Plugfest 2017: Reset handling in v3.3.0.14 may cause sockets to be unusable for 120 seconds

Observed with a customer's device. Seems as the reset causes an interruption of the graceful close of the TCP connection which causes the socket to be in the TIME_WAIT state for 120 seconds. Also, replies to reset service requests may be delayed for taht period and suddenly, when the reset is already through for a significant period of time, the reply to the reset srevice is sent.

This way, you can easily use up all sockets and render the device unreachable for a certain time. This was easily reproducible by sending the reset 5 times in a row with a host application registered and the Basic Packet Set (a.ka SetConfig). Interesting observation was, that type 1 resets were good and only type 0 resets showed the problem, so it may be a timing issue.