RTE Port for IPv4 Fragmentation
This port is built on top of pre-initialized single consumer rte_ring. In order to minimize the amount of
packets stored in the ring at any given time, the IP fragmentation functionality is executed on ring read
operation, hence this port is implemented as an input port. A regular ring_writer port can be created to
write to the same ring.
The packets written to the ring are either complete IP datagrams or jumbo frames (i.e. IP packets with
length bigger than provided MTU value). The packets read from the ring are all non-jumbo frames. The
complete IP datagrams written to the ring are not changed. The jumbo frames are fragmented into several
IP packets with length less or equal to MTU.
Definition in file rte_port_frag.h.