Classic Nano has too much RAM

[sterretje]

I was running a sketch where I needed to check the amount of available memory after certain operations/function calls. To my surprise, I got strange results.

The below (trimmed down) sketch reported more memory than the Nano actually has. The freeMemory() function was taken from Measuring Memory Usage | Memories of an Arduino | Adafruit Learning System

Sketch

void setup()
{
  Serial.begin(115200);
  Serial.println(F("String (capital S) memory test"));

  Serial.print(F("Free RAM = "));
  Serial.println(freeMemory());
}

void loop()
{
}

#ifdef __arm__
// should use uinstd.h to define sbrk but Due causes a conflict
extern "C" char* sbrk(int incr);
#else   // __ARM__
extern char *__brkval;
#endif  // __arm__

#ifdef __arm__
// should use uinstd.h to define sbrk but Due causes a conflict
extern "C" char* sbrk(int incr);
#else   // __ARM__
extern char *__brkval;
#endif  // __arm__

int freeMemory()
{
  char top;
#ifdef __arm__
  return &top - reinterpret_cast<char*>(sbrk(0));
#elif defined(CORE_TEENSY) || (ARDUINO > 103 && ARDUINO != 151)
  return &top - __brkval;
#else   // __arm__
  return __brkval ? &top - __brkval : &top - __malloc_heap_start;
#endif  // __arm__
}

Output

String (capital S) memory test
Free RAM = 2299

Anybody any idea?

[johnwasser]

All of the arm and ARM references make me think that this version of freeMemory() was intended for ARM processors, not the AVR processor on the classic Nano.

[pylon]

Remove these two lines and try again:

@sterretje

#elif defined(CORE_TEENSY) || (ARDUINO > 103 && ARDUINO != 151)
  return &top - __brkval;

[PaulRB]

I think @pylon may be on to something...

#elif defined(CORE_TEENSY) || (ARDUINO > 103 && ARDUINO != 151)

If @sterretje is using, for example, IDE 1.8, then the above will give the result

return &top - __brkval;

but what if __brkval is zero? That possibility is not checked, but in the other version

return __brkval ? &top - __brkval : &top - __malloc_heap_start;

it does check if __brkval is zero, and uses __malloc_heap_start instead in that case.

So maybe that line that is being used is incorrect, a bug, because it does not check for __brkval being zero.

If RAM begins at, say, address 256 then &top could be around 256+2048 = 2304 less a few bytes = 2299?

[sterretje]

Thanks people for the replies

@pylon

Remove these two lines and try again:

A more sensible 1845 bytes free

If I ever feel the need to dig deeper, ...