A simple and cross-platform GCC setup for nRF51-series development for Windows, OS X and Linux
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
examples/pca10001
template
.gitignore
LICENSE
README.md

README.md

nrf51-pure-gcc-setup

A simple and cross-platform GCC setup for nRF51 development. I've developed it on Linux, but should also work fine on OS X and on Windows. Beware that neither of those are regularly used by me, though. Feel free to submit a pull request if you find something that doesn't work.

This was started (long) before the official SDK got GCC support, but I've kept maintaining it, since I don't really like the official Makefiles...

The currently supported SDK version is 6.0.0. Support for some other versions are available as branches.

Usage

Make a pure-gcc subfolder in your project folder (i.e. alongside the arm/gcc/iar folders from the SDK) and copy an example Makefile from the examples/ folder into it. Make sure that the SDK_PATH and TEMPLATE_PATH points towards your nRF51 SDK installation and the template/ subfolder of your clone of this repository respectively.

Project options

All source files, if they are in the application folder (../ from the Makefile) or somewhere in the SDK should be possible to include directly, without a path. If you want to add some other path to the search paths, add it to LIBRARY_PATHS (header files) or SOURCE_PATHS.

Make sure that the USE_SOFTDEVICE variable is set to whichever softdevice is currently programmed on your board, lowercase. The value is directly used in paths, so if it's uppercase, you'll most likely end up with files not being found, at least on OS X/Linux. Currently S110, S120, S210 and S310 should be supported, although only S110 have been well-tested.

If you use the boards.h file, make sure that the BOARD define is set to the board you're currently using.

If you have multiple J-Links connected to your system, you should set the SEGGER_SERIAL variable to the serial number of your J-Link, so that the programming Just Works (tm). It seems Segger on Linux isn't capable of giving you a selection of J-Links, as on Windows.

If you want to use the GDB functionality with multiple J-Links, you should make sure that all projects have a unique GDB port number defined in their project Makefile.

You can override most of the defines from the other Makefiles by setting them in your project file, since most of them use ?=.

Targets

Most of the targets provided should be self explanatory, but some may use some extra explanation:

erase-all:

Does an erase all of a chip.

recover:

Provides equal functionality to that of nrfjprog / nRFgo Studio on Windows.

flash-softdevice SOFTDEVICE=$(PATH_TO_SOFTDEVICE_WITHOUT_SPACES)

Can be used to flash a softdevice to a chip. The path to the softdevice hex needs to be without spaces, due to Make limitations.

startdebug:

Starts a J-Link GDB Server in a separate terminal window, and then GDB also in a separate window. If you change the code, you can then make directly from gdb, and do load to run the new code.