Install Mbed OS components

This section explains how to install all the relevant Mbed OS components to run the unit test suite on Arm Fast Models. Follow these instructions to set up Mbed OS and related tools:

  1. Create a directory.

    Note: The default character limit for the path name on Windows 10 is 260 characters for any path. The path name cannot have any spaces in it, or compiling the Mbed OS tests fails.

  2. Perform a git clone to install the main Mbed OS library that containing source code and test files. Verify that the version is at or above 1.8.2. Use these commands:

    git clone https://github.com/ARMmbed/mbed-os
    Mbed –version
  3. Perform a git clone to install the Greentea test harness. Verify that the version is at or above 1.4.0, and that the mbedgt --fm output says that you need an argument. Use the following code:

    git clone https://github.com/ARMmbed/mbed-os-tools/       
    cd mbed-os-tools/greentea/packages/mbed-greentea
    sudo python setup.py install
    cd ../../../..
    mbedgt --version
    mbedgt --fm
  4. Download and run the installer to install the Mbed CLI interface on Windows. Run the following command on Linux and Mac to install with the Python pip:

    sudo pip install mbed-cli
  5. Perform a git clone with the following command to install the mbed-fastmodel-agent:
    git clone

    https://github.com/ARMmbed/mbed-fastmodel-agent

    The mbed-fastmodel-agent allows Arm Fast Models to be used as targets for the Mbed OS tests and software.
  6. Edit the file mbed-fastmodel-agent/fm_agent/settings.json to set up the mbed-fastmodel-agent correctly. Change the following variables:
    • Change these variables under GLOBAL > Windows or GLOBAL >Linux, depending on your system type:

      “model_lib_path”: <path-to-FM-install>/examples/LISA/FVP_MPS2/Build_Cortex-M3/Win64-Release-VC2015
      “PyCADI_path”: <path-to-FM-install>/lib/python27
    • Change the model_lib variable for your OS under FVP_MPS2_M3 to match either the .dll name for Windows or .so name for Linux referring to the generated Fast Model binary. For example, on Windows use “model_lib”: “cadi_system_Win64-Release-VC2015.dll”
  7. Build the agent and verify that the program can find the Fast Model executable correctly. Use these commands:

    cd mbed-fastmodel-agent
    python setup.py install
    cd ..
    mbedfm
    You should then see an output like what you can see in the following screenshot, indicating that the Mbed tools can locate the Cortex-M3 Fast Model executable:

    Fast Model executable output screenshot

Previous Next