Lazuli is a preemptive multitasking RTOS (Real-Time Operating System/kernel), targeting AVR microcontrollers. It allows to program and run multiple independent tasks in a real-time context on AVR microcontrollers. In its basic configuration, Lazuli RTOS has a low memory footprint (< 4 kilobytes of ROM, < 200 bytes of static RAM). Lazuli RTOS can be used to develop embedded applications or firmwares that have strong real-time constraints or demand multitasking, on minimal hardware.

AVR MCUs are widely used in embedded, industrial devices and applications, and are also known to be used on Arduino boards. Applications written on top of Lazuli RTOS suit industrial systems as well as hobbyists creations.

Lazuli RTOS is distributed in its source code form, to be built and statically linked with your own final executable. This allows the kernel to be statically configured, and benefit from compiler optimizations.

The project is hosted on GitHub at https://github.com/randruc/Lazuli

• Preemptive multitasking of user tasks
• Real-time scheduling
• No MMU required
• "ROMable": all the system can fit in ROM
• Highly configurable, with a modular architecture
• A "wait for event" programming model
• A complete and containerized development environment
• A complete documentation: API documentation, user and kernel documentation

For now, the Lazuli kernel provides the following functionalities:

• Rate Monotonic Scheduling
• Priority Round-Robin Scheduling (POSIX's SCHED_RR)
• Mutexes
• AVR USART driver, including a complete printf() implementation
• Software timers

The Lazuli project has very specific goals that make it unique. These are:

• Free software - This project is developed with and can be built using only free software.
• Pure ANSI C (89) - Lazuli is coded in pure C89 for all of its C parts. It means that no compiler-specific code nor extensions to the C language are used. This choice has been made the allow to code to be ported easily (if needed in the future) to many platforms and architectures, and to be compiled by the largest number of compilers. Another reason is that C89 is understood by all C code-checking tools.
• Well documented - The Lazuli project aims to be a well-documented project. All the code is carefully documented and comes with good documentation about how it works and how to use it.

Lazuli RTOS currently runs on the ATmega328p MCU (that is used on the Arduino) but it should be easily portable to other AVR platforms.

Although Lazuli tends to reach a high level of code quality and stability, it does not suit safety-critical systems as it is not certified for those specific uses. For more information about what are safety-critical systems, please read https://en.wikipedia.org/wiki/Safety-critical_system

The documentation for the project can be read at https://lazuli.readthedocs.io/en/latest/

The API documentation can be read at https://randruc.github.io/Lazuli/doxygen/latest/

Two different tools are used to document the project:

• sphinx to generate user documentation from files written in reStructuredText. The sources are in the doc/ directory.
• doxygen to generate API documentation from C code comments.

Example programs that use the Lazuli RTOS can be found in the directory example-programs/.

Here is shown the classic blinking LED. The Hello World of embedded systems. It is heavily commented to serve as a introduction to the Lazuli API.

#include <stdint.h>

#include <Lazuli/lazuli.h>
#include <Lazuli/sys/arch/AVR/registers.h>

/*
 * This is the Blink task. It simply blinks the built-in LED on Arduino
 * platforms.
 * This task is scheduled in real-time. It is configured to blink with an exact
 * period of 1 second.
 */
void
Blink()
{
  /* On the Arduino, this pin corresponds to the built-in LED */
  const uint8_t ledPin = 0x20;

  DDRB |= ledPin;   /* Set the pin to be an output pin        */
  PORTB &= ~ledPin; /* The initial state of the pin will be 0 */

  /* Now this is the main loop of this task */
  for (;;) {
    /* Wait for the next real-time activation of the task */
    Lz_Task_WaitActivation();

    PINB |= ledPin; /* Toggle the pin */
  }
}

void
main(void)
{
  /* Allocate a configuration object on the stack                      */
  Lz_TaskConfiguration configuration;

  /* Initialize the configuration object with default values           */
  Lz_TaskConfiguration_Init(&configuration);

  /* Configure the Blink task to be cyclic real-time (RMS scheduling)  */
  configuration.schedulingPolicy = CYCLIC_RT;

  /* The Blink task has a period of 25 time slices.                    */
  /* Our platform has a 16 MHz clock, and the system clock resolution  */
  /* frequency is configured to 50 Hzへるつ. This is an arbitrary value that */
  /* can be configured by the user.                                    */
  /* With a system clock resolution frequency set to 50 Hzへるつ, the system */
  /* clock period is then 1 / 50 = 0.02 second.                        */
  /* So 0.02 * 25 = 0.5 second, which corresponds to the half period   */
  /* of our task.                                                      */
  configuration.period = 25;

  /* The Blink task has a completion of 10 time slices (arbitrary here */
  /* because our task does almost nothing).                            */
  configuration.completion = 10;

  /* Register the Blink task to run with the parameters above          */
  Lz_RegisterTask(Blink, &configuration);

  /* Run the system                                                    */
  Lz_Run();
}

Lazuli has no dependency on any other existing code. You can simply write your own code, build the system, upload it to the target MCU and it just runs!

To develop with Lazuli, it is strongly recommended to use the Lazuli Docker image. You will benefit from a complete development environment, provided with all the necessary tools.

Lazuli RTOS comes with a complete and containerized development environment provided as a Docker image. This image includes all the tools that are necessary to build your own project using Lazuli RTOS. It includes compilers and linkers, build tools, binary utilities, man pages, etc.

The official Lazuli Docker image can be pulled from https://hub.docker.com/r/randruc/lazuli

Read more on how to set up the development environment in the official documentation: https://lazuli.readthedocs.io/en/latest/set_up_environment.html

Lazuli development environment container starting

Man pages in Lazuli development environment container

The Lazuli kernel is fully configurable. The build system relies on CMake. With the help of ccmake, configuration can also be done interactively in the console.

Read more in the official documentation: https://lazuli.readthedocs.io/en/latest/developing_your_project.html

Configuration using ccmake

Building using cmake

On AVR MCUs, avrdude can be used to upload the final binary to the target machine. The script scripts/AVR/upload.sh can be used for that. It takes the HEX file as a parameter.

Interaction with the serial line can be done with the help of GNU screen. The script scripts/serial.sh can be used to interact with the USB serial line.

The Issues tab (https://github.com/randruc/Lazuli/issues) of the project must be used to report troubleshooting or to make suggestions.

Contributions are welcome!

The project is hosted on GitHub (at https://github.com/randruc/Lazuli), and GitHub is used to manage it all: pull request, issues, etc. It doesn't matter if you wish to fix a bug, implement or suggest new features, or fix a typo/spelling mistake: any kind of contribution is welcome!

Read more on how to contribute in the official documentation: https://lazuli.readthedocs.io/en/latest/kernel/contributing.html

You can also use the Issues tab of the project to ask questions, suggest features without coding, or anything you wish!

All the project is distributed under the GNU General Public License v3.0 only. A full copy of this license is available in the file LICENSES/GPL-3.0-only.txt.