The old mcnt packet processing logic would result in blocks being filled out of
order and was also not very reliable when processing packets that arrived
significantly later than the pipeline was currently processing.

These changes address these issues making it so that the pipeline can now handle
late packets and restart at the mcnt it is currently seeing at the network. This
also greatly reduces the amount of bad blocks seen by in RTBF mode because the
logic better handles placing packets and we don't advance mcnts unnecessarily
resulting in marking blocks as filled to early.