address some possible shutdown races in xfrin

there were two failures during observed in testing, both occurring
when 'rndc halt' was run rather than 'rndc stop' - the latter dumps
zone contents to disk and presumably introduced enough delay to
prevent the races:

- a failure when the zone was shut down and called dns_xfrin_detach()
  before the xfrin had finished connecting; the connect timeout
  terminated without detaching its handle
- a failure when the tcpdns socket timer fired after the outerhandle
  had already been cleared.

this commit incidentally addresses a failure observed in mutexatomic
due to a variable having been initialized incorrectly.

lib/isc/netmgr/tcpdns.c

@@ -98,7 +98,9 @@ dnstcp_readtimeout(uv_timer_t *timer) {
 	REQUIRE(sock->tid == isc_nm_tid());
 
 	/* Close the TCP connection; its closure should fire ours. */
-	isc_nmhandle_detach(&sock->outerhandle);
+	if (sock->outerhandle != NULL) {
+		isc_nmhandle_detach(&sock->outerhandle);
+	}
 }
 
 /*