From mboxrd@z Thu Jan 1 00:00:00 1970 From: Chuck Anderson Subject: Re: [E1000-devel] ixgbe: Unsupported SFP+ modules on 10Gbit/s X520-DA2 NIC? Date: Wed, 18 Jan 2012 20:12:21 -0500 Message-ID: <20120119011221.GR5069@angus.ind.WPI.EDU> References: <1326886258.19261.25.camel@probook> <20120118091351.000052fc@unknown> <9B4A1B1917080E46B64F07F2989DADD621FC38@ORSMSX102.amr.corp.intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii To: "netdev@vger.kernel.org" Return-path: Received: from MAIL1.WPI.EDU ([130.215.36.91]:46868 "EHLO MAIL1.WPI.EDU" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755254Ab2ASBhX (ORCPT ); Wed, 18 Jan 2012 20:37:23 -0500 Received: from MAIL1.WPI.EDU (MAIL1.WPI.EDU [130.215.36.91]) by MAIL1.WPI.EDU (8.14.5/8.14.5) with ESMTP id q0J1CMFL018457 for ; Wed, 18 Jan 2012 20:12:22 -0500 Received: from SMTP.WPI.EDU (SMTP.WPI.EDU [130.215.36.186]) by MAIL1.WPI.EDU (8.14.5/8.14.5) with ESMTP id q0J1CM1H018454 for ; Wed, 18 Jan 2012 20:12:22 -0500 Received: from angus.ind.WPI.EDU (ANGUS.IND.WPI.EDU [130.215.130.21]) by SMTP.WPI.EDU (8.14.4/8.14.4) with ESMTP id q0J1CLUZ017372 for ; Wed, 18 Jan 2012 20:12:22 -0500 (envelope-from cra@WPI.EDU) Received: from angus.ind.WPI.EDU (angus.ind.WPI.EDU [127.0.0.1]) by angus.ind.WPI.EDU (8.14.2/8.14.2) with ESMTP id q0J1CL2R014986 for ; Wed, 18 Jan 2012 20:12:21 -0500 Received: (from cra@localhost) by angus.ind.WPI.EDU (8.14.2/8.14.2/Submit) id q0J1CL1i014985 for netdev@vger.kernel.org; Wed, 18 Jan 2012 20:12:21 -0500 Content-Disposition: inline In-Reply-To: <9B4A1B1917080E46B64F07F2989DADD621FC38@ORSMSX102.amr.corp.intel.com> Sender: netdev-owner@vger.kernel.org List-ID: On Wed, Jan 18, 2012 at 10:21:58PM +0000, Fujinaka, Todd wrote: > That's up to you. There's "locked" and there's "locked". I'm surprised > that Benny and Jesper haven't looked at the driver to see where the > messages come from. > > We have a short list of optical modules that have been tested with our > cards. The problem with standards is that there's always some wiggle > room and you won't know if something really works until you try it. No > matter how large the company, we are still constrained as far as what we > can do in a day and testing every module we can find just wasn't one of > the things that was approved for us to try. I don't buy that argument. We have Ethernet standards and we have IP standards and we have SFP/SFP+ standards. Did you test your 1000Base-T copper Ethernet cards with every vendor of Ethernet hardware? If not, did you lock them out to talk to only "pre-approved" Ethernet switches? Would you have done so if there was a way to technically do so (perhaps via LLDP)? What about USB keyboards/mice? Maybe Intel's chipsets can be locked so only Intel USB keyboards work...and then we can all stop buying Intel hardware. The hardware and drivers should not be enforcing specific optics. If a user buys a crap optic, then that is their problem. Just like if they plug a crappy Cat3 RJ45 cable between the 1000Base-T NIC and the switch.