Logstash crashes when auto reloading configuration and syslog input plugin has incoming connection

When using the syslog input plugin and using the config auto reload, LS crashes with the following error:

[2017-02-08T09:49:49,873][FATAL][logstash.runner ] An unexpected error occurred! {:error=>#, :backtrace=>["org/jruby/RubyIO.java:3565:in `each'", "/usr/share/logstash/vendor/bundle/jruby/1.9/gems/logstash-input-syslog-3.2.0/lib/logstash/inputs/syslog.rb:182:in `tcp_receiver'", "/usr/share/logstash/vendor/bundle/jruby/1.9/gems/logstash-input-syslog-3.2.0/lib/logstash/inputs/syslog.rb:167:in `tcp_listener'"]}

- Version: 5.1.2
- Operating System: Ubuntu 16.04
- Steps to Reproduce:
* Configure syslog input plugin and auto reload
* Wait for incoming connection from rsyslog
* Change a configuration file

The failure points to logstash-input-syslog-3.2.0/lib/logstash/inputs/syslog.rb, especially tcp_listener at line 167 and tcp_receiver at line 182. Start by reproducing the reported sequence with syslog input, auto reload, an incoming rsyslog connection, and a configuration change. Done means reloading no longer crashes Logstash with Errno::EBADF.

Written by the indexing model from the issue text.