From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stephen Hemminger Subject: Re: [PATCH 7/8] ethdev: break ethernet driver and pci_driver connection Date: Tue, 10 Jan 2017 10:03:21 -0800 Message-ID: <20170110100321.4d53be0a@xeon-e3> References: <20170107181756.1944-1-sthemmin@microsoft.com> <20170107181756.1944-8-sthemmin@microsoft.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: dev@dpdk.org, Stephen Hemminger To: Jan Blunck Return-path: Received: from mail-pf0-f175.google.com (mail-pf0-f175.google.com [209.85.192.175]) by dpdk.org (Postfix) with ESMTP id 7EAC65320 for ; Tue, 10 Jan 2017 19:03:29 +0100 (CET) Received: by mail-pf0-f175.google.com with SMTP id y143so29783513pfb.0 for ; Tue, 10 Jan 2017 10:03:29 -0800 (PST) In-Reply-To: List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" On Tue, 10 Jan 2017 17:11:15 +0100 Jan Blunck wrote: > On Sat, Jan 7, 2017 at 7:17 PM, Stephen Hemminger > wrote: > > There are multiple buses and device types now. Therefore it no longer > > makes sense that PCI driver information is part of the Ethernet driver > > structure. > > The Ethernet driver itself doesn't over alot of value from an > abstraction point of view. Its questionable if there ever will be an > Ethernet driver that is able to operate on different types of > low-level devices. The virtual devices are anyway able to operate > without an Ethernet driver structure. Most of that functionality > should get moved either into the bus abstraction or the low-level > device probe function. I agree that that 'struct eth_driver' is not adding a lot now. It should really be all folded back into 'struct rte_driver'. The concept of init, uninit and private data are all generic and not really specific to ethernet in anyway. If we kill off eth_driver then PCI devices only have rte_pci_driver and VMBUS can have rte_vmbus_driver.