RIOT/doc/doxygen/src/getting-started.md

Getting started

[TOC]

Downloading RIOT code

You can obtain the latest RIOT code from our Github repository either by downloading the latest tarball or by cloning the git repository.

In order to clone the RIOT repository, you need the Git revision control system and run the following command:

git clone git://github.com/RIOT-OS/RIOT.git

Compiling RIOT

Required Software for Development

A set of common tools and a toolchain for the hardware you target needs to be installed first.

Choosing an Operating System for the Development PC

Most of the RIOT OS developers are using Linux on their development PCs, so you can expect the most streamlined experience here. Other POSIX-compliant OSes such as current versions of Mac OS or the various BSD flavours will also be fine - however, we rely on users to report bugs regarding tooling incompatibilities here. So expect occasional issues for the development branch and please help testing during the feature freeze period, if you develop on Mac OS or BSD.

Native development on Windows machines is not officially supported. What works well is using Linux in a virtual machine, but at much lower performance than running Linux natively. For development using the Windows Subsystem for Linux (WSL) is a good option (installation instructions here), but it is (as of October 2021) not possible to directly access USB devices from Linux. As a result, accessing the serial connection to a board running RIOT, flashing it, and on-chip-debugging from WSL will not be possible. (It is possible to pass through the file system of USB storage device. This should allow flashing boards that have an UF2 compatible from within WSL, but this has not been tested yet.) Hence, WSL users will have to use native Windows tools for accessing the serial connection and flashing the board.

Common Tools

The following tools are required or useful regardless of the architecture and board you are developing for:

• Essential system development tools (GNU Make GCC, standard C library headers)
• git
• GDB in the multiarch variant (alternatively: install for each architecture you target the corresponding GDB package)
• unzip or p7zip
• wget or curl
• python3
• pyserial (linux distro package often named python3-serial or py3-serial)
• Doxygen for building the documentation

@note For each architecture a default tool for flashing and on-chip debugging is listed below - in most cases OpenOCD. However, some boards use different tools, e.g. because a bootloader is installed that allows flashing via the serial interface. Check the board documentation for any details on this. If that documentation contains no info about a flashing tool, the default tool for its architecture is used.

@details Running BOARD=<INSERT_TARGET_BOARD_HERE> make info-programmers-supported in your application folder lists the programmers supported by RIOT for the given board.

Architecture: ARM7 and ARM Cortex M*

• GCC, binutils, and newlib for arm-none-eabi
  • Alternatively: Install docker and export BUILD_IN_DOCKER=1
• OpenOCD for debugging/flashing (most boards)
  • Some boards use UF2 based bootloaders, which require auto-mounting to work with make flash
  • Some boards default to using J-Link for flashing/debugging. Either install that or export PROGRAMMER=openocd to just use OpenOCD instead
• Optional: picolibc for arm-none-eabi to link against picolibc instead of newlib
• Optional: clang to build with TOOLCHAIN=llvm
• Optional: GDB multiarch for debugging
  • If no multiarch package is available, use GDB for arm-none-eabi instead

Architecture: Xtensa

ESP32

• @ref esp32_toolchain "Toolchain for ESP32"
• esptool for flashing
• Optional: OpenOCD and GDB (multiarch version) for @ref esp32_jtag_debugging "debugging via JTAG"

ESP8266

• @ref esp8266_toolchain "Toolchain for ESP8266"
• esptool for flashing
• Optional: GDB (multiarch version) for @ref esp8266_esp_gdbstub "debugging via the gdbstub" interface for the ESP8266

Architecture: AVR

• GCC and binutils for AVR and avrlibc
  • Alternatively: Install docker and export BUILD_IN_DOCKER=1
• avrdude for flashing
• Optional: AVaRICE and GDB (multiarch version) for debugging

Architecture: RISC-V

• GCC, binutils, and newlib for RISC-V (target triple should start with riscv and end with -none-elf or -unknown-elf. Note that most packages are multilib, e.g. riscv64-unknown-elf will likely work fine for 32 bit RISC-V boards)
  • Alternatively: Install docker and export BUILD_IN_DOCKER=1
• OpenOCD for debugging/flashing (some new boards might require a patched version of OpenOCD or a recent build from the git sources
• Optional: picolibc to link against picolibc instead of newlib (recommended)
• Optional: clang to build with TOOLCHAIN=llvm
• Optional: GDB multiarch for debugging

Architecture: MSP430

• GCC, binutils, and newlib for MSP430
  • Alternatively: Install docker and export BUILD_IN_DOCKER=1
• mspdebug for flashing/debugging
  • Optional: MSP Debug Stack for additional board support
• Optional: GDB multiarch for debugging

Architecture: MIPS

• GCC, binutils, and newlib for MIPS
  • Alternatively: Install docker and export BUILD_IN_DOCKER=1
• Check board documentation for flashing and debugging

Architecture: native

• On 64 bit systems: multilib versions for your host compilers, standard C library, and development headers
  • Alternatively: Compile with BUILD_IN_DOCKER=1. Note that for running the executable you still need a multilib system (or 32 bit Linux) with glibc a standard C library.
• A C library supporting the deprecated POSIX.1-2001 ucontext library (e.g. glibc, FreeBSD's libc, Mac OS's libc)
• Optional: GDB for debugging. (Prefer the multiarch version, this will also work for other boards)

The build system

RIOT uses GNU make as build system. The simplest way to compile and link an application with RIOT, is to set up a Makefile providing at least the following variables:

• APPLICATION: should contain the (unique) name of your application
• BOARD: specifies the platform the application should be built for by default
• RIOTBASE: specifies the path to your copy of the RIOT repository (note, that you may want to use $(CURDIR) here, to give a relative path)

Additionally it has to include the Makefile.include, located in RIOT's root directory:

# a minimal application Makefile
APPLICATION = mini-makefile
BOARD ?= native
RIOTBASE ?= $(CURDIR)/../RIOT

include $(RIOTBASE)/Makefile.include

You can use Make's ?= operator in order to allow overwriting variables from the command line. For example, you can easily specify the target platform, using the sample Makefile, by invoking make like this:

make BOARD=iotlab-m3

Besides typical targets like clean, all, or doc, RIOT provides the special targets flash and term to invoke the configured flashing and terminal tools for the specified platform. These targets use the variable PORT for the serial communication to the device. Neither this variable nor the targets flash and term are mandatory for the native port.

For the native port, PORT has a special meaning: it is used to identify the tap interface if the netdev_tap module is used. The target debug can be used to invoke a debugger on some platforms. For the native port the additional targets such as all-valgrind and valgrind exist. Refer to cpu/native/README.md for additional information

Some RIOT directories contain special Makefiles like Makefile.base, Makefile.include or Makefile.dep. The first one can be included into other Makefiles to define some standard targets. The files called Makefile.include are used in boards and cpu to append target specific information to variables like INCLUDES, setting the include paths. Makefile.dep serves to define dependencies.

Unless specified otherwise, make will create an elf-file as well as an Intel hex file in the bin folder of your application directory.

Learn more about the build system in the Wiki

Building and executing an example

RIOT provides a number of examples in the examples/ directory. Every example has a README that documents its usage and its purpose. You can build them by typing

make BOARD=samr21-xpro

or

make all BOARD=samr21-xpro

into your shell.

To flash the application to a board just type

make flash BOARD=samr21-xpro

You can then access the board via the serial interface:

make term BOARD=samr21-xpro

If you are using multiple boards you can use the PORT macro to specify the serial interface:

make term BOARD=samr21-xpro PORT=/dev/ttyACM1

Note that the PORT macro has a slightly different semantic in native. Here it is used to provide the name of the TAP interface you want to use for the virtualized networking capabilities of RIOT.

We use pyterm as the default terminal application. It is shipped with RIOT in the dist/tools/pyterm/ directory. If you choose to use another terminal program you can set TERMPROG (and if need be the TERMFLAGS) macros:

make -C examples/gnrc_networking/ term \
    BOARD=samr21-xpro \
    TERMPROG=gtkterm \
    TERMFLAGS="-s 115200 -p /dev/ttyACM0 -e"

Configuring an application

Many modules in RIOT offer configuration options that will be considered during compile-time.They are modeled as macros that can be overridden by the user. Currently there are two ways of doing this: using CFLAGS or via @ref kconfig-in-riot "Kconfig" (the last one is currently only possible for a subset of modules).

For instructions on how to configure via CFLAGS check the @ref config "identified compile-time configurations". To learn how to use Kconfig in RIOT, please refer to the @ref kconfig-users-guide.

Default configurations

When devices have a common access interface, having a default configuration to enable them across platforms, without having to explicitly specify which modules to include, comes in handy. For this, two pseudomodules are defined:

• saul_default: will enable all the drivers of sensors and actuators that are present in the target platform.

• netdev_default: will enable all the drivers of network devices present in the target platform.

Use Docker to build RIOT

Docker is a platform that allows packaging software into containers that can easily be run on any Linux that has Docker installed.

You can download a RIOT Docker container from the Docker Hub and then use that to build your project making use of all toolchains that we've preinstalled in the container.

Setup

Installing docker

To use the RIOT docker build image, the Docker application needs to be installed on your system. To install Docker, depending on your operating system, use sudo apt-get install docker or a variant.

The user on your computer requires permission to access and use docker. There are two ways to manage this:

• Your OS distribution may create a group called docker. If so, then adding yourself to that group (and logging out and in again) should grant you permission.
• Execute docker with sudo. This is in fact the most secure and recommended setup (see here, here, here and here). No extra setup steps are needed. make should be instructed to use sudo by setting DOCKER="sudo docker" in the command line.

Finally, download the pre-built RIOT Docker container:

# docker pull riot/riotbuild

This will take a while. If it finishes correctly, you can then use the toolchains contained in the Docker container: (from the riot root):

$ docker run --rm -i -t -u $UID -v $(pwd):/data/riotbuild riot/riotbuild ./dist/tools/compile_test/compile_test.py

Usage

The RIOT build system provides support for using the Docker container to build RIOT projects, so you do not need to type the long docker command line every time:

(from the directory you would normally run make, e.g. examples/default)

$ make BUILD_IN_DOCKER=1

If your user does not have permissions to access the Docker daemon:

$ make BUILD_IN_DOCKER=1 DOCKER="sudo docker"

to always use Docker for building, set BUILD_IN_DOCKER=1 (and if necessary DOCKER="sudo docker") in the environment:

$ export BUILD_IN_DOCKER=1

running make without specifying BUILD_IN_DOCKER=1 will still use Docker (because of the environment variable)

Troubleshooting

On some Ubuntu versions a make with BUILD_IN_DOCKER=1 can't resolve the host name of for example github.com. To fix this add the file /etc/docker/daemon.json with the address of your DNS Server.

Generating compile_commands.json e.g. for code completion in IDEs

A compile_commands.json for the selected board can be generated by running inside the application folder the following:

$ make compile-commands

This target will honor the variables controlling the build process such as BOARD, TOOLCHAIN, DEVELHELP, etc. just like the usual build process. This works without actual compilation. By default, the compile_commands.json is placed in the RIOT base directory. This behavior can be overwritten using the COMPILE_COMMANDS_PATH variable by specifying the full absolute path (including file name) of the compile_commands.json instead.

Note: By default, the built-in include search directories of GCC will be explicitly added and flags incompatible with clangd will be dropped. This will allow using clangd as language server out of the box. If this is not desired, run export COMPILE_COMMANDS_FLAGS="" to turn modification of the compile commands off. For a list of available flags, run ./dist/tools/compile_commands/compile_commands.py --help in the RIOT base directory.

Using the native port with networking

If you compile RIOT for the native cpu and include the netdev_tap module, you can specify a network interface like this: PORT=tap0 make term

Setting up a tap network

There is a shell script in RIOT/dist/tools/tapsetup called tapsetup which you can use to create a network of tap interfaces.

USAGE

To create a bridge and two (or count at your option) tap interfaces:

    sudo ./dist/tools/tapsetup/tapsetup [-c [<count>]]