diff --git a/README.md b/README.md
index 1ea3768..b13db5e 100644
--- a/README.md
+++ b/README.md
@@ -37,14 +37,9 @@ sudo ip link set eip0 up
 sudo ip addr add 192.0.2.1/30 dev eip0
 ```
 
-MTUを省略した場合、Overlay MTUはUnderlayの経路MTUからIPv6、EtherIP、
-Ethernetヘッダーの合計56 byteを差し引いた値に自動調整されます。例えば
-Underlay MTUが1500の場合、Overlay MTUは1444になり、通常の送信では外側IPv6
-フラグメントを生成しません。
-
-明示的に大きなMTUを設定することもできます。その場合は、カプセル化後の
-パケットが実際のUnderlay経路MTUを超えたときだけ送信元でIPv6フラグメントを
-生成します。
+Overlay MTUのデフォルトは1500です。Underlayの経路MTUも1500の場合、
+カプセル化後の外側IPv6パケットはMTUを超えるため、送信元で経路MTU以下の
+IPv6フラグメントに分割します。
 
 ```sh
 sudo ip link set eip0 mtu 1500
diff --git a/etherip6.c b/etherip6.c
index 8747e7c..81860e2 100644
--- a/etherip6.c
+++ b/etherip6.c
@@ -33,7 +33,6 @@
 #define ETHERIP6_DEFAULT_HOP_LIMIT 64
 #define ETHERIP6_MAX_MTU 9000
 #define ETHERIP6_ENCAP_HLEN (sizeof(struct ipv6hdr) + ETHERIP6_HLEN)
-#define ETHERIP6_MTU_OVERHEAD (ETHERIP6_ENCAP_HLEN + ETH_HLEN)
 
 struct etherip6_tunnel {
 	struct list_head list;
@@ -142,35 +141,6 @@ static struct etherip6_tunnel *etherip6_lookup_rx(struct net *net,
 	return etherip6_lookup_unique_rx(net, local, remote, iif, false, false);
 }
 
-static unsigned int etherip6_underlay_mtu(struct net *net,
-					  const struct etherip6_tunnel *tun)
-{
-	struct dst_entry *dst;
-	struct flowi6 fl6 = {};
-	unsigned int mtu = 0;
-
-	fl6.flowi6_proto = IPPROTO_ETHERIP;
-	fl6.flowi6_oif = tun->link;
-	fl6.daddr = tun->remote;
-	fl6.saddr = tun->local;
-
-	dst = ip6_route_output(net, NULL, &fl6);
-	if (!dst->error)
-		mtu = dst_mtu(dst);
-	dst_release(dst);
-
-	return mtu;
-}
-
-static unsigned int etherip6_overlay_mtu(unsigned int underlay_mtu)
-{
-	if (underlay_mtu <= ETHERIP6_MTU_OVERHEAD + ETH_MIN_MTU)
-		return ETH_MIN_MTU;
-
-	return min_t(unsigned int, underlay_mtu - ETHERIP6_MTU_OVERHEAD,
-		     ETHERIP6_MAX_MTU);
-}
-
 static netdev_tx_t etherip6_xmit(struct sk_buff *skb, struct net_device *dev)
 {
 	struct etherip6_tunnel *tun = netdev_priv(dev);
@@ -222,8 +192,15 @@ static netdev_tx_t etherip6_xmit(struct sk_buff *skb, struct net_device *dev)
 
 	skb->protocol = htons(ETH_P_IPV6);
 	skb->dev = dst->dev;
-	/* IPv6 source fragmentation is needed only above the underlay PMTU. */
-	skb->ignore_df = skb->len > dst_mtu(dst);
+	/*
+	 * This skb used to contain an inner Ethernet frame.  Clear its control
+	 * block before handing it to IPv6, then pin source fragmentation to the
+	 * current underlay PMTU.  ip6_local_out() adds the Fragment Header and
+	 * splits oversized packets in ip6_finish_output().
+	 */
+	memset(IP6CB(skb), 0, sizeof(*IP6CB(skb)));
+	IP6CB(skb)->frag_max_size = dst_mtu(dst);
+	skb->ignore_df = 1;
 	skb_dst_set(skb, dst);
 
 	stats = this_cpu_ptr(tun->stats);
@@ -351,15 +328,6 @@ static int etherip6_newlink(struct net_device *dev,
 		return -EINVAL;
 	}
 
-	/* Keep the default overlay packets below the current underlay PMTU. */
-	if (!params->tb[IFLA_MTU]) {
-		unsigned int underlay_mtu;
-
-		underlay_mtu = etherip6_underlay_mtu(dev_net(dev), tun);
-		if (underlay_mtu)
-			dev->mtu = etherip6_overlay_mtu(underlay_mtu);
-	}
-
 	tun->dev = dev;
 	atomic_long_set(&tun->tx_dropped, 0);
 	atomic_long_set(&tun->rx_dropped, 0);