From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from mail-ee0-x233.google.com ([2a00:1450:4013:c00::233]) by bombadil.infradead.org with esmtps (Exim 4.80.1 #2 (Red Hat Linux)) id 1WfFLc-00033t-Q8 for barebox@lists.infradead.org; Tue, 29 Apr 2014 21:13:53 +0000 Received: by mail-ee0-f51.google.com with SMTP id c13so736118eek.10 for ; Tue, 29 Apr 2014 14:13:30 -0700 (PDT) Date: Tue, 29 Apr 2014 23:13:14 +0200 From: Beniamino Galvani Message-ID: <20140429211313.GA15720@gmail.com> References: <1398591044-3616-1-git-send-email-b.galvani@gmail.com> <20140428072627.GN5858@pengutronix.de> <20140428205452.GA29091@gmail.com> <20140429070557.GB5858@pengutronix.de> MIME-Version: 1.0 Content-Disposition: inline In-Reply-To: <20140429070557.GB5858@pengutronix.de> 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 00/11] ARM: add initial support for Rockchip boards To: Sascha Hauer Cc: barebox@lists.infradead.org, Heiko Stuebner On Tue, Apr 29, 2014 at 09:05:58AM +0200, Sascha Hauer wrote: > On Mon, Apr 28, 2014 at 10:54:53PM +0200, Beniamino Galvani wrote: > > On Mon, Apr 28, 2014 at 09:26:27AM +0200, Sascha Hauer wrote: > > > Hi Beniamino, > > > > > > On Sun, Apr 27, 2014 at 11:30:33AM +0200, Beniamino Galvani wrote: > > > > This series adds an initial support for Rockchip SoCs and has been > > > > tested on a Radxa Rock board, on which I'm able to load a kernel from > > > > the network and boot it [1]. > > > > > > > > At the moment Barebox must be chainloaded from the Rockchip binary > > > > bootloader which performs low-level initializations and loads Barebox > > > > from the "boot" partition on the NAND. > > > > > > > > Barebox should be written using the same procedure used for kernels: > > > > it must be prepared with the mkimage tool and then written with > > > > rkflashkit. > > > > > > > > There is a u-boot code released by Rockchip [2] which probably > > > > includes all the low-level initializations but I'm not brave enough to > > > > try it. > > > > > > > > The patchset adds ethernet and pinctrl drivers, PLL and clocks > > > > initialization, and code to power on the external PHY of the board > > > > through the PMIC. > > > > > > Awesome! I'm happy to see barebox support for one of the more popular > > > Linux ARM architectures. > > > > > > The patches look quite good and there's not much to be done to > > > make them ready for merging. > > > > > > I'm just on the way to merge the Linux devicetree files into barebox > > > and use them where possible so we do not duplicate the devicetrees in > > > barebox. Could you post a followup once to base the rockchip dts files > > > on the Linux dts files once I have everything in place? I saw that you > > > already use the mainline dts files, but these do not contain the > > > ethernet nodes for example. > > > > Ok, I will do. Just a question: my series is against -next because it > > requires commit 6720ad6f16db5839a72aa8b53e89918a4f0059bd "clk: move > > of_clk_get_parent_name() to common clk code", while your dts branch > > derives from master and doesn't have this commit. Which branch should > > I base the next version on? > > I reshuffled the patches in -next so that your patches have the correct > dependencies. Thanks. > > > > > > Also I'd like to let rockchip use the multi image mechanism. This is not > > > much work when you know what to do. You could change it yourself or I do > > > the change for you, but in this case, could you give the result a test? > > > > I'm not familiar with multi-image support but I would be glad to try > > it if you're willing to implement it. > > Ok, I pushed everything to -next. The multi image conversion is in the > attached patch (also included in -next). Try building rockchip_defconfig > and start images/barebox-radxa-rock.img. If it works then we are lucky, > if not we'll have to work out a way to debug the patch. The image works. There is only a small change (below) to do in board file to postpone the call of hostname_init() until the dt root node is set, otherwise the check on board compatible always returns false. > > BTW I noticed the upstream radxa rock dts file does not have a board > specific compatible. I added this in my patch, but this should be done > upstream. CC'ing Heiko, who is the Linux Rockchip maintainer. Beniamino -----------------------------8<------------------------------------- >From 829e4a021e537d704e49aaa2b885ea9fb5484007 Mon Sep 17 00:00:00 2001 From: Beniamino Galvani Date: Tue, 29 Apr 2014 20:11:46 +0200 Subject: [PATCH] ARM: radxa-rock: call hostname_init() after DT initialization --- arch/arm/boards/radxa-rock/board.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/arm/boards/radxa-rock/board.c b/arch/arm/boards/radxa-rock/board.c index 691f243..3d9b5be 100644 --- a/arch/arm/boards/radxa-rock/board.c +++ b/arch/arm/boards/radxa-rock/board.c @@ -84,4 +84,4 @@ static int hostname_init(void) return 0; } -core_initcall(hostname_init); +postcore_initcall(hostname_init); -- 1.7.10.4 _______________________________________________ barebox mailing list barebox@lists.infradead.org http://lists.infradead.org/mailman/listinfo/barebox