Add AVR compatibility header for missing math.h definitions.

- Define float versions of math functions to the double versions
  on AVR (eg. #define sinf sin).
- These macros appear to be missing in older versions of avr-libs.
- Include AP_Math.h rather than math.h to get these definitions.

libraries/AC_PID/AC_PID.cpp

@@ -3,7 +3,7 @@
 /// @file	AC_PID.cpp
 /// @brief	Generic PID algorithm
 
-#include <math.h>
+#include <AP_Math.h>
 #include "AC_PID.h"
 
 // Examples for _filter:

libraries/APM_Control/AP_PitchController.cpp

@@ -7,7 +7,7 @@
 //	License as published by the Free Software Foundation; either
 //	version 2.1 of the License, or (at your option) any later version.
 
-#include <math.h>
+#include <AP_Math.h>
 #include <AP_HAL.h>
 #include <AP_Common.h>
 #include "AP_PitchController.h"

libraries/APM_Control/AP_RollController.cpp

@@ -7,7 +7,7 @@
 //	License as published by the Free Software Foundation; either
 //	version 2.1 of the License, or (at your option) any later version.
 
-#include <math.h>
+#include <AP_Math.h>
 #include <AP_HAL.h>
 
 #include "AP_RollController.h"

libraries/APM_Control/AP_YawController.cpp

@@ -7,7 +7,7 @@
 //	License as published by the Free Software Foundation; either
 //	version 2.1 of the License, or (at your option) any later version.
 
-#include <math.h>
+#include <AP_Math.h>
 #include <AP_HAL.h>
 #include "AP_YawController.h"
 

libraries/APM_PI/APM_PI.cpp

@@ -3,7 +3,7 @@
 /// @file	ACM_PI.cpp
 /// @brief	Generic PI algorithm
 
-#include <math.h>
+#include <AP_Math.h>
 
 #include "APM_PI.h"
 

libraries/AP_Baro/AP_Baro.cpp

@@ -8,7 +8,7 @@
  *   of the License, or (at your option) any later version.
  */
 
-#include <math.h>
+#include <AP_Math.h>
 #include <AP_Common.h>
 #include <AP_Baro.h>
 #include <AP_HAL.h>

libraries/AP_Compass/AP_Compass_HMC5843.cpp

@@ -14,7 +14,7 @@
  */
 
 // AVR LibC Includes
-#include <math.h>
+#include <AP_Math.h>
 #include <AP_HAL.h>
 
 #include "AP_Compass_HMC5843.h"

libraries/AP_HAL/utility/Print.cpp

@@ -23,9 +23,8 @@
 #include <stdio.h>
 #include <string.h>
 #include <stdint.h>
-#include <math.h>
-
 
+#include <AP_Math.h>
 #include "../AP_HAL_Namespace.h"
 #include "Print.h"
 using namespace AP_HAL;

libraries/AP_HAL_AVR_SITL/sitl_barometer.cpp

@@ -7,6 +7,8 @@
  */
 
 #include <AP_HAL.h>
+#include <AP_Math.h>
+
 #if CONFIG_HAL_BOARD == HAL_BOARD_AVR_SITL
 
 #include "AP_HAL_AVR_SITL.h"

libraries/AP_Math/AP_Math.h

@@ -8,6 +8,9 @@
 #include <AP_Common.h>
 #include <AP_Param.h>
 #include <math.h>
+#ifdef __AVR__
+# include <AP_Math_AVR_Compat.h>
+#endif
 #include <stdint.h>
 #include "rotations.h"
 #include "vector2.h"

libraries/AP_Math/AP_Math_AVR_Compat.h

@@ -0,0 +1,153 @@
+// -*- tab-width: 4; Mode: C++; c-basic-offset: 4; indent-tabs-mode: t -*-
+
+#ifndef AP_MATH_AVR_COMPAT_H
+#define AP_MATH_AVR_COMPAT_H
+
+// This file defines the floating-point version of standard C math
+// functions on doubles, if they are not present in avr-libc.
+
+#ifndef cosf
+# define cosf cos
+#endif
+
+#ifndef sinf
+# define sinf sin
+#endif
+
+#ifndef tanf
+# define tanf tan
+#endif
+
+#ifndef fabsf
+# define fabsf fabs
+#endif
+
+#ifndef fmodf
+# define fmodf fmod
+#endif
+
+#ifndef sqrtf
+# define sqrtf sqrt
+#endif
+
+#ifndef cbrtf
+# define cbrtf cbrt
+#endif
+
+#ifndef hypotf
+# define hypotf hypot
+#endif
+
+#ifndef squaref
+# define squaref square
+#endif
+
+#ifndef floorf
+# define floorf floor
+#endif
+
+#ifndef ceilf
+# define ceilf ceil
+#endif
+
+#ifndef frexpf
+# define frexpf frexp
+#endif
+
+#ifndef ldexpf
+# define ldexpf ldexp
+#endif
+
+#ifndef expf
+# define expf exp
+#endif
+
+#ifndef coshf
+# define coshf cosh
+#endif
+
+#ifndef sinhf
+# define sinhf sinh
+#endif
+
+#ifndef tanhf
+# define tanhf tanh
+#endif
+
+#ifndef acosf
+# define acosf acos
+#endif
+
+#ifndef asinf
+# define asinf asin
+#endif
+
+#ifndef atanf
+# define atanf atan
+#endif
+
+#ifndef atan2f
+# define atan2f atan2
+#endif
+
+#ifndef logf
+# define logf log
+#endif
+
+#ifndef log10f
+# define log10f log10
+#endif
+
+#ifndef powf
+# define powf pow
+#endif
+
+#ifndef isnanf
+# define isnanf isnan
+#endif
+
+#ifndef isinff
+# define isinff isinf
+#endif
+
+#ifndef isfinitef
+# define isfinitef isfinite
+#endif
+
+#ifndef copysignf
+# define copysignf copysign
+#endif
+
+#ifndef signbitf
+# define signbitf signbit
+#endif
+
+#ifndef fdimf
+# define fdimf fdim
+#endif
+
+#ifndef fmaf
+# define fmaf fma
+#endif
+
+#ifndef fminf
+# define fminf fmin
+#endif
+
+#ifndef truncf
+# define truncf trunc
+#endif
+
+#ifndef roundf
+# define roundf round
+#endif
+
+#ifndef lroundf
+# define lroundf lround
+#endif
+
+#ifndef lrintf
+# define lrintf lrint
+#endif
+
+#endif  // !defined AP_MATH_AVR_COMPAT_H

libraries/AP_PerfMon/AP_PerfMon.cpp

@@ -1,5 +1,5 @@
 
-#include <math.h>
+#include <AP_Math.h>
 #include <FastSerial.h>
 #include "Arduino.h"
 #include "AP_PerfMon.h"
@@ -216,4 +216,4 @@ void AP_PerfMon::DisplayAndClear(uint32_t display_after_seconds)
         DisplayResults();
         ClearAll();
     }
-}
+}

libraries/Filter/DerivativeFilter.cpp

@@ -11,7 +11,7 @@
 /// See http://www.holoborodko.com/pavel/numerical-methods/numerical-derivative/smooth-low-noise-differentiators/
 //
 #include <inttypes.h>
-#include <math.h>
+#include <AP_Math.h>
 #include <Filter.h>
 #include <DerivativeFilter.h>