From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ferruh Yigit Subject: Re: [PATCH 0/3] ethdev: add helper functions to get eth_dev and dev private data Date: Thu, 10 Mar 2016 13:37:51 +0000 Message-ID: <56E178AF.9000207@intel.com> References: <1455718817-14171-1-git-send-email-ferruh.yigit@intel.com> <2720050.b5GnNWtRZF@xps13> Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: quoted-printable Cc: dev@dpdk.org To: Thomas Monjalon Return-path: Received: from mga14.intel.com (mga14.intel.com [192.55.52.115]) by dpdk.org (Postfix) with ESMTP id 9B84A2BA0 for ; Thu, 10 Mar 2016 14:37:54 +0100 (CET) In-Reply-To: <2720050.b5GnNWtRZF@xps13> 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 3/10/2016 12:00 AM, Thomas Monjalon wrote: > 2016-02-17 14:20, Ferruh Yigit: >> This is to provide abstraction and reduce global variable access. >> >> Global variable rte_eth_devices kept exported to not break ABI. >> >> Bonding driver not selected on purpose, just it seems it is using=20 >> rte_eth_devices heavily. >> >> There are a few more usage in drivers but they left as it is because t= hey >> are in fast path code. >=20 > What is the benefit of these functions if you do not plan to remove the > global variables later? >=20 - Better discipline to abstract, function call can be costly than direct access, but I believe this is no problem outside of the fast path. - Helper functions for common tasks, like get private data by port. > Anyway this discussion targets the release 16.07. >=20 There is no urgency for these. Thanks, ferruh