inside: mark traffic-out once per superpacket, not per segment

sendInsideEncrypt ran connectionManager.Out for every segment -- up to
~45 extra atomic stores per TSO superpacket, all inside writeLock when
boring crypto serializes encryption. One mark in sendInsideMessage
covers the whole superpacket on both the direct and relay paths.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014ugV2edVqoz3tBvq9J6yWp

inside.go

@@ -126,7 +126,6 @@ func (f *Interface) sendInsideEncrypt(hostinfo *HostInfo, ci *ConnectionState, s
 	c := ci.messageCounter.Add(1)
 
 	out := header.Encode(scratch, header.Version, header.Message, 0, hostinfo.remoteIndexId, c)
-	f.connectionManager.Out(hostinfo)
 
 	out, encErr := ci.eKey.EncryptDanger(out, out, seg, c, nb)
 	if noiseutil.EncryptLockNeeded {
@@ -159,6 +158,11 @@ func (f *Interface) sendInsideMessage(hostinfo *HostInfo, pkt tio.Packet, nb []b
 		return
 	}
 
+	// One traffic-out mark covers every segment of the superpacket; doing it
+	// per segment in sendInsideEncrypt paid an atomic store up to ~45 extra
+	// times per TSO packet, inside writeLock under boring crypto.
+	f.connectionManager.Out(hostinfo)
+
 	remote := hostinfo.GetRemote()
 	ecnEnabled := f.ecnEnabled.Load()
 	if hostinfo.lastRebindCount != f.rebindCount {