STM32 CMake Projects

Personal test repository for STM32 projects using the stm32-cmake repository.

git submodule init
git submodule update --init --recursive

Prerequisites

If you have not set up the prerequisites yet, go to the prerequisites chapter for detailed instructions on how to install these.

Here is a brief overview of the required tools to develop software for the STM32H7 microcontroller:

1. CMake build system generator installed
2. Build system like Ninja Build or Make installed.
3. Bare-Metal ARM toolchain installed
4. Recommended for application code development: Eclipse for C/C++ installed with the Eclipse MCU plugin
5. OpenOCD installed for Eclipse debugging
6. STM32 USB drivers installed, separate steps for Windows or Linux

Building the Projects

The project expect the folder containing the toolchain binaries to be in the environment path.

On Windows, when using MinGW64, add -G "MinGW Makefiles" to use MinGW makefiles or -G "Ninja to use the Ninja build system. You can flash the generated binary file to the connect board via drag-and-drop or you can use Eclipse and OpenOCD to perform debugging with the .elf file.

Multi-Board Projects

Supply board: -DBUILD_H743ZI=ON and/or -DBUILD_F303VC=ON.

Minimal Project

cd projects
mkdir build && cd build
cmake -D<BOARD_SELECT> ../minimal
cmake --build . -j

FreeRTOS Project

cd projects
mkdir build && cd build
cmake -D<BOARD_SELECT> ../freertos
cmake --build . -j

H743ZI

LwIP UDP Echoserver Project

cd projects
mkdir build && cd build
cmake ../h743zi/lwip-raw-udp-echoserver
cmake --build . -j

Setting up Prerequisites

CMake

Linux

sudo apt-get install cmake

Windows

On Windows, you can use pacman -S mingw-w64-x86_64-cmake, but you can also install the Windows CMake via the installer. It is recommended to pick the install option Add CMake to system PATH for all users to CMake can be used from the command line. Please note that you need to add the Windows CMake path to the MinGW64 path manually if you want to use it in CMake.

Cross-Compiler

The instuctions here specify how to install and use a specific version of the xPacks cross-compiler but you can use any other ARM cross-compiler which can generate bare-metal code, usually denoted by the arm-none-eabi cross-compiler triplet.

If you don't want to install nodejs you may go with the four-command manual installation.

Windows

On Windows, it is recommended to perform the xpm and toolchain installation from the Windows command line.The simple way required npm, which can be installed by installing NodeJS. Make sure npm can be run from the command line by adding the folder containing npm.exe to the system path and running the following command

npm install --global xpm@latest
xpm install --global @xpack-dev-tools/arm-none-eabi-gcc@11.2.1-1.1.1 --verbose

On Windows, the toolchain binaries will be located in a folder like this

C:\Users\<User>\AppData\Roaming\xPacks\@xpack-dev-tools\arm-none-eabi-gcc\<version>\.content\bin

You can now run the following commands in the repository root:

xpm init
xpm install @xpack-dev-tools/arm-none-eabi-gcc@11.2.1-1.1.1

to create symlinks to the toolchain in ./xpacks/.bin. You can now set up the environment by using . load_path.sh or the following command

export PATH="$(pwd)/xpacks/.bin":$PATH

On Windows, you can use git bash to perform the step above.

Alternatively, you can use the graphical system environmental variables editor to add the .bin path to the system variables permanently or use the appriate command for CMD or PowerShell to update the PATH variable.

Linux

Install npm and nodejs first. Example for Ubuntu according to this guide.

curl -sL https://deb.nodesource.com/setup_14.x | sudo -E bash -
sudo apt-get update && sudo apt-get install nodejs

Check that npm is installed with npm --version. Then xpm and the cross-compiler are installed.

sudo npm install --global xpm@latest
xpm install --global @xpack-dev-tools/arm-none-eabi-gcc@11.2.1-1.1.1 --verbose

You can now run the following commands in the repository root:

xpm init
xpm install @xpack-dev-tools/arm-none-eabi-gcc@11.2.1-1.1.1

to create symlinks to the toolchain in ./xpacks/.bin. You can now set up the environment by using . load_path.sh or the following command

export PATH="$(pwd)/xpacks/.bin":$PATH

Build System

It is recommended to use ninja or make as the software build system.

Windows

It is recommended to use the Ninja build system. Download the ninja executable and place it somewhere. You need to add the folder containing the ninja executable to the system environmental variables so it can be used in the build process. You can test whether ninja works by running ninja --version in the command line.

Alternatively or additionally, you can also install mingw32-make which comes bundled with MinGW64. Make sure to add the binaries path to the Windows path during installation. Otherwise, you need to add msys64/mingw64/bin to the Windows path so you can run installed binaries from the command line. Open the MinGW64 shell and run the following commands.

pacman -S mingw-w64-x86_64-toolchain mingw-w64-x86_64-make mingw-w64-x86_64-cmake

You can test successfull installation with mingw32-make -v from the Windows Command Line.

Linux

On Linux, make is pre-installed and it is recommended to use it directly. You can install ninja with

sudo apt-get install ninja-build

USB Drivers

Windows

Install the STM32 USB drivers.

Linux

Install the USB drivers on Linux. On Ubuntu, you can run the following command to install it:

sudo apt-get install stlink-tools

Debugging the application

Some basic OpenOCD launch configurations were provided. It is expected that the user starts the OpenOCD application.

For example, to open an OpenOCD session specifically for the STM32H743ZI-Nucleo, perform the following steps

cd projects/h743zi
openocd

Now you can flash the application with a GDB application like arm-none-eabi-gdb or gdb-multiarch. You can try this in the command line as well (here with gdb-multiarch):

cd projects
gdb-multiarch -q -x h743zi/openocd.gdb build/stm32.elf

Testing the LwIP UDP Echoserver

Connect the STM32H743ZIT to your router via Ethernet. The STM32 device should now get an IP address via DHCP. Check that the IP address in your router configuration or retrieve it from the serial console. The server is listening on requests on port 7. You can test it with the following command, using netcat:

echo "Hello World" | nc -w1 -u <ip-address> 7

Using VS Code

VS code files were supplied to perform debugging and building with VS code after the build system was generated. The only prerequsite is to install the Cortex-Debug plugin. You can do this in the command line with

code --install-extension marus25.cortex-debug