From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: by yocto-www.yoctoproject.org (Postfix, from userid 118) id BE017E00D47; Sun, 28 Feb 2016 23:12:35 -0800 (PST) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on yocto-www.yoctoproject.org X-Spam-Level: X-Spam-Status: No, score=-6.9 required=5.0 tests=BAYES_00,RCVD_IN_DNSWL_HI autolearn=ham version=3.3.1 X-Spam-HAM-Report: * -1.9 BAYES_00 BODY: Bayes spam probability is 0 to 1% * [score: 0.0000] * -5.0 RCVD_IN_DNSWL_HI RBL: Sender listed at http://www.dnswl.org/, high * trust * [192.55.52.88 listed in list.dnswl.org] Received: from mga01.intel.com (mga01.intel.com [192.55.52.88]) by yocto-www.yoctoproject.org (Postfix) with ESMTP id 826A6E00CA2 for ; Sun, 28 Feb 2016 23:12:31 -0800 (PST) Received: from fmsmga002.fm.intel.com ([10.253.24.26]) by fmsmga101.fm.intel.com with ESMTP; 28 Feb 2016 23:12:32 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.22,519,1449561600"; d="scan'208";a="925981552" Received: from gtan23-mobl3.gar.corp.intel.com (HELO peggleto-mobl.ger.corp.intel.com) ([10.255.165.53]) by fmsmga002.fm.intel.com with ESMTP; 28 Feb 2016 23:12:28 -0800 From: Paul Eggleton To: Gary Thomas Date: Mon, 29 Feb 2016 20:12:26 +1300 Message-ID: <3845777.DZXZ3IeRVd@peggleto-mobl.ger.corp.intel.com> Organization: Intel Corporation User-Agent: KMail/4.14.10 (Linux/4.3.5-300.fc23.x86_64; KDE/4.14.17; x86_64; ; ) In-Reply-To: <56D3BFA9.9080200@mlbassoc.com> References: <2956615.y8hyrZheM7@peggleto-mobl.ger.corp.intel.com> <56D3BB9C.8090105@gmail.com> <56D3BFA9.9080200@mlbassoc.com> MIME-Version: 1.0 Cc: yocto@yoctoproject.org Subject: Re: [Openembedded-architecture] Standalone image writer X-BeenThere: yocto@yoctoproject.org X-Mailman-Version: 2.1.13 Precedence: list List-Id: Discussion of all things Yocto Project List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 29 Feb 2016 07:12:35 -0000 Content-Transfer-Encoding: 7Bit Content-Type: text/plain; charset="us-ascii" On Mon, 29 Feb 2016 04:48:57 Gary Thomas wrote: > On 02/29/2016 04:31 AM, Trevor Woerner wrote: > > On 02/28/16 20:46, Khem Raj wrote: > >>> On Feb 28, 2016, at 4:17 PM, Paul Eggleton > >>> wrote: > >>> > >>> (changing subject line so people don't miss this slight tangent) > >>> > >>> I'd almost completely forgotten, but the part of what we wrote for Hob > >>> to > >>> write images to a USB stick or SD card (bitbake/bin/image-writer) is > >>> actually a standalone application. There's not a lot to it - pick your > >>> image file and then the device you want to write it to; it has some > >>> logic in it to not let you accidentally write to devices that aren't > >>> USB sticks. > >>> > >>> A few points: > >>> > >>> - I'm not sure very many people know this tool exists, so it's likely > >>> it's not being widely used. Having said that it is a nice simple UI > >>> that does the job. > >>> > >>> - It shares some code with Hob, but mostly not the bits with code > >>> quality > >>> issues, though it is still GTK+ 2 based. > >>> > >>> - It doesn't support the advanced SD card writing functionality that has > >>> been implented within wic over the last few releases, which is pretty > >>> important for devices where a special partition layout is expected by > >>> the bootloader. > >>> > >>> - Toaster can't really have this functionality in it because it's web > >>> based > >>> and the web server might not be running locally, so writing to a local > >>> USB > >>> stick or SD card isn't going to be practical from there; about all it > >>> could do is provide instructions on how to write the image once you've > >>> downloaded it. > >>> > >>> - I believe there are other equivalent tools out there that various > >>> distros > >>> use for taking a downloadable ISO image and writing it to a USB stick. I > >>> haven't done a survey to find out if if any of them work in quite the > >>> same way; I know some of them actually unpack the image and then > >>> re-create a filesystem on the device, which isn't the right thing for > >>> our images. > >>> > >>> - We do have a command-line equivalent in the form of > >>> scripts/contrib/ddimage in OE-Core. Of course it's command-line and > >>> thus less friendly but it does the job (and also has some safeguards > >>> against writing to the wrong device). > >>> > >>> So what do we do with this? We have two choices really: > >>> > >>> A) Drop bitbake/bin/imagewriter along with Hob; we could potentially > >>> resurrect it again later if desired > >>> > >>> B) Preserve it along with the shared modules that it requires > >>> > >>> Thoughts? > >> > >> Option A. There are USB writer tool in your favorite distros and OSes, > >> and I have seen people use those tools more often. > > > > Some platforms, e.g. Minnow, use their own image writing script (i.e. > > mkefidisk.sh). If there are other platforms which need some special tool > > that isn't generically available from the host and the hob tool could be > > made to accommodate them all, I'd prefer to see one tool that can do all > > of them instead of a bunch of tools, one for each platform. > > As Paul mentioned, each platform seems to want to do this "their > own way" which is troublesome at best for embedded systems. I think > it would be much better to maintain one way that works with what we > are building (OE based embedded devices) rather than relying on the > whims of others. So, if it's not too hard, I'd prefer that this > ability be kept, even as a stand-alone script. Well, arguably the "one way that works" is supposed to be wic as far as the actual image writing is concerned. There's a bug open to ensure we support all of the reference platforms [1], and I'd encourage anyone developing a BSP with unusual partition/formatting requirements to work out how those can be supported in wic if they aren't already so we can get away from these BSP- specific classes and scripts. There isn't a frontend UI for wic that I am aware of though - as I mentioned earlier image-writer has no support for it, it's just doing a straight dd to the device. Cheers, Paul [1] https://bugzilla.yoctoproject.org/show_bug.cgi?id=8719 -- Paul Eggleton Intel Open Source Technology Centre