The radio init should wait for a PLL lock IRQ, then subsequently clear it, it prevent timing-dependent radio instability after reset.

Some files were included without other headers and modules on which they depend:

attitude.c : no sys_clock.h
cam.c : no led.h
controller.c : no digital filter module
pid-hw.c : no pid.h

I suggest we establish a requirement for the imageproc-lib that the code set in 'master' should be compilable as a library project, with no warning or errors.

There's a new hardware pid module (pid-hw.c/h) that supersedes the float-based software one (controller.c/h). We need to define the name for an abstraction layer (pid.c/h? pid_cont.c/h?) and then the names of the hardware and software sub-modules (incl. both float and integer versions). Please suggest names below.

There are very few LEDs on our boards and we've recently started toggling them during the normal operation of some modules (e.g. radio). The usage of LEDs within traps (e.g. battery) makes a bit more sense, although we should always allow a project to easily override this (i.e. by setting its own callback).

I suggest we stop controlling LEDs from library modules, so that a particular user/project can freely make use of all LEDs. When developing a module, though, you can feel free to pepper it with LED toggles, but these should be removed once a module is committed as stable.

Let me know what you think.

cc. @Humhu, @ryanjulian, @abuchan, @dhaldane, @apullin, @ronf-ucb.

Since this dsp module implements just a digital filter, renaming it should be fairly trivial, but entails renaming the internal functions in the process. Please suggest a filename below.

The stock of ITG3200 gyroscope chips in the lab has been depleted, and we are probably not going to order more. The MPU6000 and 6050 will take the place of the the ITG3200 for now.
A driver that will function the same as the ITG3200 driver, insofar as retrieving gyroscope values, needs to be written, tested, and pushed ASAP.

Assign Humphrey and Ryan to this issue.