stunner: re-do how Stunner works

It used to make assumptions based on having Anycast IPs that are super
near. Now we're intentionally going to a bunch of different distant
IPs to measure latency.

Also, optimize how the hairpin detection works. No need to STUN on
that socket. Just use that separate socket for sending, once we know
the other UDP4 socket's endpoint. The trick is: make our test probe
also a STUN packet, so it fits through magicsock's existing STUN
routing.

This drops netcheck from ~5 seconds to ~250-500ms.

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>

netcheck/netcheck_test.go

@@ -0,0 +1,31 @@
+// Copyright (c) 2020 Tailscale Inc & AUTHORS All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package netcheck
+
+import (
+	"net"
+	"testing"
+
+	"tailscale.com/stun"
+)
+
+func TestHairpinSTUN(t *testing.T) {
+	c := &Client{
+		hairTX:      stun.NewTxID(),
+		gotHairSTUN: make(chan *net.UDPAddr, 1),
+	}
+	req := stun.Request(c.hairTX)
+	if !stun.Is(req) {
+		t.Fatal("expected STUN message")
+	}
+	if !c.handleHairSTUN(req, nil) {
+		t.Fatal("expected true")
+	}
+	select {
+	case <-c.gotHairSTUN:
+	default:
+		t.Fatal("expected value")
+	}
+}