natecarlson / rm5xx-modem-buildenv

How to set up a build environment for the modem

Want to build binaries that you can run on your Quectel RM520? Here you go!

Running Docker image

I've tried many ways to build binaries for the modem, and had all sorts of issues with cross-compilation/etc. I finally have a working setup.. using Ubuntu 20.04 ARMHL running in Docker with qemu emulating the ARM instructions. Here's how you do it, assuming your host OS is a recent version of Ubuntu (in my case, 22.04 running on WSL2.)

• Install Docker, as explained at https://docs.docker.com/engine/install/ubuntu/
• Install qemu-user-static and binfmt-support: sudo apt install qemu-user-static binfmt-support
• Register the qemu-user-static interpreters with Docker: docker run --rm --privileged multiarch/qemu-user-static --reset -p yes

Now, your Docker install should be set up to run arm binaries. You can either use the plain Ubuntu image and add what you need:

docker run --name rm520-modem-buildenv -it arm32v7/ubuntu:20.04

Or, you can use my pre-built image, which has the build dependencies already installed:

docker run --name rm520-modem-buildenv -it ghcr.io/natecarlson/rm520-modem-buildenv:main

Either image will dump you to a bash prompt; when you exit the container, the container will be stopped. You can restart it with docker start rm520-modem-buildenv and attach to it with docker attach rm520-modem-buildenv. If you want to remove it, use docker rm rm520-modem-buildenv.

If you want an additional shell on a running instance, you can use docker exec -it rm520-modem-buildenv bash.

Compiling something

Let's say you would like to compile MicroPython for the modem. Here's how you do it, starting at the root prompt you get after starting the container.. (reference git clone https://github.com/micropython/micropython; you can add -j<NUMTHREADS> to the make commands to parallel build.)

git clone https://github.com/micropython/micropython
cd micropython
make -C mpy-cross/ # Build the required compiler
cd ports/unix/ # Since the module runs on Linux, use the Unix port
make submodules # Pull in the necessary submodules
make # Actually build the code

The compilation should end with something like:

LINK build-standard/micropython
   text    data     bss     dec     hex filename
 414905   32168    6112  453185   6ea41 build-standard/micropython

At which point, you can scp the module to a system that is connected to the modem via adb. (I use a Raspberry Pi 4 running Ubuntu 22.04, which is connected to the modem via USB.) If that's not an option for you, when running the docker image, you can map a local directory to somewhere on the container, and then you'll be able to access the file locally; let's say you want to mount the local directory 'rm520-build' as /root/build on the container, and start in that directory. You will need to either specify the relative or absolute path to rm520-build, like -v /home/nc/tmp/rm520-build:/root/build or -v c:\Users\nc\tmp\rm520-build:/root/build depending on your OS.

docker run -v /path/to/rm520-build:/root/build -w /root/build --name rm520-modem-buildenv -it ghcr.io/natecarlson/rm520-modem-buildenv:main