memcached could extend the error information for address in use

When memcached fails to bind to the port due to address in use it could use /proc/net/tcp and /proc/net/tcp6 to find the socket currently listening to the port and then use /proc/<pid>/fd to try to look up the pid which binds to the port and add that to the error message in addition to "address in use"

This would help understanding why a system fails to start.