From mboxrd@z Thu Jan 1 00:00:00 1970 From: Peter Korsgaard Date: Thu, 08 Mar 2012 22:19:21 +0100 Subject: [Buildroot] BeagleBone patches In-Reply-To: (Frank Hunleth's message of "Thu, 8 Mar 2012 08:50:11 -0500") References: Message-ID: <87ipieeqp2.fsf@macbook.be.48ers.dk> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net >>>>> "Frank" == Frank Hunleth writes: Hi, Frank> I had been sitting on some patches for making br builds for the Frank> BeagleBone. I finally got the time to update and clean them up, Frank> and they should show up to the list soon. I hope that this is Frank> better late than never. Great. I have a beaglebone as well now, but I haven't had much time to play with it yet. Frank> There were some things that I wasn't sure about: Frank> 1. I could not avoid adding a couple files under a Frank> board/beaglebone directory. There aren't that many other Frank> configurations that do this, so I wasn't quite sure how everyone Frank> would like BeagleBone referenced. Please let me know if the Frank> folder structure should change. It's probably not a problem, but post your patches and I'll review in detail. Frank> 2. You'll see that MLO gets built with U-boot rather than with Frank> the X-Loader package. I'm not familiar with why the two were Frank> integrated, but it seems to be the way that TI is doing it now Frank> with the bone so I added an option. Yes, I believe MLO in u-boot is the "new" way, and X-loader is going away long term. Frank> 3. There were a few Linux kernels to choose from. The one that I used Frank> is the currently released version from the beagleboard group (I just Frank> verified a couple days ago). Do you mean git://github.com/hvaibhav/am335x-linux.git? Only very basic am335x support is in mainline yet, so that imho makes sense at the moment. Frank> Rather than commit the kernel configuration that I use with my Frank> project, I pulled in the one from the released kernel. This Frank> seemed more "right", but I did have to modify it to compile in Frank> EXT2 so that it could mount the generated rootfs. I also Frank> modified the kernel config to compress the kernel with gzip Frank> rather than lzo to avoid requiring the person compiling the code Frank> to install lzop. Hmm, that sounds like a buildroot bug. We already build liblzo for the host, so building lzop as well wouldn't be a big deal. Frank> The latter is probably completely trivial, but I have found br's Frank> minimalist dependencies really nice in the past and didn't want Frank> to add another. The one thing that I don't like about this Frank> kernel configuration is that enables an enormous number of Frank> features. I'm guessing that this is an Angstrom-ism, and I Frank> wanted to pass it by the list before spending any time modifying Frank> it. Please post the patches, that will make it easier to give concrete feedback. -- Bye, Peter Korsgaard