From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Wiles, Keith" Subject: Re: [PATCH 3/7] eal: move virtual device probing into a bus Date: Wed, 15 Feb 2017 17:22:05 +0000 Message-ID: <356328A2-E1A0-43F9-B413-62482FEF17E8@intel.com> References: <1487152929-23627-1-git-send-email-jblunck@infradead.org> <1487152929-23627-4-git-send-email-jblunck@infradead.org> <485202cf-3644-5ee9-45b5-d443569bf5b5@nxp.com> <7843A545-14B6-4CD9-9689-ABD123B85890@intel.com> <7F1E2AAB-BDF3-4164-8BCE-117B6DA5FF2E@intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Cc: Shreyansh Jain , "dev@dpdk.org" To: Jan Blunck Return-path: Received: from mga09.intel.com (mga09.intel.com [134.134.136.24]) by dpdk.org (Postfix) with ESMTP id 891AD282 for ; Wed, 15 Feb 2017 18:22:08 +0100 (CET) In-Reply-To: <7F1E2AAB-BDF3-4164-8BCE-117B6DA5FF2E@intel.com> Content-Language: en-US Content-ID: <1A9DF2D599802D4C913C83B2A605CB45@intel.com> 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 Feb 15, 2017, at 11:10 AM, Wiles, Keith wrote: >=20 >=20 >> On Feb 15, 2017, at 11:06 AM, Jan Blunck wrote: >>=20 >> On Wed, Feb 15, 2017 at 3:22 PM, Wiles, Keith wr= ote: >>>=20 >>>> On Feb 15, 2017, at 8:15 AM, Shreyansh Jain w= rote: >>>>=20 >>>>=20 >>>> Just ignore this comment - I am misunderstood something. >>>>=20 >>>> But another question: Is there specific reason VDEV should be register= ed/scanned *after* other devices? Is there some specific problem if we do o= therwise? (I think this is should be done, but I don't have a specific reas= on). >>>=20 >>=20 >> Just for context: the vdev's are probed after the physical devices >> because of commit f4ce209a ("eal: postpone vdev initialization"). >>=20 >>> Does the bonding driver which uses physical devices need to be register= ed after physical ones? In Pktgen I noticed the vdev after the physical por= ts and I could not blacklist them as the bonding driver needed them, which = caused the bonding ports to have a greater port number. In the case of pktg= en the bonding ports were up around 8 or 10 and caused the display to not s= how the bonding ports. This is really just a usability problem for the deve= loper using Pktgen. I would like to see the vdev devices first, but as long= as the drivers (like bonding) are fine with them being first. >>>=20 >>=20 >> The bonding devargs might specify slaves that get attached during >> device probe. If the referenced devices are physical interfaces we >> need to probe them first. This is really a chicken-egg-problem. >>=20 >> Maybe you could improve the usability in your case and sort the >> virtual devices first or even hide enslaved ports? >=20 > The port numbering comes from DPDK and I use that directly, was trying to= avoid a translation of real port to Pktgen port :-( >=20 > Regards, > Keith >=20 Looking at the bonding driver it does not attempt to access the physical po= rts until bond_ethdev_configure call (I believe). This means the vdev devic= es should be following the same flow and moving them to before the physical= probes would be fine, right? Regards, Keith