From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from metis.ext.pengutronix.de ([2001:67c:670:201:290:27ff:fe1d:cc33]) by bombadil.infradead.org with esmtps (Exim 4.89 #1 (Red Hat Linux)) id 1eqwGL-0006aL-5z for barebox@lists.infradead.org; Wed, 28 Feb 2018 07:35:26 +0000 Date: Wed, 28 Feb 2018 08:34:39 +0100 From: Sascha Hauer Message-ID: <20180228073439.bnqfnncm4zvhxnzw@pengutronix.de> References: <20180226082349.31568-1-s.hauer@pengutronix.de> <20180226101201.k7ync5dx3yaynslg@localhost.localdomain> <20180227083430.ze66vwf42ytipjsf@pengutronix.de> <20180227194358.bpppmhqcunm2gb6a@localhost> MIME-Version: 1.0 Content-Disposition: inline In-Reply-To: <20180227194358.bpppmhqcunm2gb6a@localhost> List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Sender: "barebox" Errors-To: barebox-bounces+u.kleine-koenig=pengutronix.de@lists.infradead.org Subject: Re: [PATCH] Pass barebox version to kernel To: Peter Mamonov Cc: Barebox List On Tue, Feb 27, 2018 at 10:43:59PM +0300, Peter Mamonov wrote: > On Tue, Feb 27, 2018 at 09:34:30AM +0100, Sascha Hauer wrote: > > Hi Peter, > > > > On Mon, Feb 26, 2018 at 01:12:02PM +0300, Peter Mamonov wrote: > > > Hi, Sasha, > > > > > > On Mon, Feb 26, 2018 at 09:23:49AM +0100, Sascha Hauer wrote: > > > > When userspace is interested in the barebox version it has currently no > > > > way of reliably reading it. Add it to the kernel command line as it's an > > > > established way to pass information from the bootloader to the kernel. > > > > If CONFIG_FLEXIBLE_BOOTARGS is enabled then the barebox version is > > > > passed in the "bootloader.version=" variable. > > > > > > Some time ago we solved a similar problem: a number of parameters including > > > barebox version, MAC address (which may be random due to the lack of a NIC > > > EEPROM) and some vendor specific parameters are passed to the kernel via DTB. > > > A dedicated command was implemented which can either patch the existing DTB or > > > generate an overlay DTB. In the latter case the overlay DTB is passed to the > > > kernel with the help of a new `bootm` option. Of course the latter approach > > > requires support on the kernel side. > > > > We could of course pass the barebox version in the /chosen node. That > > would require a of_register_fixup(). Why would we need an extra command > > for that? > > Well, it allows some extra flexibility: either original DTB is patched or a > separate DTB blob is generated. However there is no strict need for a command. > > My actual point is that passing various bootloader stuff to a kernel via DTB > feels like a cleaner solution, rather than using kernel cmdline for that > purpose. I tend to buy that argument. Especially when the stuff the bootloader wants to pass becomes more and more then the device tree seems like a good place. The downside is that not all boards have a devicetree (not counting legacy boards here, but for example UEFI boards). Any other opinions? We could also do both. If we only pass the version then this would be ok I guess, but I have no idea where this leads to. Sascha -- Pengutronix e.K. | | Industrial Linux Solutions | http://www.pengutronix.de/ | Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 | _______________________________________________ barebox mailing list barebox@lists.infradead.org http://lists.infradead.org/mailman/listinfo/barebox