I think I'm running around in circles. First off here is what I'm trying to accomplish - to get linux source codes, cross-compile (targeting ARMv6 (the raspberry pi one)) running off an x86 host, generate a .img file, mount the file on an SD card and boot my pi off it.
I spent some time under the 'developer' section of http://archlinuxarm.org
and following the links, eventually found myself on wiki.archlinux.org. It was a rude shock when the FAQ mentioned that the information was only meant for x86 or 64-x86 systems! Figured things out soon enough; so no worries there.
I might need to review my linux kernel knowhow, but intuitively this is what I feel needs to be done (host is an x86 virtualbox system running ubuntu server 13.04, chosen for the lack of GUI);
1. setup the arm cross-compiler toolchain on my host
Some instructions are available here (http://archlinuxarm.org/developers/dist ... -compiling
) and I think I saw something elsewhere too, but can't find the link now. It was a blog and I remember reading that it would take about an hour to just compile everything needed. I remember it demonstrating how to cross-compile a simple 'hello world' application on an x86 host and copy it over to the ARM system to run.
2. configure the kernel sources to use the cross-compiler instead of the default one and create an image.
I'm going to assume that the output of this step will be the .img file I need to load on the SD card. Problem is; I'm lost on how to configure things to use the cross-compiler created in the earlier step. At the moment, I'm searching the internet for answers.
Any hints on where I should look would be appreciated. Also; I understand that the Raspberry Pi isn't fully open-source and that some sections of the main chip (such as the GPU) are closed-source and only available in binary blobs (I think). I don't mind linking those in as extra modules, but I really want to compile as much of the final .IMG as I can from raw source code.