diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index e68c8eb4ea8e..4e2caaf8b5da 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/vxlan.c
@@ -135,7 +135,7 @@ struct vxlan_dev {
 	__u16		  port_max;
 	__u8		  tos;		/* TOS override */
 	__u8		  ttl;
-	u32		  flags;	/* VXLAN_F_* below */
+	u32		  flags;	/* VXLAN_F_* in vxlan.h */
 
 	struct work_struct sock_work;
 	struct work_struct igmp_join;
@@ -150,13 +150,6 @@ struct vxlan_dev {
 	struct hlist_head fdb_head[FDB_HASH_SIZE];
 };
 
-#define VXLAN_F_LEARN	0x01
-#define VXLAN_F_PROXY	0x02
-#define VXLAN_F_RSC	0x04
-#define VXLAN_F_L2MISS	0x08
-#define VXLAN_F_L3MISS	0x10
-#define VXLAN_F_IPV6	0x20 /* internal flag */
-
 /* salt for hash table */
 static u32 vxlan_salt __read_mostly;
 static struct workqueue_struct *vxlan_wq;
@@ -1601,18 +1594,11 @@ __be16 vxlan_src_port(__u16 port_min, __u16 port_max, struct sk_buff *skb)
 }
 EXPORT_SYMBOL_GPL(vxlan_src_port);
 
-static int handle_offloads(struct sk_buff *skb)
+static inline struct sk_buff *vxlan_handle_offloads(struct sk_buff *skb,
+						    bool udp_csum)
 {
-	if (skb_is_gso(skb)) {
-		int err = skb_unclone(skb, GFP_ATOMIC);
-		if (unlikely(err))
-			return err;
-
-		skb_shinfo(skb)->gso_type |= SKB_GSO_UDP_TUNNEL;
-	} else if (skb->ip_summed != CHECKSUM_PARTIAL)
-		skb->ip_summed = CHECKSUM_NONE;
-
-	return 0;
+	int type = udp_csum ? SKB_GSO_UDP_TUNNEL_CSUM : SKB_GSO_UDP_TUNNEL;
+	return iptunnel_handle_offloads(skb, udp_csum, type);
 }
 
 #if IS_ENABLED(CONFIG_IPV6)
@@ -1629,10 +1615,9 @@ static int vxlan6_xmit_skb(struct vxlan_sock *vs,
 	int min_headroom;
 	int err;
 
-	if (!skb->encapsulation) {
-		skb_reset_inner_headers(skb);
-		skb->encapsulation = 1;
-	}
+	skb = vxlan_handle_offloads(skb, !udp_get_no_check6_tx(vs->sock->sk));
+	if (IS_ERR(skb))
+		return -EINVAL;
 
 	skb_scrub_packet(skb, xnet);
 
@@ -1666,27 +1651,14 @@ static int vxlan6_xmit_skb(struct vxlan_sock *vs,
 	uh->source = src_port;
 
 	uh->len = htons(skb->len);
-	uh->check = 0;
 
 	memset(&(IPCB(skb)->opt), 0, sizeof(IPCB(skb)->opt));
 	IPCB(skb)->flags &= ~(IPSKB_XFRM_TUNNEL_SIZE | IPSKB_XFRM_TRANSFORMED |
 			      IPSKB_REROUTED);
 	skb_dst_set(skb, dst);
 
-	if (!skb_is_gso(skb) && !(dst->dev->features & NETIF_F_IPV6_CSUM)) {
-		__wsum csum = skb_checksum(skb, 0, skb->len, 0);
-		skb->ip_summed = CHECKSUM_UNNECESSARY;
-		uh->check = csum_ipv6_magic(saddr, daddr, skb->len,
-					    IPPROTO_UDP, csum);
-		if (uh->check == 0)
-			uh->check = CSUM_MANGLED_0;
-	} else {
-		skb->ip_summed = CHECKSUM_PARTIAL;
-		skb->csum_start = skb_transport_header(skb) - skb->head;
-		skb->csum_offset = offsetof(struct udphdr, check);
-		uh->check = ~csum_ipv6_magic(saddr, daddr,
-					     skb->len, IPPROTO_UDP, 0);
-	}
+	udp6_set_csum(udp_get_no_check6_tx(vs->sock->sk), skb,
+		      saddr, daddr, skb->len);
 
 	__skb_push(skb, sizeof(*ip6h));
 	skb_reset_network_header(skb);
@@ -1702,10 +1674,6 @@ static int vxlan6_xmit_skb(struct vxlan_sock *vs,
 	ip6h->daddr	  = *daddr;
 	ip6h->saddr	  = *saddr;
 
-	err = handle_offloads(skb);
-	if (err)
-		return err;
-
 	ip6tunnel_xmit(skb, dev);
 	return 0;
 }
@@ -1721,10 +1689,9 @@ int vxlan_xmit_skb(struct vxlan_sock *vs,
 	int min_headroom;
 	int err;
 
-	if (!skb->encapsulation) {
-		skb_reset_inner_headers(skb);
-		skb->encapsulation = 1;
-	}
+	skb = vxlan_handle_offloads(skb, !vs->sock->sk->sk_no_check_tx);
+	if (IS_ERR(skb))
+		return -EINVAL;
 
 	min_headroom = LL_RESERVED_SPACE(rt->dst.dev) + rt->dst.header_len
 			+ VXLAN_HLEN + sizeof(struct iphdr)
@@ -1756,11 +1723,9 @@ int vxlan_xmit_skb(struct vxlan_sock *vs,
 	uh->source = src_port;
 
 	uh->len = htons(skb->len);
-	uh->check = 0;
 
-	err = handle_offloads(skb);
-	if (err)
-		return err;
+	udp_set_csum(vs->sock->sk->sk_no_check_tx, skb,
+		     src, dst, skb->len);
 
 	return iptunnel_xmit(vs->sock->sk, rt, skb, src, dst, IPPROTO_UDP,
 			     tos, ttl, df, xnet);
@@ -2405,7 +2370,7 @@ static void vxlan_del_work(struct work_struct *work)
  * could be used for both IPv4 and IPv6 communications, but
  * users may set bindv6only=1.
  */
-static struct socket *create_v6_sock(struct net *net, __be16 port)
+static struct socket *create_v6_sock(struct net *net, __be16 port, u32 flags)
 {
 	struct sock *sk;
 	struct socket *sock;
@@ -2442,18 +2407,25 @@ static struct socket *create_v6_sock(struct net *net, __be16 port)
 
 	/* Disable multicast loopback */
 	inet_sk(sk)->mc_loop = 0;
+
+	if (flags & VXLAN_F_UDP_ZERO_CSUM6_TX)
+		udp_set_no_check6_tx(sk, true);
+
+	if (flags & VXLAN_F_UDP_ZERO_CSUM6_RX)
+		udp_set_no_check6_rx(sk, true);
+
 	return sock;
 }
 
 #else
 
-static struct socket *create_v6_sock(struct net *net, __be16 port)
+static struct socket *create_v6_sock(struct net *net, __be16 port, u32 flags)
 {
 		return ERR_PTR(-EPFNOSUPPORT);
 }
 #endif
 
-static struct socket *create_v4_sock(struct net *net, __be16 port)
+static struct socket *create_v4_sock(struct net *net, __be16 port, u32 flags)
 {
 	struct sock *sk;
 	struct socket *sock;
@@ -2486,18 +2458,24 @@ static struct socket *create_v4_sock(struct net *net, __be16 port)
 
 	/* Disable multicast loopback */
 	inet_sk(sk)->mc_loop = 0;
+
+	if (!(flags & VXLAN_F_UDP_CSUM))
+		sock->sk->sk_no_check_tx = 1;
+
 	return sock;
 }
 
 /* Create new listen socket if needed */
 static struct vxlan_sock *vxlan_socket_create(struct net *net, __be16 port,
-					      vxlan_rcv_t *rcv, void *data, bool ipv6)
+					      vxlan_rcv_t *rcv, void *data,
+					      u32 flags)
 {
 	struct vxlan_net *vn = net_generic(net, vxlan_net_id);
 	struct vxlan_sock *vs;
 	struct socket *sock;
 	struct sock *sk;
 	unsigned int h;
+	bool ipv6 = !!(flags & VXLAN_F_IPV6);
 
 	vs = kzalloc(sizeof(*vs), GFP_KERNEL);
 	if (!vs)
@@ -2509,9 +2487,9 @@ static struct vxlan_sock *vxlan_socket_create(struct net *net, __be16 port,
 	INIT_WORK(&vs->del_work, vxlan_del_work);
 
 	if (ipv6)
-		sock = create_v6_sock(net, port);
+		sock = create_v6_sock(net, port, flags);
 	else
-		sock = create_v4_sock(net, port);
+		sock = create_v4_sock(net, port, flags);
 	if (IS_ERR(sock)) {
 		kfree(vs);
 		return ERR_CAST(sock);
@@ -2549,12 +2527,12 @@ static struct vxlan_sock *vxlan_socket_create(struct net *net, __be16 port,
 
 struct vxlan_sock *vxlan_sock_add(struct net *net, __be16 port,
 				  vxlan_rcv_t *rcv, void *data,
-				  bool no_share, bool ipv6)
+				  bool no_share, u32 flags)
 {
 	struct vxlan_net *vn = net_generic(net, vxlan_net_id);
 	struct vxlan_sock *vs;
 
-	vs = vxlan_socket_create(net, port, rcv, data, ipv6);
+	vs = vxlan_socket_create(net, port, rcv, data, flags);
 	if (!IS_ERR(vs))
 		return vs;
 
@@ -2587,7 +2565,7 @@ static void vxlan_sock_work(struct work_struct *work)
 	__be16 port = vxlan->dst_port;
 	struct vxlan_sock *nvs;
 
-	nvs = vxlan_sock_add(net, port, vxlan_rcv, NULL, false, vxlan->flags & VXLAN_F_IPV6);
+	nvs = vxlan_sock_add(net, port, vxlan_rcv, NULL, false, vxlan->flags);
 	spin_lock(&vn->sock_lock);
 	if (!IS_ERR(nvs))
 		vxlan_vs_add_dev(nvs, vxlan);
@@ -2711,6 +2689,17 @@ static int vxlan_newlink(struct net *net, struct net_device *dev,
 	if (data[IFLA_VXLAN_PORT])
 		vxlan->dst_port = nla_get_be16(data[IFLA_VXLAN_PORT]);
 
+	if (data[IFLA_VXLAN_UDP_CSUM] && nla_get_u8(data[IFLA_VXLAN_UDP_CSUM]))
+		vxlan->flags |= VXLAN_F_UDP_CSUM;
+
+	if (data[IFLA_VXLAN_UDP_ZERO_CSUM6_TX] &&
+	    nla_get_u8(data[IFLA_VXLAN_UDP_ZERO_CSUM6_TX]))
+		vxlan->flags |= VXLAN_F_UDP_ZERO_CSUM6_TX;
+
+	if (data[IFLA_VXLAN_UDP_ZERO_CSUM6_RX] &&
+	    nla_get_u8(data[IFLA_VXLAN_UDP_ZERO_CSUM6_RX]))
+		vxlan->flags |= VXLAN_F_UDP_ZERO_CSUM6_RX;
+
 	if (vxlan_find_vni(net, vni, vxlan->dst_port)) {
 		pr_info("duplicate VNI %u\n", vni);
 		return -EEXIST;
@@ -2774,7 +2763,10 @@ static size_t vxlan_get_size(const struct net_device *dev)
 		nla_total_size(sizeof(__u32)) +	/* IFLA_VXLAN_AGEING */
 		nla_total_size(sizeof(__u32)) +	/* IFLA_VXLAN_LIMIT */
 		nla_total_size(sizeof(struct ifla_vxlan_port_range)) +
-		nla_total_size(sizeof(__be16))+ /* IFLA_VXLAN_PORT */
+		nla_total_size(sizeof(__be16)) + /* IFLA_VXLAN_PORT */
+		nla_total_size(sizeof(__u8)) + /* IFLA_VXLAN_UDP_CSUM */
+		nla_total_size(sizeof(__u8)) + /* IFLA_VXLAN_UDP_ZERO_CSUM6_TX */
+		nla_total_size(sizeof(__u8)) + /* IFLA_VXLAN_UDP_ZERO_CSUM6_RX */
 		0;
 }
 
@@ -2834,7 +2826,13 @@ static int vxlan_fill_info(struct sk_buff *skb, const struct net_device *dev)
 			!!(vxlan->flags & VXLAN_F_L3MISS)) ||
 	    nla_put_u32(skb, IFLA_VXLAN_AGEING, vxlan->age_interval) ||
 	    nla_put_u32(skb, IFLA_VXLAN_LIMIT, vxlan->addrmax) ||
-	    nla_put_be16(skb, IFLA_VXLAN_PORT, vxlan->dst_port))
+	    nla_put_be16(skb, IFLA_VXLAN_PORT, vxlan->dst_port) ||
+	    nla_put_u8(skb, IFLA_VXLAN_UDP_CSUM,
+			!!(vxlan->flags & VXLAN_F_UDP_CSUM)) ||
+	    nla_put_u8(skb, IFLA_VXLAN_UDP_ZERO_CSUM6_TX,
+			!!(vxlan->flags & VXLAN_F_UDP_ZERO_CSUM6_TX)) ||
+	    nla_put_u8(skb, IFLA_VXLAN_UDP_ZERO_CSUM6_RX,
+			!!(vxlan->flags & VXLAN_F_UDP_ZERO_CSUM6_RX)))
 		goto nla_put_failure;
 
 	if (nla_put(skb, IFLA_VXLAN_PORT_RANGE, sizeof(ports), &ports))
diff --git a/include/linux/netdev_features.h b/include/linux/netdev_features.h
index c26d0ec2ef3a..e5a589435e2b 100644
--- a/include/linux/netdev_features.h
+++ b/include/linux/netdev_features.h
@@ -42,9 +42,11 @@ enum {
 	NETIF_F_TSO6_BIT,		/* ... TCPv6 segmentation */
 	NETIF_F_FSO_BIT,		/* ... FCoE segmentation */
 	NETIF_F_GSO_GRE_BIT,		/* ... GRE with TSO */
+	NETIF_F_GSO_GRE_CSUM_BIT,	/* ... GRE with csum with TSO */
 	NETIF_F_GSO_IPIP_BIT,		/* ... IPIP tunnel with TSO */
 	NETIF_F_GSO_SIT_BIT,		/* ... SIT tunnel with TSO */
 	NETIF_F_GSO_UDP_TUNNEL_BIT,	/* ... UDP TUNNEL with TSO */
+	NETIF_F_GSO_UDP_TUNNEL_CSUM_BIT,/* ... UDP TUNNEL with TSO & CSUM */
 	NETIF_F_GSO_MPLS_BIT,		/* ... MPLS segmentation */
 	/**/NETIF_F_GSO_LAST =		/* last bit, see GSO_MASK */
 		NETIF_F_GSO_MPLS_BIT,
@@ -111,6 +113,7 @@ enum {
 #define NETIF_F_RXFCS		__NETIF_F(RXFCS)
 #define NETIF_F_RXALL		__NETIF_F(RXALL)
 #define NETIF_F_GSO_GRE		__NETIF_F(GSO_GRE)
+#define NETIF_F_GSO_GRE_CSUM	__NETIF_F(GSO_GRE_CSUM)
 #define NETIF_F_GSO_IPIP	__NETIF_F(GSO_IPIP)
 #define NETIF_F_GSO_SIT		__NETIF_F(GSO_SIT)
 #define NETIF_F_GSO_UDP_TUNNEL	__NETIF_F(GSO_UDP_TUNNEL)
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 7a9beeb1c458..c705808bef9c 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -345,6 +345,10 @@ enum {
 	SKB_GSO_UDP_TUNNEL = 1 << 9,
 
 	SKB_GSO_MPLS = 1 << 10,
+
+	SKB_GSO_UDP_TUNNEL_CSUM = 1 << 11,
+
+	SKB_GSO_GRE_CSUM = 1 << 12,
 };
 
 #if BITS_PER_LONG > 32
@@ -567,7 +571,8 @@ struct sk_buff {
 	 * headers if needed
 	 */
 	__u8			encapsulation:1;
-	/* 6/8 bit hole (depending on ndisc_nodetype presence) */
+	__u8			encap_hdr_csum:1;
+	/* 5/7 bit hole (depending on ndisc_nodetype presence) */
 	kmemcheck_bitfield_end(flags2);
 
 #if defined CONFIG_NET_DMA || defined CONFIG_NET_RX_BUSY_POLL
@@ -2988,6 +2993,7 @@ static inline struct sec_path *skb_sec_path(struct sk_buff *skb)
 struct skb_gso_cb {
 	int	mac_offset;
 	int	encap_level;
+	__u16	csum_start;
 };
 #define SKB_GSO_CB(skb) ((struct skb_gso_cb *)(skb)->cb)
 
@@ -3012,6 +3018,28 @@ static inline int gso_pskb_expand_head(struct sk_buff *skb, int extra)
 	return 0;
 }
 
+/* Compute the checksum for a gso segment. First compute the checksum value
+ * from the start of transport header to SKB_GSO_CB(skb)->csum_start, and
+ * then add in skb->csum (checksum from csum_start to end of packet).
+ * skb->csum and csum_start are then updated to reflect the checksum of the
+ * resultant packet starting from the transport header-- the resultant checksum
+ * is in the res argument (i.e. normally zero or ~ of checksum of a pseudo
+ * header.
+ */
+static inline __sum16 gso_make_checksum(struct sk_buff *skb, __wsum res)
+{
+	int plen = SKB_GSO_CB(skb)->csum_start - skb_headroom(skb) -
+	    skb_transport_offset(skb);
+	__u16 csum;
+
+	csum = csum_fold(csum_partial(skb_transport_header(skb),
+				      plen, skb->csum));
+	skb->csum = res;
+	SKB_GSO_CB(skb)->csum_start -= plen;
+
+	return csum;
+}
+
 static inline bool skb_is_gso(const struct sk_buff *skb)
 {
 	return skb_shinfo(skb)->gso_size;
diff --git a/include/net/gre.h b/include/net/gre.h
index 70046a0b0b89..b53182018743 100644
--- a/include/net/gre.h
+++ b/include/net/gre.h
@@ -37,9 +37,10 @@ void gre_build_header(struct sk_buff *skb, const struct tnl_ptk_info *tpi,
 		      int hdr_len);
 
 static inline struct sk_buff *gre_handle_offloads(struct sk_buff *skb,
-						  bool gre_csum)
+						  bool csum)
 {
-	return iptunnel_handle_offloads(skb, gre_csum, SKB_GSO_GRE);
+	return iptunnel_handle_offloads(skb, csum,
+					csum ? SKB_GSO_GRE_CSUM : SKB_GSO_GRE);
 }
 
 
diff --git a/include/net/ip6_checksum.h b/include/net/ip6_checksum.h
index 8ac5c21f8456..55236cb71174 100644
--- a/include/net/ip6_checksum.h
+++ b/include/net/ip6_checksum.h
@@ -82,5 +82,17 @@ static inline void tcp_v6_send_check(struct sock *sk, struct sk_buff *skb)
 }
 #endif
 
+static inline __sum16 udp_v6_check(int len,
+				   const struct in6_addr *saddr,
+				   const struct in6_addr *daddr,
+				   __wsum base)
+{
+	return csum_ipv6_magic(saddr, daddr, len, IPPROTO_UDP, base);
+}
+
+void udp6_set_csum(bool nocheck, struct sk_buff *skb,
+		   const struct in6_addr *saddr,
+		   const struct in6_addr *daddr, int len);
+
 int udp6_csum_init(struct sk_buff *skb, struct udphdr *uh, int proto);
 #endif
diff --git a/include/net/udp.h b/include/net/udp.h
index 5eb86874bcd6..2ecfc6e15609 100644
--- a/include/net/udp.h
+++ b/include/net/udp.h
@@ -147,6 +147,15 @@ static inline __wsum udp_csum(struct sk_buff *skb)
 	return csum;
 }
 
+static inline __sum16 udp_v4_check(int len, __be32 saddr,
+				   __be32 daddr, __wsum base)
+{
+	return csum_tcpudp_magic(saddr, daddr, len, IPPROTO_UDP, base);
+}
+
+void udp_set_csum(bool nocheck, struct sk_buff *skb,
+		  __be32 saddr, __be32 daddr, int len);
+
 /* hash routines shared between UDPv4/6 and UDP-Litev4/6 */
 static inline void udp_lib_hash(struct sock *sk)
 {
diff --git a/include/net/vxlan.h b/include/net/vxlan.h
index 7bb4084b1bd0..12196ce661d9 100644
--- a/include/net/vxlan.h
+++ b/include/net/vxlan.h
@@ -24,9 +24,19 @@ struct vxlan_sock {
 	struct udp_offload udp_offloads;
 };
 
+#define VXLAN_F_LEARN			0x01
+#define VXLAN_F_PROXY			0x02
+#define VXLAN_F_RSC			0x04
+#define VXLAN_F_L2MISS			0x08
+#define VXLAN_F_L3MISS			0x10
+#define VXLAN_F_IPV6			0x20
+#define VXLAN_F_UDP_CSUM		0x40
+#define VXLAN_F_UDP_ZERO_CSUM6_TX	0x80
+#define VXLAN_F_UDP_ZERO_CSUM6_RX	0x100
+
 struct vxlan_sock *vxlan_sock_add(struct net *net, __be16 port,
 				  vxlan_rcv_t *rcv, void *data,
-				  bool no_share, bool ipv6);
+				  bool no_share, u32 flags);
 
 void vxlan_sock_release(struct vxlan_sock *vs);
 
diff --git a/include/uapi/linux/if_link.h b/include/uapi/linux/if_link.h
index 622e7910b8cc..b38534895db5 100644
--- a/include/uapi/linux/if_link.h
+++ b/include/uapi/linux/if_link.h
@@ -319,6 +319,9 @@ enum {
 	IFLA_VXLAN_PORT,	/* destination port */
 	IFLA_VXLAN_GROUP6,
 	IFLA_VXLAN_LOCAL6,
+	IFLA_VXLAN_UDP_CSUM,
+	IFLA_VXLAN_UDP_ZERO_CSUM6_TX,
+	IFLA_VXLAN_UDP_ZERO_CSUM6_RX,
 	__IFLA_VXLAN_MAX
 };
 #define IFLA_VXLAN_MAX	(__IFLA_VXLAN_MAX - 1)
diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index 3f6c7e8be8a4..05f4bef2ce12 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -2885,7 +2885,9 @@ struct sk_buff *skb_segment(struct sk_buff *head_skb,
 	if (unlikely(!proto))
 		return ERR_PTR(-EINVAL);
 
-	csum = !!can_checksum_protocol(features, proto);
+	csum = !head_skb->encap_hdr_csum &&
+	    !!can_checksum_protocol(features, proto);
+
 	__skb_push(head_skb, doffset);
 	headroom = skb_headroom(head_skb);
 	pos = skb_headlen(head_skb);
@@ -2983,6 +2985,8 @@ struct sk_buff *skb_segment(struct sk_buff *head_skb,
 			nskb->csum = skb_copy_and_csum_bits(head_skb, offset,
 							    skb_put(nskb, len),
 							    len, 0);
+			SKB_GSO_CB(nskb)->csum_start =
+			    skb_headroom(nskb) + offset;
 			continue;
 		}
 
@@ -3052,6 +3056,8 @@ perform_csum_check:
 			nskb->csum = skb_checksum(nskb, doffset,
 						  nskb->len - doffset, 0);
 			nskb->ip_summed = CHECKSUM_NONE;
+			SKB_GSO_CB(nskb)->csum_start =
+			    skb_headroom(nskb) + doffset;
 		}
 	} while ((offset += len) < head_skb->len);
 
diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c
index 0e9bb08a91e4..d5e6836cf772 100644
--- a/net/ipv4/af_inet.c
+++ b/net/ipv4/af_inet.c
@@ -1254,10 +1254,12 @@ static struct sk_buff *inet_gso_segment(struct sk_buff *skb,
 		       SKB_GSO_DODGY |
 		       SKB_GSO_TCP_ECN |
 		       SKB_GSO_GRE |
+		       SKB_GSO_GRE_CSUM |
 		       SKB_GSO_IPIP |
 		       SKB_GSO_SIT |
 		       SKB_GSO_TCPV6 |
 		       SKB_GSO_UDP_TUNNEL |
+		       SKB_GSO_UDP_TUNNEL_CSUM |
 		       SKB_GSO_MPLS |
 		       0)))
 		goto out;
diff --git a/net/ipv4/gre_demux.c b/net/ipv4/gre_demux.c
index fbfd829f4049..4e9619bca732 100644
--- a/net/ipv4/gre_demux.c
+++ b/net/ipv4/gre_demux.c
@@ -84,7 +84,8 @@ void gre_build_header(struct sk_buff *skb, const struct tnl_ptk_info *tpi,
 			ptr--;
 		}
 		if (tpi->flags&TUNNEL_CSUM &&
-		    !(skb_shinfo(skb)->gso_type & SKB_GSO_GRE)) {
+		    !(skb_shinfo(skb)->gso_type &
+		      (SKB_GSO_GRE|SKB_GSO_GRE_CSUM))) {
 			*ptr = 0;
 			*(__sum16 *)ptr = csum_fold(skb_checksum(skb, 0,
 								 skb->len, 0));
diff --git a/net/ipv4/gre_offload.c b/net/ipv4/gre_offload.c
index f1d32280cb54..24deb3928b9e 100644
--- a/net/ipv4/gre_offload.c
+++ b/net/ipv4/gre_offload.c
@@ -42,6 +42,7 @@ static struct sk_buff *gre_gso_segment(struct sk_buff *skb,
 				  SKB_GSO_DODGY |
 				  SKB_GSO_TCP_ECN |
 				  SKB_GSO_GRE |
+				  SKB_GSO_GRE_CSUM |
 				  SKB_GSO_IPIP)))
 		goto out;
 
@@ -55,6 +56,8 @@ static struct sk_buff *gre_gso_segment(struct sk_buff *skb,
 		goto out;
 
 	csum = !!(greh->flags & GRE_CSUM);
+	if (csum)
+		skb->encap_hdr_csum = 1;
 
 	if (unlikely(!pskb_may_pull(skb, ghl)))
 		goto out;
@@ -94,10 +97,13 @@ static struct sk_buff *gre_gso_segment(struct sk_buff *skb,
 				}
 			}
 
-			greh = (struct gre_base_hdr *)(skb->data);
+			skb_reset_transport_header(skb);
+
+			greh = (struct gre_base_hdr *)
+			    skb_transport_header(skb);
 			pcsum = (__be32 *)(greh + 1);
 			*pcsum = 0;
-			*(__sum16 *)pcsum = csum_fold(skb_checksum(skb, 0, skb->len, 0));
+			*(__sum16 *)pcsum = gso_make_checksum(skb, 0);
 		}
 		__skb_push(skb, tnl_hlen - ghl);
 
diff --git a/net/ipv4/ip_tunnel_core.c b/net/ipv4/ip_tunnel_core.c
index 847e69cbff7e..f4c987bb7e94 100644
--- a/net/ipv4/ip_tunnel_core.c
+++ b/net/ipv4/ip_tunnel_core.c
@@ -135,6 +135,14 @@ struct sk_buff *iptunnel_handle_offloads(struct sk_buff *skb,
 		return skb;
 	}
 
+	/* If packet is not gso and we are resolving any partial checksum,
+	 * clear encapsulation flag. This allows setting CHECKSUM_PARTIAL
+	 * on the outer header without confusing devices that implement
+	 * NETIF_F_IP_CSUM with encapsulation.
+	 */
+	if (csum_help)
+		skb->encapsulation = 0;
+
 	if (skb->ip_summed == CHECKSUM_PARTIAL && csum_help) {
 		err = skb_checksum_help(skb);
 		if (unlikely(err))
diff --git a/net/ipv4/tcp_offload.c b/net/ipv4/tcp_offload.c
index b92b81718ca4..4e86c59ec7f7 100644
--- a/net/ipv4/tcp_offload.c
+++ b/net/ipv4/tcp_offload.c
@@ -57,10 +57,12 @@ struct sk_buff *tcp_gso_segment(struct sk_buff *skb,
 			       SKB_GSO_TCP_ECN |
 			       SKB_GSO_TCPV6 |
 			       SKB_GSO_GRE |
+			       SKB_GSO_GRE_CSUM |
 			       SKB_GSO_IPIP |
 			       SKB_GSO_SIT |
 			       SKB_GSO_MPLS |
 			       SKB_GSO_UDP_TUNNEL |
+			       SKB_GSO_UDP_TUNNEL_CSUM |
 			       0) ||
 			     !(type & (SKB_GSO_TCPV4 | SKB_GSO_TCPV6))))
 			goto out;
@@ -97,9 +99,7 @@ struct sk_buff *tcp_gso_segment(struct sk_buff *skb,
 		th->check = newcheck;
 
 		if (skb->ip_summed != CHECKSUM_PARTIAL)
-			th->check =
-			     csum_fold(csum_partial(skb_transport_header(skb),
-						    thlen, skb->csum));
+			th->check = gso_make_checksum(skb, ~th->check);
 
 		seq += mss;
 		if (copy_destructor) {
@@ -133,8 +133,7 @@ struct sk_buff *tcp_gso_segment(struct sk_buff *skb,
 	th->check = ~csum_fold((__force __wsum)((__force u32)th->check +
 				(__force u32)delta));
 	if (skb->ip_summed != CHECKSUM_PARTIAL)
-		th->check = csum_fold(csum_partial(skb_transport_header(skb),
-						   thlen, skb->csum));
+		th->check = gso_make_checksum(skb, ~th->check);
 out:
 	return segs;
 }
diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c
index e07d52b8617a..8d8c33d84c9a 100644
--- a/net/ipv4/udp.c
+++ b/net/ipv4/udp.c
@@ -762,6 +762,43 @@ void udp4_hwcsum(struct sk_buff *skb, __be32 src, __be32 dst)
 }
 EXPORT_SYMBOL_GPL(udp4_hwcsum);
 
+/* Function to set UDP checksum for an IPv4 UDP packet. This is intended
+ * for the simple case like when setting the checksum for a UDP tunnel.
+ */
+void udp_set_csum(bool nocheck, struct sk_buff *skb,
+		  __be32 saddr, __be32 daddr, int len)
+{
+	struct udphdr *uh = udp_hdr(skb);
+
+	if (nocheck)
+		uh->check = 0;
+	else if (skb_is_gso(skb))
+		uh->check = ~udp_v4_check(len, saddr, daddr, 0);
+	else if (skb_dst(skb) && skb_dst(skb)->dev &&
+		 (skb_dst(skb)->dev->features & NETIF_F_V4_CSUM)) {
+
+		BUG_ON(skb->ip_summed == CHECKSUM_PARTIAL);
+
+		skb->ip_summed = CHECKSUM_PARTIAL;
+		skb->csum_start = skb_transport_header(skb) - skb->head;
+		skb->csum_offset = offsetof(struct udphdr, check);
+		uh->check = ~udp_v4_check(len, saddr, daddr, 0);
+	} else {
+		__wsum csum;
+
+		BUG_ON(skb->ip_summed == CHECKSUM_PARTIAL);
+
+		uh->check = 0;
+		csum = skb_checksum(skb, 0, len, 0);
+		uh->check = udp_v4_check(len, saddr, daddr, csum);
+		if (uh->check == 0)
+			uh->check = CSUM_MANGLED_0;
+
+		skb->ip_summed = CHECKSUM_UNNECESSARY;
+	}
+}
+EXPORT_SYMBOL(udp_set_csum);
+
 static int udp_send_skb(struct sk_buff *skb, struct flowi4 *fl4)
 {
 	struct sock *sk = skb->sk;
@@ -2491,7 +2528,11 @@ struct sk_buff *skb_udp_tunnel_segment(struct sk_buff *skb,
 	int tnl_hlen = skb_inner_mac_header(skb) - skb_transport_header(skb);
 	__be16 protocol = skb->protocol;
 	netdev_features_t enc_features;
-	int outer_hlen;
+	int udp_offset, outer_hlen;
+	unsigned int oldlen;
+	bool need_csum;
+
+	oldlen = (u16)~skb->len;
 
 	if (unlikely(!pskb_may_pull(skb, tnl_hlen)))
 		goto out;
@@ -2503,6 +2544,10 @@ struct sk_buff *skb_udp_tunnel_segment(struct sk_buff *skb,
 	skb->mac_len = skb_inner_network_offset(skb);
 	skb->protocol = htons(ETH_P_TEB);
 
+	need_csum = !!(skb_shinfo(skb)->gso_type & SKB_GSO_UDP_TUNNEL_CSUM);
+	if (need_csum)
+		skb->encap_hdr_csum = 1;
+
 	/* segment inner packet. */
 	enc_features = skb->dev->hw_enc_features & netif_skb_features(skb);
 	segs = skb_mac_gso_segment(skb, enc_features);
@@ -2513,10 +2558,11 @@ struct sk_buff *skb_udp_tunnel_segment(struct sk_buff *skb,
 	}
 
 	outer_hlen = skb_tnl_header_len(skb);
+	udp_offset = outer_hlen - tnl_hlen;
 	skb = segs;
 	do {
 		struct udphdr *uh;
-		int udp_offset = outer_hlen - tnl_hlen;
+		int len;
 
 		skb_reset_inner_headers(skb);
 		skb->encapsulation = 1;
@@ -2527,31 +2573,20 @@ struct sk_buff *skb_udp_tunnel_segment(struct sk_buff *skb,
 		skb_reset_mac_header(skb);
 		skb_set_network_header(skb, mac_len);
 		skb_set_transport_header(skb, udp_offset);
+		len = skb->len - udp_offset;
 		uh = udp_hdr(skb);
-		uh->len = htons(skb->len - udp_offset);
+		uh->len = htons(len);
 
-		/* csum segment if tunnel sets skb with csum. */
-		if (protocol == htons(ETH_P_IP) && unlikely(uh->check)) {
-			struct iphdr *iph = ip_hdr(skb);
+		if (need_csum) {
+			__be32 delta = htonl(oldlen + len);
+
+			uh->check = ~csum_fold((__force __wsum)
+					       ((__force u32)uh->check +
+						(__force u32)delta));
+			uh->check = gso_make_checksum(skb, ~uh->check);
 
-			uh->check = ~csum_tcpudp_magic(iph->saddr, iph->daddr,
-						       skb->len - udp_offset,
-						       IPPROTO_UDP, 0);
-			uh->check = csum_fold(skb_checksum(skb, udp_offset,
-							   skb->len - udp_offset, 0));
 			if (uh->check == 0)
 				uh->check = CSUM_MANGLED_0;
-
-		} else if (protocol == htons(ETH_P_IPV6)) {
-			struct ipv6hdr *ipv6h = ipv6_hdr(skb);
-			u32 len = skb->len - udp_offset;
-
-			uh->check = ~csum_ipv6_magic(&ipv6h->saddr, &ipv6h->daddr,
-						     len, IPPROTO_UDP, 0);
-			uh->check = csum_fold(skb_checksum(skb, udp_offset, len, 0));
-			if (uh->check == 0)
-				uh->check = CSUM_MANGLED_0;
-			skb->ip_summed = CHECKSUM_NONE;
 		}
 
 		skb->protocol = protocol;
diff --git a/net/ipv4/udp_offload.c b/net/ipv4/udp_offload.c
index 88b4023ecfcf..7b1840110173 100644
--- a/net/ipv4/udp_offload.c
+++ b/net/ipv4/udp_offload.c
@@ -56,7 +56,8 @@ static struct sk_buff *udp4_ufo_fragment(struct sk_buff *skb,
 	__wsum csum;
 
 	if (skb->encapsulation &&
-	    skb_shinfo(skb)->gso_type & SKB_GSO_UDP_TUNNEL) {
+	    (skb_shinfo(skb)->gso_type &
+	     (SKB_GSO_UDP_TUNNEL|SKB_GSO_UDP_TUNNEL_CSUM))) {
 		segs = skb_udp_tunnel_segment(skb, features);
 		goto out;
 	}
@@ -71,8 +72,10 @@ static struct sk_buff *udp4_ufo_fragment(struct sk_buff *skb,
 
 		if (unlikely(type & ~(SKB_GSO_UDP | SKB_GSO_DODGY |
 				      SKB_GSO_UDP_TUNNEL |
+				      SKB_GSO_UDP_TUNNEL_CSUM |
 				      SKB_GSO_IPIP |
-				      SKB_GSO_GRE | SKB_GSO_MPLS) ||
+				      SKB_GSO_GRE | SKB_GSO_GRE_CSUM |
+				      SKB_GSO_MPLS) ||
 			     !(type & (SKB_GSO_UDP))))
 			goto out;
 
diff --git a/net/ipv6/ip6_checksum.c b/net/ipv6/ip6_checksum.c
index da26224a5993..9a4d7322fb22 100644
--- a/net/ipv6/ip6_checksum.c
+++ b/net/ipv6/ip6_checksum.c
@@ -84,3 +84,41 @@ int udp6_csum_init(struct sk_buff *skb, struct udphdr *uh, int proto)
 					   ip6_compute_pseudo);
 }
 EXPORT_SYMBOL(udp6_csum_init);
+
+/* Function to set UDP checksum for an IPv6 UDP packet. This is intended
+ * for the simple case like when setting the checksum for a UDP tunnel.
+ */
+void udp6_set_csum(bool nocheck, struct sk_buff *skb,
+		   const struct in6_addr *saddr,
+		   const struct in6_addr *daddr, int len)
+{
+	struct udphdr *uh = udp_hdr(skb);
+
+	if (nocheck)
+		uh->check = 0;
+	else if (skb_is_gso(skb))
+		uh->check = ~udp_v6_check(len, saddr, daddr, 0);
+	else if (skb_dst(skb) && skb_dst(skb)->dev &&
+		 (skb_dst(skb)->dev->features & NETIF_F_IPV6_CSUM)) {
+
+		BUG_ON(skb->ip_summed == CHECKSUM_PARTIAL);
+
+		skb->ip_summed = CHECKSUM_PARTIAL;
+		skb->csum_start = skb_transport_header(skb) - skb->head;
+		skb->csum_offset = offsetof(struct udphdr, check);
+		uh->check = ~udp_v6_check(len, saddr, daddr, 0);
+	} else {
+		__wsum csum;
+
+		BUG_ON(skb->ip_summed == CHECKSUM_PARTIAL);
+
+		uh->check = 0;
+		csum = skb_checksum(skb, 0, len, 0);
+		uh->check = udp_v6_check(len, saddr, daddr, csum);
+		if (uh->check == 0)
+			uh->check = CSUM_MANGLED_0;
+
+		skb->ip_summed = CHECKSUM_UNNECESSARY;
+	}
+}
+EXPORT_SYMBOL(udp6_set_csum);
diff --git a/net/ipv6/ip6_offload.c b/net/ipv6/ip6_offload.c
index b2f091566f88..65eda2a8af48 100644
--- a/net/ipv6/ip6_offload.c
+++ b/net/ipv6/ip6_offload.c
@@ -97,9 +97,11 @@ static struct sk_buff *ipv6_gso_segment(struct sk_buff *skb,
 		       SKB_GSO_DODGY |
 		       SKB_GSO_TCP_ECN |
 		       SKB_GSO_GRE |
+		       SKB_GSO_GRE_CSUM |
 		       SKB_GSO_IPIP |
 		       SKB_GSO_SIT |
 		       SKB_GSO_UDP_TUNNEL |
+		       SKB_GSO_UDP_TUNNEL_CSUM |
 		       SKB_GSO_MPLS |
 		       SKB_GSO_TCPV6 |
 		       0)))
diff --git a/net/ipv6/udp_offload.c b/net/ipv6/udp_offload.c
index b261ee8b83fc..0ae3d98f83e0 100644
--- a/net/ipv6/udp_offload.c
+++ b/net/ipv6/udp_offload.c
@@ -63,7 +63,9 @@ static struct sk_buff *udp6_ufo_fragment(struct sk_buff *skb,
 		if (unlikely(type & ~(SKB_GSO_UDP |
 				      SKB_GSO_DODGY |
 				      SKB_GSO_UDP_TUNNEL |
+				      SKB_GSO_UDP_TUNNEL_CSUM |
 				      SKB_GSO_GRE |
+				      SKB_GSO_GRE_CSUM |
 				      SKB_GSO_IPIP |
 				      SKB_GSO_SIT |
 				      SKB_GSO_MPLS) ||
@@ -76,7 +78,8 @@ static struct sk_buff *udp6_ufo_fragment(struct sk_buff *skb,
 		goto out;
 	}
 
-	if (skb->encapsulation && skb_shinfo(skb)->gso_type & SKB_GSO_UDP_TUNNEL)
+	if (skb->encapsulation && skb_shinfo(skb)->gso_type &
+	    (SKB_GSO_UDP_TUNNEL|SKB_GSO_UDP_TUNNEL_CSUM))
 		segs = skb_udp_tunnel_segment(skb, features);
 	else {
 		/* Do software UFO. Complete and fill in the UDP checksum as HW cannot
diff --git a/net/l2tp/l2tp_core.c b/net/l2tp/l2tp_core.c
index 379558014b60..bea259043205 100644
--- a/net/l2tp/l2tp_core.c
+++ b/net/l2tp/l2tp_core.c
@@ -1095,33 +1095,6 @@ static int l2tp_xmit_core(struct l2tp_session *session, struct sk_buff *skb,
 	return 0;
 }
 
-#if IS_ENABLED(CONFIG_IPV6)
-static void l2tp_xmit_ipv6_csum(struct sock *sk, struct sk_buff *skb,
-				int udp_len)
-{
-	struct ipv6_pinfo *np = inet6_sk(sk);
-	struct udphdr *uh = udp_hdr(skb);
-
-	if (udp_get_no_check6_tx(sk))
-		skb->ip_summed = CHECKSUM_NONE;
-	else if (!skb_dst(skb) || !skb_dst(skb)->dev ||
-	    !(skb_dst(skb)->dev->features & NETIF_F_IPV6_CSUM)) {
-		__wsum csum = skb_checksum(skb, 0, udp_len, 0);
-		skb->ip_summed = CHECKSUM_UNNECESSARY;
-		uh->check = csum_ipv6_magic(&np->saddr, &sk->sk_v6_daddr, udp_len,
-					    IPPROTO_UDP, csum);
-		if (uh->check == 0)
-			uh->check = CSUM_MANGLED_0;
-	} else {
-		skb->ip_summed = CHECKSUM_PARTIAL;
-		skb->csum_start = skb_transport_header(skb) - skb->head;
-		skb->csum_offset = offsetof(struct udphdr, check);
-		uh->check = ~csum_ipv6_magic(&np->saddr, &sk->sk_v6_daddr,
-					     udp_len, IPPROTO_UDP, 0);
-	}
-}
-#endif
-
 /* If caller requires the skb to have a ppp header, the header must be
  * inserted in the skb data before calling this function.
  */
@@ -1133,7 +1106,6 @@ int l2tp_xmit_skb(struct l2tp_session *session, struct sk_buff *skb, int hdr_len
 	struct flowi *fl;
 	struct udphdr *uh;
 	struct inet_sock *inet;
-	__wsum csum;
 	int headroom;
 	int uhlen = (tunnel->encap == L2TP_ENCAPTYPE_UDP) ? sizeof(struct udphdr) : 0;
 	int udp_len;
@@ -1182,33 +1154,17 @@ int l2tp_xmit_skb(struct l2tp_session *session, struct sk_buff *skb, int hdr_len
 		uh->dest = inet->inet_dport;
 		udp_len = uhlen + hdr_len + data_len;
 		uh->len = htons(udp_len);
-		uh->check = 0;
 
 		/* Calculate UDP checksum if configured to do so */
 #if IS_ENABLED(CONFIG_IPV6)
 		if (sk->sk_family == PF_INET6 && !tunnel->v4mapped)
-			l2tp_xmit_ipv6_csum(sk, skb, udp_len);
+			udp6_set_csum(udp_get_no_check6_tx(sk),
+				      skb, &inet6_sk(sk)->saddr,
+				      &sk->sk_v6_daddr, udp_len);
 		else
 #endif
-		if (sk->sk_no_check_tx)
-			skb->ip_summed = CHECKSUM_NONE;
-		else if ((skb_dst(skb) && skb_dst(skb)->dev) &&
-			 (!(skb_dst(skb)->dev->features & NETIF_F_V4_CSUM))) {
-			skb->ip_summed = CHECKSUM_COMPLETE;
-			csum = skb_checksum(skb, 0, udp_len, 0);
-			uh->check = csum_tcpudp_magic(inet->inet_saddr,
-						      inet->inet_daddr,
-						      udp_len, IPPROTO_UDP, csum);
-			if (uh->check == 0)
-				uh->check = CSUM_MANGLED_0;
-		} else {
-			skb->ip_summed = CHECKSUM_PARTIAL;
-			skb->csum_start = skb_transport_header(skb) - skb->head;
-			skb->csum_offset = offsetof(struct udphdr, check);
-			uh->check = ~csum_tcpudp_magic(inet->inet_saddr,
-						       inet->inet_daddr,
-						       udp_len, IPPROTO_UDP, 0);
-		}
+		udp_set_csum(sk->sk_no_check_tx, skb, inet->inet_saddr,
+			     inet->inet_daddr, udp_len);
 		break;
 
 	case L2TP_ENCAPTYPE_IP:
diff --git a/net/mpls/mpls_gso.c b/net/mpls/mpls_gso.c
index 851cd880b0c0..6b38d083e1c9 100644
--- a/net/mpls/mpls_gso.c
+++ b/net/mpls/mpls_gso.c
@@ -33,6 +33,7 @@ static struct sk_buff *mpls_gso_segment(struct sk_buff *skb,
 				  SKB_GSO_DODGY |
 				  SKB_GSO_TCP_ECN |
 				  SKB_GSO_GRE |
+				  SKB_GSO_GRE_CSUM |
 				  SKB_GSO_IPIP |
 				  SKB_GSO_MPLS)))
 		goto out;
diff --git a/net/openvswitch/vport-vxlan.c b/net/openvswitch/vport-vxlan.c
index a93efa3f64c3..0edbd95c60e7 100644
--- a/net/openvswitch/vport-vxlan.c
+++ b/net/openvswitch/vport-vxlan.c
@@ -122,7 +122,7 @@ static struct vport *vxlan_tnl_create(const struct vport_parms *parms)
 	vxlan_port = vxlan_vport(vport);
 	strncpy(vxlan_port->name, parms->name, IFNAMSIZ);
 
-	vs = vxlan_sock_add(net, htons(dst_port), vxlan_rcv, vport, true, false);
+	vs = vxlan_sock_add(net, htons(dst_port), vxlan_rcv, vport, true, 0);
 	if (IS_ERR(vs)) {
 		ovs_vport_free(vport);
 		return (void *)vs;