/9pnet_rdma.ko] undefined! Good catch. > config NET_9P_RDMA > - depends on NET_9P && INFINIBAND && EXPERIMENTAL > + depends on INET && INFINIBAND && EXPERIMENTAL I wonder if there's a way to make this depend on INFINIBAND_ADDR_TRANS instead of INET, since that's the real dependency to get the rdma_* symbols. However INFINIBAND_ADDR_TRANS is a bool that gets set to y iff INET and INFINIBAND are set, so it's not trivial to make a tristate depend on it. Maybe this patch is fine as it is, since the Kconfig dependencies are unlikely to change anyway... - R.