The primary header in bundle protocol contains some offset values(destination
scheme offset, destination ssp offset, source scheme offset, source ssp offset,
etc). These are the offsets within the dictionary if the length of the
dictionary length is greater than 0. But if the dictionary length is 0, then
these offset refer to node number and service number respectively(according to
compressed bundle header encoding). For example if destination scheme offset is
2 and the destination ssp offset is 1, then the destination
EID(<node_number>.<service_number>) is 2.1.
Currently the dtn dissector will consider these offsets to be actual offsets in
the dictionary even if the dictionary length is 0. So the values for the
EID's(destination, source, report, custodian) and their schemes are junk
values. For example if the destination scheme offset is 2 and the destination
ssp offset is 1 and the dictionary length is 0(which means the dictionary is
empty), then the destination scheme is 2 bytes after the beginning of the
metadata block(field after dictionary) and destination is 1 byte after the
beginning of the metadata block.
svn path=/trunk/; revision=30682
Within the attached diff file are two source files, packet-dtn.h and
packet-dtn.c. Their function is to decode Bundle Protocol PDUs sent using the
UDP or TCP Convergence Layers. These protocols have been released by the
Internet Research Task Force and are described in RFC 4838 and RFC 5050.
Detailed information on DTN can be obtained at www.dtnrg.org.
svn path=/trunk/; revision=29010
Anders Broman