From mboxrd@z Thu Jan 1 00:00:00 1970 From: =?UTF-8?Q?Simon_K=c3=a5gstr=c3=b6m?= Subject: Re: [PATCH 1/7] lib/librte_ether: Add 2/2.5/25/50Gbps link speeds Date: Thu, 3 Mar 2016 11:22:49 +0100 Message-ID: <56D81079.4030905@netinsight.net> References: <1456954614-119304-1-git-send-email-stephen.hurd@broadcom.com> <1456978137-98097-2-git-send-email-stephen.hurd@broadcom.com> <56D7ED89.8060903@netinsight.net> <2201521.3lr0Br3EYS@xps13> Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: quoted-printable Cc: dev@dpdk.org To: Thomas Monjalon , Stephen Hurd Return-path: Received: from ernst.netinsight.se (ernst.netinsight.se [194.16.221.21]) by dpdk.org (Postfix) with SMTP id 9F47E2B94 for ; Thu, 3 Mar 2016 11:22:55 +0100 (CET) In-Reply-To: <2201521.3lr0Br3EYS@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 2016-03-03 10:28, Thomas Monjalon wrote: > 2016-03-03 08:53, Simon K=E5gstr=F6m: >> I realize this is a more general question, but is it really meaningful >> to have macros for all possible link speeds? We're working on a PMD >> driver with a channelized interface exposed as DPDK ports. Each channe= l >> can be configured with an arbitrary speed, so e.g., 1337 Mbps is also >> possible. >=20 > What is the benefit? Why not negotiate the maximum capability of the pe= er? Communication is channelized over a backplane, and each channel has a specific (and configurable) capacity. >> To me, it would seem more natural to just have a number in mbits for t= he >> link speed. >=20 > Please jump in the thread initiated by Marc Sune months ago. OK. I haven't been following the DPDK mailing list for a while, so I wasn't aware of this. // Simon