Improve PrrtBlock internal lists efficiency

Currently, PrrtBlock uses linked lists internally to store packets. Accesses are linear from the start. There should be a more efficient algorithm and data structure for this.

Run gprof to look for obvious performance issues.