MicroPython external C modules¶
When developing modules for use with MicroPython you may find you run into limitations with the Python environment, often due to an inability to access certain hardware resources or Python speed limitations.
If your limitations can’t be resolved with suggestions in Maximising MicroPython speed, writing some or all of your module in C (and/or C++ if implemented for your port) is a viable option.
If your module is designed to access or work with commonly available hardware or libraries please consider implementing it inside the MicroPython source tree alongside similar modules and submitting it as a pull request. If however you’re targeting obscure or proprietary systems it may make more sense to keep this external to the main MicroPython repository.
This chapter describes how to compile such external modules into the MicroPython executable or firmware image.
An alternative approach is to use Native machine code in .mpy files which allows writing custom C code that is placed in a .mpy file, which can be imported dynamically in to a running MicroPython system without the need to recompile the main firmware.
Structure of an external C module¶
A MicroPython user C module is a directory with the following files:
*.hsource code files for your module.
These will typically include the low level functionality being implemented and the MicroPython binding functions to expose the functions and module(s).
Currently the best reference for writing these functions/modules is to find similar modules within the MicroPython tree and use them as examples.
micropython.mkcontains the Makefile fragment for this module.
$(USERMOD_DIR)is available in
micropython.mkas the path to your module directory. As it’s redefined for each c module, is should be expanded in your
micropython.mkto a local make variable, eg
EXAMPLE_MOD_DIR := $(USERMOD_DIR)
micropython.mkmust add your modules source files relative to your expanded copy of
SRC_USERMOD += $(EXAMPLE_MOD_DIR)/example.c
If you have custom compiler options (like
-Ito add directories to search for header files), these should be added to
CFLAGS_USERMODfor C code and to
CXXFLAGS_USERMODfor C++ code.
See below for full usage example.
This simple module named
cexample provides a single function
cexample.add_ints(a, b) which adds the two integer args together and returns
the result. It can be found in the MicroPython source tree
in the examples directory
and has a source file and a Makefile fragment with content as descibed above:
micropython/ └──examples/ └──usercmodule/ └──cexample/ ├── examplemodule.c └── micropython.mk
Refer to the comments in these 2 files for additional explanation.
Next to the
cexample module there’s also
works in the same way but shows one way of mixing C and C++ code
Compiling the cmodule into MicroPython¶
To build such a module, compile MicroPython (see getting started), applying 2 modifications:
USER_C_MODULESset to the directory containing all modules you want included (not to the module itself). For building the example modules which come with MicroPython, set
examples/usercmoduledirectory. For your own projects it’s more convenient to keep custom code out of the main source tree so a typical project directory structure will look like this:
my_project/ ├── modules/ │ └──example1/ │ ├──example1.c │ └──micropython.mk │ └──example2/ │ ├──example2.c │ └──micropython.mk └── micropython/ ├──ports/ ... ├──stm32/ ...
USER_C_MODULESset to the
all modules found in this directory will be compiled, but only those which are explicitly enabled will be availabe for importing. Enabling a module is done by setting the preprocessor define from its module registration to 1. For example if the source code defines the module with
MP_REGISTER_MODULE(MP_QSTR_cexample, example_user_cmodule, MODULE_CEXAMPLE_ENABLED);
MODULE_CEXAMPLE_ENABLEDhas to be set to 1 to make the module available. This can be done by adding
makecommand, or editing
#define MODULE_CEXAMPLE_ENABLED (1)
Note that the exact method depends on the port as they have different structures. If not done correctly it will compile but importing will fail to find the module.
To sum up, here’s how the
cexample module from the
directory can be built for the unix port:
cd micropython/ports/unix make USER_C_MODULES=../../examples/usercmodule CFLAGS_EXTRA=-DMODULE_CEXAMPLE_ENABLED=1 all
The build output will show the modules found:
... Including User C Module from ../../examples/usercmodule/cexample Including User C Module from ../../examples/usercmodule/cppexample ...
Or for your own project with a directory structure as shown above, including both modules and building the stm32 port for example:
cd my_project/micropython/ports/stm32 make USER_C_MODULES=../../../modules \ CFLAGS_EXTRA="-DMODULE_EXAMPLE1_ENABLED=1 -DMODULE_EXAMPLE2_ENABLED=1" all
Module usage in MicroPython¶
Once built into your copy of MicroPython, the module can now be accessed in Python just like any other builtin module, e.g.
import cexample print(cexample.add_ints(1, 3)) # should display 4