SurfaceFlinger里面一个算法类，不知道什么作用。该如何解决

SurfaceFlinger里面一个算法类，不知道什么作用。
class Tokenizer
{
public:
  Tokenizer();
  Tokenizer(const Tokenizer& other);
  ~Tokenizer();

  uint32_t acquire();
  status_t reserve(uint32_t token);
  status_t release(uint32_t token);
  bool isAcquired(uint32_t token) const;

  void dump() const;

  struct run_t {
  run_t() {};
  run_t(uint32_t f, uint32_t l) : first(f), length(l) {}
  uint32_t first;
  uint32_t length;
  };
private:
  ssize_t _indexOrderOf(uint32_t token, size_t* order=0) const;
  ssize_t _insertTokenAt(uint32_t token, size_t index);
  Vector<run_t> mRanges;
};

}; // namespace android


Tokenizer::Tokenizer()
{
}

Tokenizer::Tokenizer(const Tokenizer& other)
  : mRanges(other.mRanges)
{
}

Tokenizer::~Tokenizer()
{
}

uint32_t Tokenizer::acquire()
{
  if (!mRanges.size() || mRanges[0].first) {
  _insertTokenAt(0,0);
  return 0;
  }
   
  // just extend the first run
  const run_t& run = mRanges[0];
  uint32_t token = run.first + run.length;
  _insertTokenAt(token, 1);
  return token;
}

bool Tokenizer::isAcquired(uint32_t token) const
{
  return (_indexOrderOf(token) >= 0);
}

status_t Tokenizer::reserve(uint32_t token)
{
  size_t o;
  const ssize_t i = _indexOrderOf(token, &o);
  if (i >= 0) {
  return BAD_VALUE; // this token is already taken
  }
  ssize_t err = _insertTokenAt(token, o);
  return (err<0) ? err : status_t(NO_ERROR);
}

status_t Tokenizer::release(uint32_t token)
{
  const ssize_t i = _indexOrderOf(token);
  if (i >= 0) {
  const run_t& run = mRanges[i];
  if ((token >= run.first) && (token < run.first+run.length)) {
  // token in this range, we need to split
  run_t& run = mRanges.editItemAt(i);
  if ((token == run.first) || (token == run.first+run.length-1)) {
  if (token == run.first) {
  run.first += 1;
  }
  run.length -= 1;
  if (run.length == 0) {
  // XXX: should we systematically remove a run that's empty?
  mRanges.removeItemsAt(i);
  }
  } else {
  // split the run
  run_t new_run;
  new_run.first = token+1;
  new_run.length = run.first+run.length - new_run.first;
  run.length = token - run.first;
  mRanges.insertAt(new_run, i+1);
  }
  return NO_ERROR;
  }
  }
  return NAME_NOT_FOUND;
}

ssize_t Tokenizer::_indexOrderOf(uint32_t token, size_t* order) const
{
  // binary search
  ssize_t err = NAME_NOT_FOUND;
  ssize_t l = 0;
  ssize_t h = mRanges.size()-1;
  ssize_t mid;
  const run_t* a = mRanges.array();
  while (l <= h) {
  mid = l + (h - l)/2;
  const run_t* const curr = a + mid;
  int c = 0;
  if (token < curr->first) c = 1;