Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S933608AbcLGW4Z (ORCPT ); Wed, 7 Dec 2016 17:56:25 -0500 Received: from esa5.dell-outbound.iphmx.com ([68.232.153.95]:15265 "EHLO esa5.dell-outbound.iphmx.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932642AbcLGW4W (ORCPT ); Wed, 7 Dec 2016 17:56:22 -0500 DomainKey-Signature: s=smtpout; d=dell.com; c=simple; q=dns; h=Received:Received:Received:X-DKIM:Received:Received:From: To:Cc:References:In-Reply-To:Subject:Date:Message-ID: MIME-Version:Content-Type:Content-Transfer-Encoding: X-Mailer:Thread-Index:Content-Language: X-RSA-Classifications:X-Sentrion-Hostname; b=DTO4/LINluvOrRg9xaJhRmAhTkLCO5RwH6HmLu8gDqYiuetxyMb7Dt58 Ssp3+56snk7W/RNBRDm6EjKXiw24l9a97Vc3Ya1+Ne3CcQYIgYzA1IT3p Oltzp1nrq3kzYRVk4qXiLtnHDr7QuCoMl/7V+ylEsfPqyhxRB7Fa5BaX1 w=; X-DKIM: OpenDKIM Filter v2.4.3 mailuogwprd53.lss.emc.com uB7MuHVa007032 From: "Allen Hubbe" To: "'Allen Hubbe'" , "'Serge Semin'" , , , Cc: , , References: <003d01d24dc1$f5ba4d60$e12ee820$@dell.com> In-Reply-To: <003d01d24dc1$f5ba4d60$e12ee820$@dell.com> Subject: RE: [PATCH 02/22] NTB: Add peer indexed ports NTB API Date: Wed, 7 Dec 2016 17:56:02 -0500 Message-ID: <003f01d250dd$16de5630$449b0290$@dell.com> MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit X-Mailer: Microsoft Outlook 14.0 Thread-Index: AdJNwe0GkNEPhjRtSzKoPh8KwbdwOwDF6EHg Content-Language: en-us X-RSA-Classifications: public X-Sentrion-Hostname: mailuogwprd53.lss.emc.com Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1211 Lines: 23 From: Allen Hubbe > From: Serge Semin > > Add new port-index NTB API. Additionally lets get rid of Primary and > > Secondary topologies, since port-number can be effectively used instead. > > Split into two patches please. > > I see no harm to the TOPO changes, though I wonder if they are necessary. > I am leaning more toward recommending that the topo api be left alone. The topo changes to the ntb api complicate the Intel driver: - The changes add second branches where there had been just one before (need to check topo AND port now). - The changes also cause some complicated merge conflicts with https://github.com/davejiang/linux.git ntb. See RE: [PATCH 10/22] NTB Intel: Add port-related NTB API callback methods If we leave the ntb topo api as it was, then on multiport devices, if the local port is not the primary port, let it be one of potentially many secondary ports. Or, if there is no distinction between primary/secondary on some hardware, let them all be primary. This topo api doesn't have much value for existing drivers (that I know of), except for informative purposes. So, my preference for changing it would be, only if necessary, and to minimize changes otherwise. Allen