From mboxrd@z Thu Jan 1 00:00:00 1970 From: Bart Van Assche Subject: Re: net/smc and the RDMA core Date: Tue, 2 May 2017 15:37:07 +0000 Message-ID: <1493739426.2552.1.camel@sandisk.com> References: <20170501163311.GA22209@lst.de> <1493659776.2665.7.camel@sandisk.com> Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Return-path: In-Reply-To: Content-Language: en-US Content-ID: <3F299F14F648384182E3788903D885C8@namprd04.prod.outlook.com> Sender: netdev-owner@vger.kernel.org To: "hch@lst.de" , "davem@davemloft.net" , "parav@mellanox.com" , "ubraun@linux.vnet.ibm.com" Cc: "netdev@vger.kernel.org" , "linux-rdma@vger.kernel.org" List-Id: linux-rdma@vger.kernel.org On Tue, 2017-05-02 at 14:41 +0200, Ursula Braun wrote: > On 05/01/2017 07:55 PM, Parav Pandit wrote: > > Hi Bart, Ursula, Dave, > >=20 > > I am particularly concerned about SMC as address family. > > It should not be treated as address family, but rather an additional > > protocol similar for socket type SOCK_STREAM. >=20 > We tried to avoid changes of the kernel TCP code. A new address family > seemed to be a feasible way to achieve this. Hello Ursula, I agree with Parav that introducing a new address family for SMC was an unfortunate choice. As one can see in e.g. the implementation of the SCTP protocol no changes to the TCP implementation are needed to add support for a new SOCK_STREAM protocol. I think the SCTP implementation uses inet_register_protosw() to register itself dynamically. Bart.=