From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from vms173005pub.verizon.net (vms173005pub.verizon.net [206.46.173.5]) by yocto-www.yoctoproject.org (Postfix) with ESMTP id 1B8B8E0072A for ; Fri, 27 Jan 2012 12:57:19 -0800 (PST) Received: from gandalf.denix.org ([unknown] [71.163.124.23]) by vms173005.mailsrvcs.net (Sun Java(tm) System Messaging Server 7u2-7.02 32bit (built Apr 16 2009)) with ESMTPA id <0LYH002IG67IYUC2@vms173005.mailsrvcs.net> for meta-ti@yoctoproject.org; Fri, 27 Jan 2012 14:57:18 -0600 (CST) Received: by gandalf.denix.org (Postfix, from userid 1000) id 3153C202E0; Fri, 27 Jan 2012 15:57:17 -0500 (EST) Date: Fri, 27 Jan 2012 15:57:17 -0500 From: Denys Dmytriyenko To: Philip Balister Message-id: <20120127205717.GI8707@denix.org> References: <4F21D735.5000303@ti.com> <27324814-C581-4BD9-92DF-68FC5CB4CECF@dominion.thruhere.net> <7D46E86EC0A8354091174257B2FED1010BF9BA72@DLEE12.ent.ti.com> <4F22E41B.8070708@ti.com> <7D46E86EC0A8354091174257B2FED1010BF9D77A@DLEE12.ent.ti.com> <4F2304E7.7070809@ti.com> <7D46E86EC0A8354091174257B2FED1010BF9D879@DLEE12.ent.ti.com> <20120127202150.GE8707@denix.org> <4F230803.1070503@ti.com> <4F230E21.9010405@balister.org> MIME-version: 1.0 In-reply-to: <4F230E21.9010405@balister.org> User-Agent: Mutt/1.5.20 (2009-06-14) Cc: "meta-ti@yoctoproject.org" Subject: Re: [PATCH 1/3] matrix-gui-browser: port from arago overlay X-BeenThere: meta-ti@yoctoproject.org X-Mailman-Version: 2.1.13 Precedence: list List-Id: Mailing list for the meta-ti layer List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 27 Jan 2012 20:57:20 -0000 Content-type: text/plain; charset=us-ascii Content-disposition: inline On Fri, Jan 27, 2012 at 03:50:41PM -0500, Philip Balister wrote: > On 01/27/2012 03:24 PM, William Mills wrote: > > > ..... > > > >>> can > >>> agree to spliting meta-ti into two layers, a HW layer for our devices > >>> and a > >>> layer containing all of the TI recipes. > >>> > >>> But if we wanted to match the meta-intel layer way would you also > >>> propose > >>> making a layer per device? I personally find that more confusing. > >> I don't think that was Bill's message. It was simplicity. BSP only > >> layer, no > >> supplemental apps, if not absolutely required. > >> > >> Your example with lmbench is not correct - BSP layer should be simple > >> enough > >> to be used with OE-Core alone to produce a console rootfs image with > >> nothing > >> but busybox. > >> > >> How about splitting meta-ti into: > >> * BSP only > >> * SGX graphics > >> * DSP tools > >> * WiFi etc. > >> > >> And then splitting meta-arago into: > >> * Arago distro for TI SDKs > >> * Supplemental apps > >> > > > > Need to get my "YES!" in here before everyone barfs all over the > > proposal :) > > > > Chase: your right. I do not want to follow intel's example of layer per > > BSP. > > I build stuff for the USRP E100 (based on a gumstic overo). I use my own > later for BSP that provides kernel, u-boot, and image recipes. > > I still need a TI BSP layer for DSP stuff (we do not care about SGX, > although it is possible customers could). > > Do forget your customers using all this to ship products. > Philip, I don't understand what you are arguing here about or against? :) It won't change much for you, maybe just setup step a little. The proposal above is to split meta-ti and meta-arago repositories into multiple layers inside those repositories, like meta-oe already does. Your example above is a good one - having BSP, DSP and SGX in 3 separate layers allows you to enable first two w/o the need to get the second one parsed or used. -- Denys