Merged in raiden00/nuttx (pull request #668)

Master

* libdsp: initial commit

* libdsp: cosmetics

* stm32f334-disco/buckboost: use a PID controller from libdsp

* stm32_adc.h: fix typo

* stm32_dac.c: set OUTEN bit for DAC1CH2 and DAC2CH1

* stm32_hrtim: cosmetic changes

* power/motor: direction parameter is now int8 + add overload fault

* libdsp: all floats with f-sufix

    libdsp: add precision option for library

    libdsp: add debug option for library and assertions in functions

    libdsp: add current samples correction for SVM3

    libds: add some motor control specific functions

    libdsp: add basic speed observer

    libdsp: fix phase shift in SMO observer

    libdsp: add more logic to FOC

    config/sim/dsptest: add dsptest configuration

* libdsp/lib_motor.c: remove unused comparation

* libdsp/lib_observer.c: update some comments

Approved-by: GregoryN <gnutt@nuttx.org>

libs/libdsp/Kconfig

@@ -8,3 +8,23 @@ config LIBDSP
 	default n
 	---help---
 		Enable build for various DSP functions
+
+config LIBDSP_DEBUG
+	bool "Libdsp debugging"
+	depends on LIBDSP
+	default n
+	---help---
+		Enable debugging for libdsp. This option enables additional parameters
+		checking. It can drastically reduce performance and be potentially
+		dangerous to hardware, so it should be used carefully (probably only
+		at an early stage of application development).
+
+config LIBDSP_PRECISION
+	int "Libdsp precision [0/1/2]"
+	default 0
+	---help---
+		Whith this option we can select libdsp precision for
+		some of calculations. There are 3 available options:
+		0 - the fastest calculation but the lowest precision
+		1 - a little better precision than above, but slowest
+		2 - the most accuracte but the slowest one, use standard math functions.