DWT to measure time between rising edge of two pulse waves

Official Hardware Mega Due
1

i have given two square waves as input to arduino due's digital pins and trying to find the propagation delay between them. for this purpose instead of using micros() i want nanoseconds measurement hence DWT_CYCCNT is used . but the when the delay between two is zero( both are ideal square waves ) then alse the delay given by this code is 770ns. i just want your opinions on whether the appraoch to use DWT is correct for this purpose and how to improvise .

#define REG_DWT_CTRL   (*(volatile unsigned int*)0xE0001000U)  // DWT Control Register
#define REG_DWT_CYCCNT (*(volatile unsigned int*)0xE0001004U)  // DWT Cycle Count Register
#define REG_DEMCR      (*(volatile unsigned int*)0xE000EDFCU)  // Debug Exception and Monitor Control Register

#define CYCCNTENA (1 << 0)  // DWT CTRL Cycle Count Enable Bit
#define TRCENA    (1 << 24) // DEMCR Trace Enable Bit

#define INPUT_PIN 4  // Buffer input
#define OUTPUT_PIN 5 // Buffer output

volatile uint32_t start_cycles = 0;
volatile uint32_t end_cycles = 0;
volatile bool measurement_ready = false;

void inputISR() {
  start_cycles = REG_DWT_CYCCNT;  // Capture cycles at input rising edge
  measurement_ready = false;      // Reset until output triggers
}

void outputISR() {
  end_cycles = REG_DWT_CYCCNT;    // Capture cycles at output rising edge
  measurement_ready = true;       // Signal measurement is ready
}

void setup() {
  // Enable trace functionality
  REG_DEMCR |= TRCENA;
  
  // Reset cycle counter to 0 (optional)
  REG_DWT_CYCCNT = 0;
  
  // Enable the cycle counter
  REG_DWT_CTRL |= CYCCNTENA;

  pinMode(INPUT_PIN, INPUT);
  pinMode(OUTPUT_PIN, INPUT);

  // Attach interrupts for rising edges (buffer: input rise -> output rise)
  attachInterrupt(digitalPinToInterrupt(INPUT_PIN), inputISR, RISING);
  attachInterrupt(digitalPinToInterrupt(OUTPUT_PIN), outputISR, RISING);

  Serial.begin(115200);
  
}

void loop() {
  if (measurement_ready) {
    uint32_t cycle_diff = end_cycles - start_cycles;
    float delay_ns = cycle_diff * (1000.0 / 84.0); // Convert to ns (84 MHz)
    Serial.print("Propagation delay (ns): ");
    Serial.println(delay_ns);
    measurement_ready = false; // Reset for next measurement
     
  }
}
2

What additional hardware are you planning on adding. I would use an oscilloscope for the measurement.