Building Cerowrt on your own Linux machine¶
git clone git://github.com/dtaht/cerofiles.git
There will be a build_cero.sh script in that checkout.
There is also a cero_config file here. Copy cero_config to ~/.cero and then edit it to suit your preferences.
./build_cero.sh init cd wherever_you_chose in ~/.cero for the target dir make -jthe_number_of_processors_you_have_to_spare # But please see note below for an extra step
The first time you run this it will take a VERY long time to build the toolchains, etc. Subsequent runs will be much faster. AND:
NOTE With kernel.org down it is impossible to get a build to summon and compile all the needed packages. There is a temporary mirror of the needed packages available via rsync. Inside the build dir above:
mkdir dl; cd dl; rsync -av huchra.bufferbloat.net::dl .
(this is a temporary expedient)
You will probably need to install extra packages on your host system in order to get a working build (documentation, anyone?)
Given that this builds most of cerowrt against the main git head trees, your build will differ somewhat from what we are making available as release candidates at present. After we get stable, we will tag and freeze the relevant trees.
We note that due to this difference, there may be slight (or even major) differences in the default .config file. It helps to repeatedly cp env/.config onto .config and run make defconfig, and install packages, etc, until they more closely match.
We will improve this script as time goes on, and if you run into trouble, please contact the #bufferbloat irc channel and/or the mailing lists.