Protothreads

lbuild module: modm:processing:protothread

Protothreads are extremely lightweight stackless threads designed for severely memory constrained systems, such as small embedded systems or wireless sensor network nodes. Protothreads provide linear code execution for event-driven systems implemented in C. Protothreads can be used with or without an underlying operating system to provide blocking event-handlers.

Protothreads provide sequential flow of control without complex state machines or full multi-threading.

Since they implement some kind of cooperative multi-threading, Protothreads are non-preemptable. Therefore, a context switch can only take place on blocking operations, which means you don't need complex synchronization.

Protothreads are also stackless, so local variables are not preserved across context switches, and must instead become member variables of the modm::Protothread subclass

A protothread runs within a single function (modm::Protothread::run()) and cannot span over other functions. A protothread may call normal functions, but cannot block inside a called function. Blocking inside nested function calls is instead made by spawning a separate protothread for each potentially blocking function.

The protothread concept was developed by Adam Dunkels and Oliver Schmidt: http://dunkels.com/adam/pt

Originally ported to C++ for use by Hamilton Jet (www.hamiltonjet.co.nz) by Ben Hoyt, but stripped down for public release.

Example

#include <modm/processing/protothread.hpp>

using Led = GpioB0;

class BlinkingLight : public modm::pt::Protothread
{
public:
    bool
    run()
    {
        PT_BEGIN();

        // set everything up
        Led::setOutput();
        Led::set();

        while (true)
        {
            timeout.restart(100ms);
            Led::set();
            PT_WAIT_UNTIL(timeout.isExpired());

            timeout.restart(200ms);
            Led::reset();
            PT_WAIT_UNTIL(timeout.isExpired());
        }

        PT_END();
    }

private:
    modm::ShortTimeout timeout;
};
// ...

BlinkingLight light;

while (true) {
    light.run();
}

Using Fibers

Protothreads can be implemented using stackful fibers by setting the use_fiber option, which replaces the preprocessor macros and C++ implementations of this and the modm:processing:resumable module with a fiber version.

Specifically, the PT_* and RF_* macros are now forwarding their arguments unmodified and instead relying on modm::this_fiber::yield() for context switching:

#define PT_YIELD() modm::this_fiber::yield()
#define PT_WAIT_WHILE(cond) while(cond) { modm::this_fiber::yield(); }
#define PT_CALL(func) func

The modm::pt::Protothread class is implemented using modm::Fiber<> with the default stack size MODM_PROTOTHREAD_STACK_SIZE. It automatically runs the two virtual methods bool run() and bool update() if they are defined in the protothread class.

There should be no modification of the existing code necessary with the exception that you must replace the main loop calling all protothreads with the fiber scheduler:

int main()
{
    /*
    while(true)
    {
        protothread1.update();
        protothread2.update();
    }
    */
    modm::fiber::Scheduler::run();
    return 0;
}

Restrictions

If the default stack size is too low, you can set MODM_PROTOTHREAD_STACK_SIZE to a higher value, however, this will apply to all protothreads, consuming a lot more memory. Instead, we recommend refactoring the protothread into a fiber function.

All functions and macros work as expected, except for the Protothread::stop() function, which not implementable using fibers. It is left unimplemented, so that you may define it with the behavior most suitable to your use case.

See the modm:processing:resumable module for additional restrictions when calling resumable functions from a protothread.

Options

use_fiber

Implement via Fibers

Default: no
Inputs: [yes, no]
Input Dependency: yes -> modm:processing:fiber

Dependencies