Convert pointer to two dimensional array

Variable-Length Arrays can help you here:

int (*p)[sp] = (void*) m;

Now you can access the elements via p[i][j].

See it live on Coliru

About the well-definedness of the cast with respect to strict aliasing rules: it is valid, because both m and p point to compatible types.

You cannot, as m can be dereferenced only once. But if you know the dimensions of the array, you can calculate the 1D index as

element[i][j]=m[i * width + j];

If m is indeed a pointer to a 2D matrix, you can declare the arguments this way:

void printMatrix(int rows, int cols, int m[rows][cols]) {
    ...
}

You will have to assume that it is a so-called "mangled" array. In which case you can create a new array as a VLA:

void printMatrix(int *m, int ze, int sp)
{
  int array[ze][sp];
  memcpy(array, m, sizeof(int[ze][sp]));
}

Or perhaps as a dynamically allocated 2D array:

void printMatrix(int *m, int ze, int sp)
{
  size_t size = sizeof(int[ze][sp]);
  int (*array)[sp] = malloc(size);
  memcpy(array, m, size);
  ...
}

Converting the int* to an array pointer type is however questionable practice, which will cause pointer aliasing problems. The best is of course if you can rewrite the function prototype to properly use a 2D array instead.