From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stephen Hemminger Subject: Re: [PATCH v2] devargs: add blacklisting by linux interface name Date: Thu, 5 Nov 2015 11:23:58 -0800 Message-ID: <20151105112358.6cec995f@xeon-e3> References: <1443798007-20122-1-git-send-email-3chas3@gmail.com> <561CFDDB.90601@6wind.com> <1444830094.3494.59.camel@gmail.com> <7145730.ihAL5VqqiH@xps13> <1446741544.1777.3.camel@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: dev@dpdk.org To: "Charles (Chas) Williams" <3chas3@gmail.com> Return-path: Received: from mail-pa0-f43.google.com (mail-pa0-f43.google.com [209.85.220.43]) by dpdk.org (Postfix) with ESMTP id B25CE5A6C for ; Thu, 5 Nov 2015 20:23:49 +0100 (CET) Received: by pasz6 with SMTP id z6so99538880pas.2 for ; Thu, 05 Nov 2015 11:23:49 -0800 (PST) In-Reply-To: <1446741544.1777.3.camel@gmail.com> 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 Thu, 05 Nov 2015 11:39:04 -0500 "Charles (Chas) Williams" <3chas3@gmail.com> wrote: > On Wed, 2015-11-04 at 23:40 +0100, Thomas Monjalon wrote: > > 2015-10-14 09:41, Charles Williams: > > > On Tue, 2015-10-13 at 14:49 +0200, Olivier MATZ wrote: > > > > For PCI devices that have several interfaces (I think it's the case for > > > > some Mellanox boards), maybe we should not store the interface name? > > > > > > I am not sure what you mean here. If a device has multiple ethernet > > > interfaces, then it should a have seperate PCI device address space for > > > each interface (I dont know of any DPDK drivers that don't make this > > > assumption as well). > > > > mlx4 and cxgbe? > > OK, I see now. I don't know of a way to tell if a device has multiple > ports just from the pci vendor/device id without maintaining some > sort of table. > > Do these devices have multiple interfaces listed in their > /sys/devices/.../net diretory? If so, matching one of the listed > interfaces can just blacklist the whole device similar to blacklisting > by the device id. Devices with multiple ports are supposed to report the port via /sys/class/net/xxx/portid But you aren't going to be able to blacklist only one port of these devices. The two drivers would be fighting over registers and IRQ management. Plus kernel bind/unbind is by PCI id.