/*
 * Copyright (C) EdgeTX
 *
 * Based on code named
 *   opentx - https://github.com/opentx/opentx
 *   th9x - http://code.google.com/p/th9x
 *   er9x - http://code.google.com/p/er9x
 *   gruvin9x - http://code.google.com/p/gruvin9x
 *
 * License GPLv2: http://www.gnu.org/licenses/gpl-2.0.html
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 */

#include "opentx.h"
#include "timers.h"
#include "switches.h"

int8_t  virtualInputsTrims[MAX_INPUTS];
int16_t anas [MAX_INPUTS] = {0};
int16_t trims[NUM_TRIMS] = {0};
int32_t chans[MAX_OUTPUT_CHANNELS] = {0};
BeepANACenter bpanaCenter = 0;

int32_t act   [MAX_MIXERS] = {0};
SwOn    swOn  [MAX_MIXERS]; // TODO better name later...

uint8_t mixWarning;


int16_t calibratedAnalogs[NUM_CALIBRATED_ANALOGS];
int16_t channelOutputs[MAX_OUTPUT_CHANNELS] = {0};
int16_t ex_chans[MAX_OUTPUT_CHANNELS] = {0}; // Outputs (before LIMITS) of the last perMain;

#if defined(HELI)
int16_t cyc_anas[3] = {0};
#endif

// #define EXTENDED_EXPO
// increases range of expo curve but costs about 82 bytes flash

// expo-funktion:
// ---------------
// kmplot
// f(x,k)=exp(ln(x)*k/10) ;P[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20]
// f(x,k)=x*x*x*k/10 + x*(1-k/10) ;P[0,1,2,3,4,5,6,7,8,9,10]
// f(x,k)=x*x*k/10 + x*(1-k/10) ;P[0,1,2,3,4,5,6,7,8,9,10]
// f(x,k)=1+(x-1)*(x-1)*(x-1)*k/10 + (x-1)*(1-k/10) ;P[0,1,2,3,4,5,6,7,8,9,10]
// don't know what this above should be, just confusing in my opinion,

// here is the real explanation
// actually the real formula is
/*
 f(x) = exp( ln(x) * 10^k)
 if it is 10^k or e^k or 2^k etc. just defines the max distortion of the expo curve; I think 10 is useful
 this gives values from 0 to 1 for x and output; k must be between -1 and +1
 we do not like to calculate with floating point. Therefore we rescale for x from 0 to 1024 and for k from -100 to +100
 f(x) = 1024 * ( e^( ln(x/1024) * 10^(k/100) ) )
 This would be really hard to be calculated by such a microcontroller
 Therefore Thomas Husterer compared a few usual function something like x^3, x^4*something, which look similar
 Actually the formula
 f(x) = k*x^3+x*(1-k)
 gives a similar form and should have even advantages compared to a original exp curve.
 This function again expect x from 0 to 1 and k only from 0 to 1
 Therefore rescaling is needed like before:
 f(x) = 1024* ((k/100)*(x/1024)^3 + (x/1024)*(100-k)/100)
 some mathematical tricks
 f(x) = (k*x*x*x/(1024*1024) + x*(100-k)) / 100
 for better rounding results we add the 50
 f(x) = (k*x*x*x/(1024*1024) + x*(100-k) + 50) / 100

 because we now understand the formula, we can optimize it further
 --> calc100to256(k) --> eliminates /100 by replacing with /256 which is just a simple shift right 8
 k is now between 0 and 256
 f(x) = (k*x*x*x/(1024*1024) + x*(256-k) + 128) / 256
 */

// input parameters;
//  x 0 to 1024;
//  k 0 to 100;
// output between 0 and 1024
unsigned int expou(unsigned int x, unsigned int k)
{
#if defined(EXTENDED_EXPO)
  bool extended;
  if (k > 80) {
    extended=true;
  }
  else {
    k += (k>>2);  // use bigger values before extend, because the effect is anyway very very low
    extended=false;
  }
#endif

  k = calc100to256(k);

  uint32_t value = (uint32_t) x*x;
  value *= (uint32_t)k;
  value >>= 8;
  value *= (uint32_t)x;

#if defined(EXTENDED_EXPO)
  if (extended) {  // for higher values do more multiplications to get a stronger expo curve
    value >>= 16;
    value *= (uint32_t)x;
    value >>= 4;
    value *= (uint32_t)x;
  }
#endif

  value >>= 12;
  value += (uint32_t)(256-k) * x + 128;

  return value >> 8;
}

int expo(int x, int k)
{
  if (k == 0) {
    return x;
  }

  int y;
  bool neg = (x < 0);

  if (neg) {
    x = -x;
  }
  if (x > (int)RESXu) {
    x = RESXu;
  }
  if (k < 0) {
    y = RESXu - expou(RESXu-x, -k);
  }
  else {
    y = expou(x, k);
  }
  return neg ? -y : y;
}

void applyExpos(int16_t * anas, uint8_t mode, uint8_t ovwrIdx, int16_t ovwrValue)
{
  int8_t cur_chn = -1;

  for (uint8_t i=0; i<MAX_EXPOS; i++) {
    if (mode == e_perout_mode_normal) swOn[i].activeExpo = false;
    ExpoData * ed = expoAddress(i);
    if (!EXPO_VALID(ed)) break; // end of list
    if (ed->chn == cur_chn)
      continue;
    if (ed->flightModes & (1<<mixerCurrentFlightMode))
      continue;
    if (ed->srcRaw >= MIXSRC_FIRST_TRAINER && ed->srcRaw <= MIXSRC_LAST_TRAINER && !IS_TRAINER_INPUT_VALID())
      continue;
    if (getSwitch(ed->swtch)) {
      int32_t v;
      if (ed->srcRaw == ovwrIdx) {
        v = ovwrValue;
      }
      else {
        v = getValue(ed->srcRaw);
        if (ed->srcRaw >= MIXSRC_FIRST_TELEM && ed->scale > 0) {
          v = (v * 1024) / convertTelemValue(ed->srcRaw-MIXSRC_FIRST_TELEM+1, ed->scale);
        }
        v = limit<int32_t>(-1024, v, 1024);
      }
      if (EXPO_MODE_ENABLE(ed, v)) {
        if (mode == e_perout_mode_normal) swOn[i].activeExpo = true;
        cur_chn = ed->chn;

        //========== CURVE=================
        if (ed->curve.value) {
          v = applyCurve(v, ed->curve);
        }

        //========== WEIGHT ===============
        int32_t weight = GET_GVAR_PREC1(ed->weight, -100, 100, mixerCurrentFlightMode);
        v = divRoundClosest((int32_t)v * weight, 1000);

        //========== OFFSET ===============
        int32_t offset = GET_GVAR_PREC1(ed->offset, -100, 100, mixerCurrentFlightMode);
        if (offset) v += divRoundClosest(calc100toRESX(offset), 10);

        //========== TRIMS ================
        if (ed->carryTrim < TRIM_ON)
          virtualInputsTrims[cur_chn] = -ed->carryTrim - 1;
        else if (ed->carryTrim == TRIM_ON && ed->srcRaw >= MIXSRC_Rud && ed->srcRaw <= MIXSRC_Ail)
          virtualInputsTrims[cur_chn] = ed->srcRaw - MIXSRC_Rud;
        else
          virtualInputsTrims[cur_chn] = -1;
        anas[cur_chn] = v;
      }
    }
  }
}

// #define PREVENT_ARITHMETIC_OVERFLOW
// because of optimizations the reserves before overruns occurs is only the half
// this defines enables some checks the greatly improves this situation
// It should nearly prevent all overruns (is still a chance for it, but quite low)
// negative side is code cost 96 bytes flash

// we do it now half way, only in applyLimits, which costs currently 50bytes
// according opinion poll this topic is currently not very important
// the change below improves already the situation
// the check inside mixer would slow down mix a little bit and costs additionally flash
// also the check inside mixer still is not bulletproof, there may be still situations a overflow could occur
// a bulletproof implementation would take about additional 100bytes flash
// therefore with go with this compromize, interested people could activate this define

// @@@2 open.20.fsguruh ;
// channel = channelnumber -1;
// value = outputvalue with 100 mulitplied usual range -102400 to 102400; output -1024 to 1024
// changed rescaling from *100 to *256 to optimize performance
// rescaled from -262144 to 262144
int16_t applyLimits(uint8_t channel, int32_t value)
{
#if defined(OVERRIDE_CHANNEL_FUNCTION)
  if (safetyCh[channel] != OVERRIDE_CHANNEL_UNDEFINED) {
    // safety channel available for channel check
    return calc100toRESX(safetyCh[channel]);
  }
#endif

  if (isFunctionActive(FUNCTION_TRAINER_CHANNELS) && IS_TRAINER_INPUT_VALID()) {
    return ppmInput[channel] * 2;
  }

  LimitData * lim = limitAddress(channel);

  if (lim->curve) {
    // TODO we loose precision here, applyCustomCurve could work with int32_t on ARM boards...
    if (lim->curve > 0)
      value = 256 * applyCustomCurve(value/256, lim->curve-1);
    else
      value = 256 * applyCustomCurve(-value/256, -lim->curve-1);
  }

  int16_t ofs   = LIMIT_OFS_RESX(lim);
  int16_t lim_p = LIMIT_MAX_RESX(lim);
  int16_t lim_n = LIMIT_MIN_RESX(lim);

  if (ofs > lim_p) ofs = lim_p;
  if (ofs < lim_n) ofs = lim_n;

  // because the rescaling optimization would reduce the calculation reserve we activate this for all builds
  // it increases the calculation reserve from factor 20,25x to 32x, which it slightly better as original
  // without it we would only have 16x which is slightly worse as original, we should not do this

  // thanks to gbirkus, he motivated this change, which greatly reduces overruns
  // unfortunately the constants and 32bit compares generates about 50 bytes codes; didn't find a way to get it down.
  value = limit(int32_t(-RESXl*256), value, int32_t(RESXl*256));  // saves 2 bytes compared to other solutions up to now

#if defined(PPM_LIMITS_SYMETRICAL)
  if (value) {
    int16_t tmp;
    if (lim->symetrical)
      tmp = (value > 0) ? (lim_p) : (-lim_n);
    else
      tmp = (value > 0) ? (lim_p - ofs) : (-lim_n + ofs);
    value = (int32_t) value * tmp;   //  div by 1024*256 -> output = -1024..1024
#else
  if (value) {
    int16_t tmp = (value > 0) ? (lim_p - ofs) : (-lim_n + ofs);
    value = (int32_t) value * tmp;   //  div by 1024*256 -> output = -1024..1024
#endif

#ifdef CORRECT_NEGATIVE_SHIFTS
    int8_t sign = (value<0?1:0);
    value -= sign;
    tmp = value>>16;   // that's quite tricky: the shiftright 16 operation is assmbled just with addressmove; just forget the two least significant bytes;
    tmp >>= 2;   // now one simple shift right for two bytes does the rest
    tmp += sign;
#else
    tmp = value>>16;   // that's quite tricky: the shiftright 16 operation is assmbled just with addressmove; just forget the two least significant bytes;
    tmp >>= 2;   // now one simple shift right for two bytes does the rest
#endif

    ofs += tmp;  // ofs can to added directly because already recalculated,
  }

  if (ofs > lim_p)
    ofs = lim_p;
  if (ofs < lim_n)
    ofs = lim_n;
  if (lim->revert)
    ofs = -ofs; // finally do the reverse.

  return ofs;
}

// TODO same naming convention than the drawSource

getvalue_t getValue(mixsrc_t i)
{
  if (i == MIXSRC_NONE) {
    return 0;
  }
  else if (i <= MIXSRC_LAST_INPUT) {
    return anas[i-MIXSRC_FIRST_INPUT];
  }
#if defined(LUA_INPUTS)
  else if (i <= MIXSRC_LAST_LUA) {
#if defined(LUA_MODEL_SCRIPTS)
    div_t qr = div(i-MIXSRC_FIRST_LUA, MAX_SCRIPT_OUTPUTS);
    return scriptInputsOutputs[qr.quot].outputs[qr.rem].value;
#else
    return 0;
#endif
  }
#endif

  else if (i <= MIXSRC_LAST_POT + NUM_MOUSE_ANALOGS) {
    return calibratedAnalogs[i - MIXSRC_Rud];
  }

#if defined(IMU)
  else if (i == MIXSRC_TILT_X) {
    return gyro.scaledX();
  }
  else if (i == MIXSRC_TILT_Y) {
    return gyro.scaledY();
  }
#endif

  else if (i == MIXSRC_MAX) {
    return 1024;
  }

  else if (i <= MIXSRC_CYC3) {
#if defined(HELI)
    return cyc_anas[i - MIXSRC_CYC1];
#else
    return 0;
#endif
  }

  else if (i <= MIXSRC_LAST_TRIM) {
    return calc1000toRESX((int16_t)8 * getTrimValue(mixerCurrentFlightMode, i-MIXSRC_FIRST_TRIM));
  }

  // TODO : find a better define
#if defined(PCBFRSKY) || defined(PCBFLYSKY)
#if defined(FUNCTION_SWITCHES)
  else if (i >= MIXSRC_FIRST_SWITCH && i <= MIXSRC_LAST_REGULAR_SWITCH) {
    mixsrc_t sw = i - MIXSRC_FIRST_SWITCH;
    if (SWITCH_EXISTS(sw)) {
      return (switchState(3*sw) ? -1024 : (IS_CONFIG_3POS(sw) && switchState(3*sw+1) ? 0 : 1024));
    }
    else {
      return 0;
    }
  }
  else if (i >= MIXSRC_FIRST_FS_SWITCH && i <= MIXSRC_LAST_SWITCH) {
    return getFSLogicalState(i - MIXSRC_FIRST_SWITCH - NUM_REGULAR_SWITCHES) ? +1024 : -1024;
  }
#else
  else if (i >= MIXSRC_FIRST_SWITCH && i <= MIXSRC_LAST_SWITCH) {
    mixsrc_t sw = i - MIXSRC_FIRST_SWITCH;
    if (SWITCH_EXISTS(sw)) {
      return (switchState(3*sw) ? -1024 : (IS_CONFIG_3POS(sw) && switchState(3*sw+1) ? 0 : 1024));
    }
    else {
      return 0;
    }
  }
#endif
#else
  else if (i == MIXSRC_3POS) {
    return (getSwitch(SW_ID0+1) ? -1024 : (getSwitch(SW_ID1+1) ? 0 : 1024));
  }
  // don't use switchState directly to give getSwitch possibility to hack values if needed for switch warning
  else if (i < MIXSRC_SW1) {
    return getSwitch(SWSRC_THR+i-MIXSRC_THR) ? 1024 : -1024;
  }
#endif

  else if (i <= MIXSRC_LAST_LOGICAL_SWITCH) {
    return getSwitch(SWSRC_FIRST_LOGICAL_SWITCH + i - MIXSRC_FIRST_LOGICAL_SWITCH) ? 1024 : -1024;
  }
  else if (i <= MIXSRC_LAST_TRAINER) {
    int16_t x = ppmInput[i - MIXSRC_FIRST_TRAINER];
    if (i < MIXSRC_FIRST_TRAINER + NUM_CAL_PPM) {
      x -= g_eeGeneral.trainer.calib[i - MIXSRC_FIRST_TRAINER];
    }
    return x * 2;
  }
  else if (i <= MIXSRC_LAST_CH) {
    return ex_chans[i - MIXSRC_CH1];
  }

  else if (i <= MIXSRC_LAST_GVAR) {
#if defined(GVARS)
    return GVAR_VALUE(i - MIXSRC_GVAR1, getGVarFlightMode(mixerCurrentFlightMode, i - MIXSRC_GVAR1));
#else
    return 0;
#endif
  }

  else if (i == MIXSRC_TX_VOLTAGE) {
    return g_vbat100mV;
  }
  else if (i < MIXSRC_FIRST_TIMER) {
    // TX_TIME + SPARES
#if defined(RTCLOCK)
    return (g_rtcTime % SECS_PER_DAY) / 60; // number of minutes from midnight
#else
    return 0;
#endif
  }
  else if (i <= MIXSRC_LAST_TIMER) {
    return timersStates[i - MIXSRC_FIRST_TIMER].val;
  }

  else if (i <= MIXSRC_LAST_TELEM) {
    if (IS_FAI_FORBIDDEN(i)) {
      return 0;
    }
    i -= MIXSRC_FIRST_TELEM;
    div_t qr = div(i, 3);
    TelemetryItem & telemetryItem = telemetryItems[qr.quot];
    switch (qr.rem) {
      case 1:
        return telemetryItem.valueMin;
      case 2:
        return telemetryItem.valueMax;
      default:
        return telemetryItem.value;
    }
  }
  else return 0;
}

void evalInputs(uint8_t mode)
{
  BeepANACenter anaCenter = 0;

#if defined(STICK_DEAD_ZONE)
  static int16_t P_OFFSET = 0;
  static int16_t N_OFFSET = 0;
  static float aParam = 0.0f;
  static float bParam = 0.0f;
  static int16_t lastDeadZone = -1;

  if (lastDeadZone != g_eeGeneral.stickDeadZone) {
    P_OFFSET =
        (g_eeGeneral.stickDeadZone ? 2 << (g_eeGeneral.stickDeadZone - 1) : 0);
    N_OFFSET = (-1) * P_OFFSET;
    aParam = 1024.0 / (1024.0 - (float)P_OFFSET);
    bParam = 1024.0 * (aParam - 1.0f);
    lastDeadZone = g_eeGeneral.stickDeadZone;
  }
#endif

  for (uint8_t i = 0; i < NUM_STICKS + NUM_POTS + NUM_SLIDERS; i++) {
    // normalization [0..2048] -> [-1024..1024]
    uint8_t ch = (i < NUM_STICKS ? CONVERT_MODE(i) : i);
    int16_t v = anaIn(i);

    if (IS_POT_MULTIPOS(i)) {
      v -= RESX;
    }
#if !defined(SIMU)
    else {
      CalibData * calib = &g_eeGeneral.calib[i];
      v -= calib->mid;
      v = v * (int32_t) RESX / (max((int16_t) 100, (v > 0 ? calib->spanPos : calib->spanNeg)));
    }
#endif

    if (v < -RESX) v = -RESX;
    if (v >  RESX) v =  RESX;

#if defined(STICK_DEAD_ZONE)
    // dead zone invented by FlySky in my opinion it should goes into ADC
    // float calculations are not efficient
    if (g_eeGeneral.stickDeadZone && ch != THR_STICK) {
      if (v > P_OFFSET) {
        // y=ax+b
        v = (int)((aParam * (float)v) - bParam);
      } else if ((v <= P_OFFSET) && (v >= N_OFFSET)) {
        v = 0;
      } else if (v < N_OFFSET) {
        // y=ax+b
        v = (int)((aParam * (float)v) + bParam);
      }
    }
#endif

    if (g_model.throttleReversed && ch==THR_STICK) {
      v = -v;
    }

    BeepANACenter mask = (BeepANACenter)1 << ch;

    calibratedAnalogs[ch] = v; // for show in expo

    // filtering for center beep
    uint8_t tmp = (uint16_t)abs(v) / 16;
    if (mode == e_perout_mode_normal) {
      if (tmp==0 || (tmp==1 && (bpanaCenter & mask))) {
        anaCenter |= mask;
        if ((g_model.beepANACenter & mask) && !(bpanaCenter & mask) && s_mixer_first_run_done && !menuCalibrationState) {
          if (!IS_POT(i) || IS_POT_SLIDER_AVAILABLE(i)) {
            AUDIO_POT_MIDDLE(i);
          }
        }
      }
    }

    if (ch < NUM_STICKS) { // only do this for sticks
      if (mode & e_perout_mode_nosticks) {
        v = 0;
      }

      if (mode <= e_perout_mode_inactive_flight_mode && isFunctionActive(FUNCTION_TRAINER_STICK1+ch) && IS_TRAINER_INPUT_VALID()) {
        // trainer mode
        TrainerMix* td = &g_eeGeneral.trainer.mix[ch];
        if (td->mode) {
          uint8_t chStud = td->srcChn;
          int32_t vStud  = (ppmInput[chStud] - g_eeGeneral.trainer.calib[chStud]);
          vStud *= td->studWeight;
          vStud /= 50;
          switch (td->mode) {
            case TRAINER_ADD:
              // add-mode
              v = limit<int16_t>(-RESX, v+vStud, RESX);
              break;
            case TRAINER_REPL:
              // subst-mode
              v = vStud;
              break;
          }
        }
      }
      calibratedAnalogs[ch] = v;
    }
  }

#if NUM_MOUSE_ANALOGS > 0
  for (uint8_t i=0; i<NUM_MOUSE_ANALOGS; i++) {
    uint8_t ch = NUM_STICKS+NUM_POTS+NUM_SLIDERS+i;
    int16_t v = anaIn(MOUSE1+i);
    CalibData * calib = &g_eeGeneral.calib[ch];
    v -= calib->mid;
    v = v * (int32_t) RESX / (max((int16_t) 100, (v > 0 ? calib->spanPos : calib->spanNeg)));
    if (v < -RESX) v = -RESX;
    if (v >  RESX) v =  RESX;
    calibratedAnalogs[ch] = v;
  }
#endif

  /* EXPOs */
  applyExpos(anas, mode);

  /* TRIMs */
  evalTrims(); // when no virtual inputs, the trims need the anas array calculated above (when throttle trim enabled)

  if (mode == e_perout_mode_normal) {
    bpanaCenter = anaCenter;
  }
}

int getStickTrimValue(int stick, int stickValue)
{
  if (stick < 0)
    return 0;

  int trim = trims[stick];
  uint8_t thrTrimSw = g_model.getThrottleStickTrimSource() - MIXSRC_FIRST_TRIM;
  if (stick == thrTrimSw) {
    if (g_model.throttleReversed)
      trim = -trim;
    if (g_model.thrTrim) {
      trim = (g_model.extendedTrims) ? 2*TRIM_EXTENDED_MAX + trim : 2*TRIM_MAX + trim;
      trim = trim * (1024 - stickValue) / (2*RESX);
    }
  }
  return trim;
}

int getSourceTrimOrigin(int source)
{
  if (source >= MIXSRC_Rud && source <= MIXSRC_Ail)
    return source - MIXSRC_Rud;
  else if (source >= MIXSRC_FIRST_INPUT && source <= MIXSRC_LAST_INPUT)
    return virtualInputsTrims[source - MIXSRC_FIRST_INPUT];
  else
    return -1;
}

int getSourceTrimValue(int source, int stickValue=0)
{
  auto origin = getSourceTrimOrigin(source);
  if (origin >= 0) {
    return getStickTrimValue(origin, stickValue);
  }
  else {
    return 0;
  }
}

uint8_t mixerCurrentFlightMode;
void evalFlightModeMixes(uint8_t mode, uint8_t tick10ms)
{
  evalInputs(mode);

  if (tick10ms)
    evalLogicalSwitches(mode==e_perout_mode_normal);

#if defined(HELI)
  int heliEleValue = getValue(g_model.swashR.elevatorSource);
  int heliAilValue = getValue(g_model.swashR.aileronSource);
  if (g_model.swashR.value) {
    uint32_t v = ((int32_t)heliEleValue*heliEleValue + (int32_t)heliAilValue*heliAilValue);
    uint32_t q = calc100toRESX(g_model.swashR.value);
    q *= q;
    if (v>q) {
      uint16_t d = isqrt32(v);
      int16_t tmp = calc100toRESX(g_model.swashR.value);
      heliEleValue = (int32_t) heliEleValue*tmp/d;
      heliAilValue = (int32_t) heliAilValue*tmp/d;
    }
  }

#define REZ_SWASH_X(x)  ((x) - (x)/8 - (x)/128 - (x)/512)   //  1024*sin(60) ~= 886
#define REZ_SWASH_Y(x)  ((x))   //  1024 => 1024

  if (g_model.swashR.type) {
    getvalue_t vp = heliEleValue + getSourceTrimValue(g_model.swashR.elevatorSource);
    getvalue_t vr = heliAilValue + getSourceTrimValue(g_model.swashR.aileronSource);
    getvalue_t vc = 0;
    if (g_model.swashR.collectiveSource)
      vc = getValue(g_model.swashR.collectiveSource);

    vp = (vp * g_model.swashR.elevatorWeight) / 100;
    vr = (vr * g_model.swashR.aileronWeight) / 100;
    vc = (vc * g_model.swashR.collectiveWeight) / 100;

    switch (g_model.swashR.type) {
      case SWASH_TYPE_120:
        vp = REZ_SWASH_Y(vp);
        vr = REZ_SWASH_X(vr);
        cyc_anas[0] = vc - vp;
        cyc_anas[1] = vc + vp/2 + vr;
        cyc_anas[2] = vc + vp/2 - vr;
        break;
      case SWASH_TYPE_120X:
        vp = REZ_SWASH_X(vp);
        vr = REZ_SWASH_Y(vr);
        cyc_anas[0] = vc - vr;
        cyc_anas[1] = vc + vr/2 + vp;
        cyc_anas[2] = vc + vr/2 - vp;
        break;
      case SWASH_TYPE_140:
        vp = REZ_SWASH_Y(vp);
        vr = REZ_SWASH_Y(vr);
        cyc_anas[0] = vc - vp;
        cyc_anas[1] = vc + vp + vr;
        cyc_anas[2] = vc + vp - vr;
        break;
      case SWASH_TYPE_90:
        vp = REZ_SWASH_Y(vp);
        vr = REZ_SWASH_Y(vr);
        cyc_anas[0] = vc - vp;
        cyc_anas[1] = vc + vr;
        cyc_anas[2] = vc - vr;
        break;
      default:
        break;
    }
  }
#endif

  memclear(chans, sizeof(chans)); // all outputs to 0

  //========== MIXER LOOP ===============
  uint8_t lv_mixWarning = 0;

  uint8_t pass = 0;

  bitfield_channels_t dirtyChannels = (bitfield_channels_t)-1; // all dirty when mixer starts

  do {
    bitfield_channels_t passDirtyChannels = 0;

    for (uint8_t i=0; i<MAX_MIXERS; i++) {
      if (mode == e_perout_mode_normal && pass == 0)
        swOn[i].activeMix = 0;

      MixData * md = mixAddress(i);

      if (md->srcRaw == 0)
#if defined(COLORLCD)
        continue;
#else
        break;
#endif

      mixsrc_t stickIndex = md->srcRaw - MIXSRC_Rud;

      if (!(dirtyChannels & ((bitfield_channels_t)1 << md->destCh)))
        continue;

      // if this is the first calculation for the destination channel, initialize it with 0 (otherwise would be random)
      if (i == 0 || md->destCh != (md-1)->destCh)
        chans[md->destCh] = 0;

      //========== FLIGHT MODE && SWITCH =====
      bool mixCondition = (md->flightModes != 0 || md->swtch);
      delayval_t mixEnabled = (!(md->flightModes & (1 << mixerCurrentFlightMode)) && getSwitch(md->swtch)) ? DELAY_POS_MARGIN+1 : 0;

#define MIXER_LINE_DISABLE()   (mixCondition = true, mixEnabled = 0)

      if (mixEnabled && md->srcRaw >= MIXSRC_FIRST_TRAINER && md->srcRaw <= MIXSRC_LAST_TRAINER && !IS_TRAINER_INPUT_VALID()) {
        MIXER_LINE_DISABLE();
      }

#if defined(LUA_MODEL_SCRIPTS)
      // disable mixer if Lua script is used as source and script was killed
      if (mixEnabled && md->srcRaw >= MIXSRC_FIRST_LUA && md->srcRaw <= MIXSRC_LAST_LUA) {
        div_t qr = div(md->srcRaw-MIXSRC_FIRST_LUA, MAX_SCRIPT_OUTPUTS);
        if (scriptInternalData[qr.quot].state != SCRIPT_OK) {
          MIXER_LINE_DISABLE();
        }
      }
#endif

      //========== VALUE ===============
      getvalue_t v = 0;
      if (mode > e_perout_mode_inactive_flight_mode) {
        if (mixEnabled)
          v = getValue(md->srcRaw);
        else
          continue;
      }
      else {
        mixsrc_t srcRaw = MIXSRC_Rud + stickIndex;
        v = getValue(srcRaw);
        srcRaw -= MIXSRC_CH1;
        if (srcRaw <= MIXSRC_LAST_CH-MIXSRC_CH1 && md->destCh != srcRaw) {
          if (dirtyChannels & ((bitfield_channels_t)1 << srcRaw) & (passDirtyChannels|~(((bitfield_channels_t) 1 << md->destCh)-1)))
            passDirtyChannels |= (bitfield_channels_t) 1 << md->destCh;
          if (srcRaw < md->destCh || pass > 0)
            v = chans[srcRaw] >> 8;
        }
        if (!mixCondition) {
          mixEnabled = v;
        }
      }

      bool applyOffsetAndCurve = true;

      //========== DELAYS ===============
      delayval_t _swOn = swOn[i].now;
      delayval_t _swPrev = swOn[i].prev;
      bool swTog = (mixEnabled > _swOn+DELAY_POS_MARGIN || mixEnabled < _swOn-DELAY_POS_MARGIN);
      if (mode == e_perout_mode_normal && swTog) {
        if (!swOn[i].delay)
          _swPrev = _swOn;
        swOn[i].delay = (mixEnabled > _swOn ? md->delayUp : md->delayDown) * 10;
        swOn[i].now = mixEnabled;
        swOn[i].prev = _swPrev;
      }
      if (mode == e_perout_mode_normal && swOn[i].delay > 0) {
        swOn[i].delay = max<int16_t>(0, (int16_t)swOn[i].delay - tick10ms);
        if (!mixCondition)
          v = _swPrev;
        else if (mixEnabled)
          continue;
      }
      else {
        if (mode==e_perout_mode_normal) {
          swOn[i].now = swOn[i].prev = mixEnabled;
        }
        if (!mixEnabled) {
          if ((md->speedDown || md->speedUp) && md->mltpx!=MLTPX_REPL) {
            if (mixCondition) {
              v = (md->mltpx == MLTPX_ADD ? 0 : RESX);
              applyOffsetAndCurve = false;
            }
          }
          else if (mixCondition) {
            continue;
          }
        }
      }

      if (mode==e_perout_mode_normal && (!mixCondition || mixEnabled || swOn[i].delay)) {
        if (md->mixWarn)
          lv_mixWarning |= 1 << (md->mixWarn - 1);
        swOn[i].activeMix = true;
      }

      if (applyOffsetAndCurve) {
        bool applyTrims = !(mode & e_perout_mode_notrims);
        if (!applyTrims && g_model.thrTrim) {
          auto origin = getSourceTrimOrigin(md->srcRaw);
          if (origin == g_model.getThrottleStickTrimSource() - MIXSRC_FIRST_TRIM) {
            applyTrims = true;
          }
        }
        if (applyTrims && md->carryTrim == 0) {
          v += getSourceTrimValue(md->srcRaw, v);
        }
      }

      int32_t weight = GET_GVAR_PREC1(MD_WEIGHT(md), GV_RANGELARGE_NEG, GV_RANGELARGE, mixerCurrentFlightMode);
      weight = calc100to256_16Bits(weight);
      //========== SPEED ===============
      // now its on input side, but without weight compensation. More like other remote controls
      // lower weight causes slower movement

      if (mode <= e_perout_mode_inactive_flight_mode && (md->speedUp || md->speedDown)) { // there are delay values
#define DEL_MULT_SHIFT 8
        // we recale to a mult 256 higher value for calculation
        int32_t tact = act[i];
        int16_t diff = v - (tact>>DEL_MULT_SHIFT);
        if (diff) {
          // open.20.fsguruh: speed is defined in % movement per second; In menu we specify the full movement (-100% to 100%) = 200% in total
          // the unit of the stored value is the value from md->speedUp or md->speedDown * 0.1s; e.g. value 4 means 0.4 seconds
          // because we get a tick each 10msec, we need 100 ticks for one second
          // the value in md->speedXXX gives the time it should take to do a full movement from -100 to 100 therefore 200%. This equals 2048 in recalculated internal range
          if (tick10ms || !s_mixer_first_run_done) {
            // only if already time is passed add or substract a value according the speed configured
            int32_t rate = (int32_t) tick10ms << (DEL_MULT_SHIFT+11);  // = DEL_MULT*2048*tick10ms
            // rate equals a full range for one second; if less time is passed rate is accordingly smaller
            // if one second passed, rate would be 2048 (full motion)*256(recalculated weight)*100(100 ticks needed for one second)
            int32_t currentValue = ((int32_t) v<<DEL_MULT_SHIFT);
            if (diff > 0) {
              if (s_mixer_first_run_done && md->speedUp > 0) {
                // if a speed upwards is defined recalculate the new value according configured speed; the higher the speed the smaller the add value is
                int32_t newValue = tact+rate/((int16_t)10*md->speedUp);
                if (newValue<currentValue) currentValue = newValue; // Endposition; prevent toggling around the destination
              }
            }
            else {  // if is <0 because ==0 is not possible
              if (s_mixer_first_run_done && md->speedDown > 0) {
                // see explanation in speedUp
                int32_t newValue = tact-rate/((int16_t)10*md->speedDown);
                if (newValue>currentValue) currentValue = newValue; // Endposition; prevent toggling around the destination
              }
            }
            act[i] = tact = currentValue;
            // open.20.fsguruh: this implementation would save about 50 bytes code
          } // endif tick10ms ; in case no time passed assign the old value, not the current value from source
          v = (tact >> DEL_MULT_SHIFT);
        }
      }

      //========== CURVES ===============
      if (applyOffsetAndCurve && md->curve.type != CURVE_REF_DIFF && md->curve.value) {
        v = applyCurve(v, md->curve);
      }

      //========== WEIGHT ===============
      int32_t dv = (int32_t)v * weight;
      dv = divRoundClosest(dv, 10);

      //========== OFFSET / AFTER ===============
      if (applyOffsetAndCurve) {
        int32_t offset = GET_GVAR_PREC1(MD_OFFSET(md), GV_RANGELARGE_NEG, GV_RANGELARGE, mixerCurrentFlightMode);
        if (offset) dv += divRoundClosest(calc100toRESX_16Bits(offset), 10) << 8;
      }

      //========== DIFFERENTIAL =========
      if (md->curve.type == CURVE_REF_DIFF && md->curve.value) {
        dv = applyCurve(dv, md->curve);
      }

      int32_t * ptr = &chans[md->destCh]; // Save calculating address several times

      switch (md->mltpx) {
        case MLTPX_REPL:
          *ptr = dv;
          if (mode == e_perout_mode_normal) {
            for (uint8_t m=i-1; m<MAX_MIXERS && mixAddress(m)->destCh==md->destCh; m--)
              swOn[m].activeMix = false;
          }
          break;
        case MLTPX_MUL:
          // @@@2 we have to remove the weight factor of 256 in case of 100%; now we use the new base of 256
          dv >>= 8;
          dv *= *ptr;
          dv >>= RESX_SHIFT;   // same as dv /= RESXl;
          *ptr = dv;
          break;
        default: // MLTPX_ADD
          *ptr += dv; //Mixer output add up to the line (dv + (dv>0 ? 100/2 : -100/2))/(100);
          break;
      } // endswitch md->mltpx
#ifdef PREVENT_ARITHMETIC_OVERFLOW
/*
      // a lot of assumptions must be true, for this kind of check; not really worth for only 4 bytes flash savings
      // this solution would save again 4 bytes flash
      int8_t testVar=(*ptr<<1)>>24;
      if ( (testVar!=-1) && (testVar!=0 ) ) {
        // this devices by 64 which should give a good balance between still over 100% but lower then 32x100%; should be OK
        *ptr >>= 6;  // this is quite tricky, reduces the value a lot but should be still over 100% and reduces flash need
      } */


      PACK( union u_int16int32_t {
        struct {
          int16_t lo;
          int16_t hi;
        } words_t;
        int32_t dword;
      });

      u_int16int32_t tmp;
      tmp.dword=*ptr;

      if (tmp.dword<0) {
        if ((tmp.words_t.hi&0xFF80)!=0xFF80) tmp.words_t.hi=0xFF86; // set to min nearly
      }
      else {
        if ((tmp.words_t.hi|0x007F)!=0x007F) tmp.words_t.hi=0x0079; // set to max nearly
      }
      *ptr = tmp.dword;
      // this implementation saves 18bytes flash

/*      dv=*ptr>>8;
      if (dv>(32767-RESXl)) {
        *ptr=(32767-RESXl)<<8;
      } else if (dv<(-32767+RESXl)) {
        *ptr=(-32767+RESXl)<<8;
      }*/
      // *ptr=limit( int32_t(int32_t(-1)<<23), *ptr, int32_t(int32_t(1)<<23));  // limit code cost 72 bytes
      // *ptr=limit( int32_t((-32767+RESXl)<<8), *ptr, int32_t((32767-RESXl)<<8));  // limit code cost 80 bytes
#endif

    } //endfor mixers

    tick10ms = 0;
    dirtyChannels &= passDirtyChannels;

  } while (++pass < 5 && dirtyChannels);

  mixWarning = lv_mixWarning;
}



#define MAX_ACT 0xffff
uint8_t lastFlightMode = 255; // TODO reinit everything here when the model changes, no???

tmr10ms_t flightModeTransitionTime;
uint8_t   flightModeTransitionLast = 255;

void evalMixes(uint8_t tick10ms)
{
  int32_t sum_chans512[MAX_OUTPUT_CHANNELS];

  static uint16_t fp_act[MAX_FLIGHT_MODES] = {0};
  static uint16_t delta = 0;
  static uint16_t flightModesFade = 0;

  uint8_t fm = getFlightMode();

  if (lastFlightMode != fm) {
    flightModeTransitionTime = get_tmr10ms();

    if (lastFlightMode == 255) {
      fp_act[fm] = MAX_ACT;
    }
    else {
      uint8_t fadeTime = max(g_model.flightModeData[lastFlightMode].fadeOut, g_model.flightModeData[fm].fadeIn);
      uint16_t transitionMask = (0x01u << lastFlightMode) + (0x01u << fm);
      if (fadeTime) {
        flightModesFade |= transitionMask;
        delta = (MAX_ACT / 10) / fadeTime;
      }
      else {
        flightModesFade &= ~transitionMask;
        fp_act[lastFlightMode] = 0;
        fp_act[fm] = MAX_ACT;
      }
      logicalSwitchesCopyState(lastFlightMode, fm); // push last logical switches state from old to new flight mode
    }
    lastFlightMode = fm;
  }

  if (flightModeTransitionTime && get_tmr10ms() > flightModeTransitionTime+SWITCHES_DELAY()) {
    flightModeTransitionTime = 0;
    if (fm != flightModeTransitionLast) {
      if (flightModeTransitionLast != 255) {
        PLAY_PHASE_OFF(flightModeTransitionLast);
      }
      PLAY_PHASE_ON(fm);
      flightModeTransitionLast = fm;
    }
  }

  int32_t weight = 0;
  if (flightModesFade) {
    memclear(sum_chans512, sizeof(sum_chans512));
    for (uint8_t p=0; p<MAX_FLIGHT_MODES; p++) {
      if (flightModesFade & (0x01 << p)) {
        mixerCurrentFlightMode = p;
        evalFlightModeMixes(p==fm ? e_perout_mode_normal : e_perout_mode_inactive_flight_mode, p==fm ? tick10ms : 0);
        for (uint8_t i=0; i<MAX_OUTPUT_CHANNELS; i++)
          sum_chans512[i] += limit<int32_t>(-0x6fff, chans[i] >> 4, 0x6fff) * fp_act[p];
        weight += fp_act[p];
      }
    }
    assert(weight);
    mixerCurrentFlightMode = fm;
  }
  else {
    mixerCurrentFlightMode = fm;
    evalFlightModeMixes(e_perout_mode_normal, tick10ms);
  }

  //========== FUNCTIONS ===============
  // must be done after mixing because some functions use the inputs/channels values
  // must be done before limits because of the applyLimit function: it checks for safety switches which would be not initialized otherwise
  if (tick10ms) {
    requiredSpeakerVolume = g_eeGeneral.speakerVolume + VOLUME_LEVEL_DEF;
    requiredBacklightBright = g_eeGeneral.backlightBright;

    if (!g_model.noGlobalFunctions) {
      evalFunctions(g_eeGeneral.customFn, globalFunctionsContext);
    }
    evalFunctions(g_model.customFn, modelFunctionsContext);
  }

  //========== LIMITS ===============
  for (uint8_t i=0; i<MAX_OUTPUT_CHANNELS; i++) {
    // chans[i] holds data from mixer.   chans[i] = v*weight => 1024*256
    // later we multiply by the limit (up to 100) and then we need to normalize
    // at the end chans[i] = chans[i]/256 =>  -1024..1024
    // interpolate value with min/max so we get smooth motion from center to stop
    // this limits based on v original values and min=-1024, max=1024  RESX=1024
    int32_t q = (flightModesFade ? (sum_chans512[i] / weight) << 4 : chans[i]);

    ex_chans[i] = q / 256;

    int16_t value = applyLimits(i, q);  // applyLimits will remove the 256 100% basis

    channelOutputs[i] = value;  // copy consistent word to int-level
  }

  if (tick10ms && flightModesFade) {
    uint16_t tick_delta = delta * tick10ms;
    for (uint8_t p=0; p<MAX_FLIGHT_MODES; p++) {
      uint16_t flightModeMask = (0x01 << p);
      if (flightModesFade & flightModeMask) {
        if (p == fm) {
          if (MAX_ACT - fp_act[p] > tick_delta)
            fp_act[p] += tick_delta;
          else {
            fp_act[p] = MAX_ACT;
            flightModesFade -= flightModeMask;
          }
        }
        else {
          if (fp_act[p] > tick_delta)
            fp_act[p] -= tick_delta;
          else {
            fp_act[p] = 0;
            flightModesFade -= flightModeMask;
          }
        }
      }
    }
  }
}