Question

Very simple code:

signed int **ifftResults = (signed int **)malloc(sizeof(signed int *) * recordsPerBuffer);

for (int i=0; i < recordsPerBuffer; i++)
{
    ifftResults[i] = (signed int*)malloc(sizeof(signed int) * fftLength);
}

Then later:

for (int i=0; i < recordsPerBuffer; i++)
{
   free(ifftResults[i]);
}
free(ifftResults);

When I comment these lines out – no memory leak. When they are present – memory leak. Hopefully I just need another pair of eyes, cause I cannot for the life of me see what’s wrong.

Answer 1

The code presented as I’m writing this doesn’t seem to be sufficient to answer the question of “why”.

However, since you’re using C++ you can make sure that there’s no memory leak by using std::vector.

Like so:

// Allocation.
std::vector< std::vector< int > >  fftResults( recordsPerBuffer, std::vector< int >( fftLength ) );

// Usage:
fftResults[y][x] = blah;

// Deallocation: automatic.

Another way to implement a matrix is like

std::vector< int >  fftResults( recordsPerBuffer*fftLength );

and then compute the index for given (x,y).

Cheers & hth.,