[Guy Harris <guy@xxxxxxxxxxxx>]

BTW:

	1) some other dissectors for protocols running atop TCP put
	   separate top-level subtrees into the protocol tree for each
	   packet, rather than having a single tree with all the packet
	   under it;

	2) you might want the dissector to handle packets split across
	   TCP segments (and you might be able to handle that *and*
	   multiple SMPP packets per TCP segment with
	   "tcp_dissect_pdus()");

	3) for SMPP-over-X.25, should it do the "multiple packets per
	   lower-level SDU" stuff, or is there only one SMPP packet per
	   X.25 packet?