Return-Path: Authentication-Results: lists.ozlabs.org; spf=pass (mailfrom) smtp.mailfrom=fuzziesquirrel.com (client-ip=173.167.31.197; helo=bajor.fuzziesquirrel.com; envelope-from=bradleyb@fuzziesquirrel.com; receiver=) Authentication-Results: lists.ozlabs.org; dmarc=none (p=none dis=none) header.from=fuzziesquirrel.com Received: from bajor.fuzziesquirrel.com (mail.fuzziesquirrel.com [173.167.31.197]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by lists.ozlabs.org (Postfix) with ESMTPS id 44PVq648TbzDqCW; Thu, 21 Mar 2019 00:18:25 +1100 (AEDT) X-Virus-Scanned: amavisd-new at fuzziesquirrel.com Date: Wed, 20 Mar 2019 09:18:59 -0400 From: Brad Bishop To: Joseph Reynolds Cc: openbmc@lists.ozlabs.org, openbmc Subject: Re: API authentication Message-ID: <20190320131858.yziiadaogwcl2hfz@thinkpad> References: <20190318103504.zcfgdmgqfhi63jte@thinkpad> <91edbcb6aa1390130fb15b11f35e48bd@linux.vnet.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Disposition: inline In-Reply-To: <91edbcb6aa1390130fb15b11f35e48bd@linux.vnet.ibm.com> X-BeenThere: openbmc@lists.ozlabs.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Development list for OpenBMC List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 20 Mar 2019 13:18:28 -0000 On Mon, Mar 18, 2019 at 09:58:33AM -0500, Joseph Reynolds wrote: >On 2019-03-18 05:35, Brad Bishop wrote: >>I am looking for ideas on how to implement a mechanism to restrict >>access to >>specific methods of a service/API (e.g. Redfish). >> >>This would be orthogonal to role-based authorization - e.g. the >>authorization >>would be provided by someone other than the system administrator - >>e.g. the >>system manufacturer. > >Is this a followup to the March 4 OpenBMC Community call? >https://github.com/openbmc/openbmc/wiki/Weekly-Community-Telecon > >My understanding is that we would design and code the BMC with >functions for things like manufacturing test and specialized >diagnostics ... functions which could harm the device, etc. So we >want the functions to be present for manufacturing test and service >calls, but locked out for all other users. Is that what this >mechanism is for? I hadn't thought of mfg mode commands as a use case for this but on the surface it does seem to be a match.