Log Message:
Update net/scamper to 20141211d.
Patches provided by Matthew Luckie in PR pkg/50654.

ChangeLogs:
https://mailman.caida.org/pipermail/scamper-announce/2015-October/000004.html
https://mailman.caida.org/pipermail/scamper-announce/2015-December/000005.html
https://mailman.caida.org/pipermail/scamper-announce/2016-January/000006.html

tbit
* add support for initial congestion window (ICW) inferences
* add new tests to check response to packets that could have been
  sent by a blind attacker
* add a TCP fast-open implementation, with both experimental
  and official option values
* add support for testing HTTPS and BGP.  drop FTP, DNS, and SMTP
* add sc_tbitblind driver that was used for IMC 2015 paper

trace
* add tx timestamp to hop records
* add dl option, to replace dlts option removed from scamper.
* process UDP responses, if a UDP probe method is used.

ping:
* add tcp-syn ping method.
* fix memory leak when payloads are specified in ping.

sc_ipiddump
* report IPID values from traceroute measurements, where available
* report the source IP address used to probe the destination

sc_filterpolicy:
* add a new scamper driver to test systems for congruent filtering policy
  http://www.caida.org/tools/measurement/scamper/man/sc_filterpolicy.1.pdf

scamper:
* update scamper maximum PPS to 10,000 (from 1000).  Its not 2002 anymore.
* bind to requested source port with UDP sockets.
* set SO_SNDBUF once, when a probe socket is created.
* remove dlts option which was only used by traceroute.
* drop divert socket from privsep, which was not used in scamper anywhere.
* shift socket creation glue from scamper_privsep.c to
  scamper_udp4.c, scamper_icmp6.c, etc.
* fix memory leak when receiving TCP responses in tracelb.
* do not use the global address cache in tracelb: use a local one.
* in qsort with 3-way partition, do not compare items against
  themselves.
* improve performance of warts_addr_t code
* use calloc instead of malloc() -> memset(0) on systems where calloc
  is available.
* do not use the global address cache in ping: most responses are
  either from the destination, or from the same IP address, so
  optimize for that.