Question

I already have a functional embedded linux system using MontaVista. My embedded board boots linux and runs the included binaries just fine. I know how to make a small C program and compile it with a Makefile that calls arm_v5t_le-gcc instead of gcc.

That is all simple enough, but how the heck do you configure and install a “standard” linux package that requires ./configure, make, and make install?

The main conceptual problem I have is that ./configure collects data about your current system like compilers and include files and such. When I run ./configure, it is just going to collect info about my desktop linux environment, not my embedded environment.

Answer 1

I am answering my own question. It turns out you can run ./configure -h to see some options, some of which allow you to describe your embedded (or cross compile) environment.

In my specific case, I wanted to compile libvisca (a Sony RS232 camera control library) for my embedded ARM environment that runs MontaVista. I ran the following commands to get it working:

export CC=arm_v5t_le-gcc
export AR=arm_v5t_le-ar

./configure --host=arm

make

So long as you have the above compilers in your $PATH, you’re all set. Happy compiling.