Skip to content
No description, website, or topics provided.
C Shell C++ Perl Python
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
arch
board
boot
configs
docs
fs
linux
package
support
system
toolchain
.gitignore
CHANGES
COPYING
Config.in
Config.in.legacy
Makefile
Makefile.legacy
README.md
build.sh

README.md

Buildroot for the DM36x projects

This README contains basic instructions for creating a system image for DM36x boards using Buildroot.

Preparing your system

Buildroot requires several packages to be installed on the system before it can work. On Ubuntu, the following apt-get line is sufficient:

sudo apt-get install git-core bison flex g++ gettext texinfo libncurses5-dev pv libssl-dev lzop gawk

Building for Virt2Real v.1 mass board

First of all set default board configuration:

make virt2real_v1_mass_defconfig

Now you can change some options and build the world by using:

make menuconfig
make
make linux-menuconfig // for kernel configuration issues
make

Now you get uImage and FS image in output/images.

Building for Leopard Board

To build a firmware image, run:

./board/leopardboard/build.sh

This will result in a .fw file being built. This takes a while.

If you are hacking the main image, you probably want more granularity in building everything. First tell Buildroot which configuration to use:

make leopardboard_dm368_defconfig

To build everything, run:

make

The build output is stored in the output/images directory. Subsequent builds are faster. It is possible to perform partial rebuilds. See the Buildroot website for details.

Creating a MicroSD card

After the build completes successfully, you should have a .fw file. Insert a MicroSD card into an attached card reader and run:

umount /media/<partitions that were automounted from the card>
sudo sh ./leopardboard-xyz.fw -f /dev/sdc

where /dev/sdc is the location of the MicroSD card. Be sure to unmount any directories that may have been automounted before running the command or there's a high chance of creating a corrupt image.

Updating the Root FS on target

The Root FS is configured to mount read-only. This is done to prevent data corruption from unsafe shutdowns and as a reminder that all configuration and new packages have to end up in the buildroot configuration.

Since this can be annoying during debug, the rootfs can be mounted by running:

remount

Updating any of the applications

As bugs are fixed, it will be necessary to bring in new versions of packages. Buildroot is configured to not allow this to happen by default. This is good, since you can have a high confidence of being able to reproduce an old build.

If a package needs to be updated, edit the corresponding .mk file. For example, if updating foo, edit package/foo/foo.mk and change the VERSION to be the new version. The version could be a git hash tag or a named tag.

After making and testing the changes, commit the changes to the local buildroot repository. You may also want to tag the new version so that firmware files have nice looking names. (Renaming the .fw files will not change the version numbers reported in the UI)

You can’t perform that action at this time.