Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   Jacky Hu <hengqing.hu@gmail.com>
To:     hengqing.hu@gmail.com
Cc:     jacky.hu@walmart.com, jason.niesz@walmart.com,
        "David S. Miller" <davem@davemloft.net>,
        Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>,
        Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>,
        Wensong Zhang <wensong@linux-vs.org>,
        Simon Horman <horms@verge.net.au>,
        Julian Anastasov <ja@ssi.bg>,
        Pablo Neira Ayuso <pablo@netfilter.org>,
        Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>,
        Florian Westphal <fw@strlen.de>, netdev@vger.kernel.org,
        lvs-devel@vger.kernel.org, linux-kernel@vger.kernel.org,
        netfilter-devel@vger.kernel.org, coreteam@netfilter.org
Subject: [PATCH v7] ipvs: allow tunneling with gue encapsulation
Date:   Tue, 26 Mar 2019 18:31:21 +0800
Message-Id: <20190326103121.31083-1-hengqing.hu@gmail.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

ipip packets are blocked in some public cloud environments, this patch
allows gue encapsulation with the tunneling method, which would make
tunneling working in those environments.

Signed-off-by: Jacky Hu <hengqing.hu@gmail.com>
---
 include/net/ip_vs.h             |  5 ++
 include/uapi/linux/ip_vs.h      | 11 +++++
 net/netfilter/ipvs/ip_vs_ctl.c  | 35 +++++++++++++-
 net/netfilter/ipvs/ip_vs_xmit.c | 84 +++++++++++++++++++++++++++++++--
 4 files changed, 130 insertions(+), 5 deletions(-)

diff --git a/include/net/ip_vs.h b/include/net/ip_vs.h
index 047f9a5ccaad..2ac40135b576 100644
--- a/include/net/ip_vs.h
+++ b/include/net/ip_vs.h
@@ -600,6 +600,9 @@ struct ip_vs_dest_user_kern {
 
 	/* Address family of addr */
 	u16			af;
+
+	u16			tun_type;	/* tunnel type */
+	__be16			tun_port;	/* tunnel port */
 };
 
 
@@ -660,6 +663,8 @@ struct ip_vs_dest {
 	atomic_t		conn_flags;	/* flags to copy to conn */
 	atomic_t		weight;		/* server weight */
 	atomic_t		last_weight;	/* server latest weight */
+	__u16			tun_type;	/* tunnel type */
+	__be16			tun_port;	/* tunnel port */
 
 	refcount_t		refcnt;		/* reference counter */
 	struct ip_vs_stats      stats;          /* statistics */
diff --git a/include/uapi/linux/ip_vs.h b/include/uapi/linux/ip_vs.h
index 1c916b2f89dc..e34f436fc79d 100644
--- a/include/uapi/linux/ip_vs.h
+++ b/include/uapi/linux/ip_vs.h
@@ -124,6 +124,13 @@
 
 #define IP_VS_PEDATA_MAXLEN     255
 
+/* Tunnel types */
+enum {
+	IP_VS_CONN_F_TUNNEL_TYPE_IPIP = 0,	/* IPIP */
+	IP_VS_CONN_F_TUNNEL_TYPE_GUE,		/* GUE */
+	IP_VS_CONN_F_TUNNEL_TYPE_MAX,
+};
+
 /*
  *	The struct ip_vs_service_user and struct ip_vs_dest_user are
  *	used to set IPVS rules through setsockopt.
@@ -392,6 +399,10 @@ enum {
 
 	IPVS_DEST_ATTR_STATS64,		/* nested attribute for dest stats */
 
+	IPVS_DEST_ATTR_TUN_TYPE,	/* tunnel type */
+
+	IPVS_DEST_ATTR_TUN_PORT,	/* tunnel port */
+
 	__IPVS_DEST_ATTR_MAX,
 };
 
diff --git a/net/netfilter/ipvs/ip_vs_ctl.c b/net/netfilter/ipvs/ip_vs_ctl.c
index 053cd96b9c76..328683452229 100644
--- a/net/netfilter/ipvs/ip_vs_ctl.c
+++ b/net/netfilter/ipvs/ip_vs_ctl.c
@@ -831,6 +831,10 @@ __ip_vs_update_dest(struct ip_vs_service *svc, struct ip_vs_dest *dest,
 	conn_flags = udest->conn_flags & IP_VS_CONN_F_DEST_MASK;
 	conn_flags |= IP_VS_CONN_F_INACTIVE;
 
+	/* set the tunnel info */
+	dest->tun_type = udest->tun_type;
+	dest->tun_port = udest->tun_port;
+
 	/* set the IP_VS_CONN_F_NOOUTPUT flag if not masquerading/NAT */
 	if ((conn_flags & IP_VS_CONN_F_FWD_MASK) != IP_VS_CONN_F_MASQ) {
 		conn_flags |= IP_VS_CONN_F_NOOUTPUT;
@@ -987,6 +991,13 @@ ip_vs_add_dest(struct ip_vs_service *svc, struct ip_vs_dest_user_kern *udest)
 		return -ERANGE;
 	}
 
+	if (udest->tun_type == IP_VS_CONN_F_TUNNEL_TYPE_GUE) {
+		if (udest->tun_port == 0) {
+			pr_err("%s(): tunnel port is zero\n", __func__);
+			return -EINVAL;
+		}
+	}
+
 	ip_vs_addr_copy(udest->af, &daddr, &udest->addr);
 
 	/* We use function that requires RCU lock */
@@ -1051,6 +1062,13 @@ ip_vs_edit_dest(struct ip_vs_service *svc, struct ip_vs_dest_user_kern *udest)
 		return -ERANGE;
 	}
 
+	if (udest->tun_type == IP_VS_CONN_F_TUNNEL_TYPE_GUE) {
+		if (udest->tun_port == 0) {
+			pr_err("%s(): tunnel port is zero\n", __func__);
+			return -EINVAL;
+		}
+	}
+
 	ip_vs_addr_copy(udest->af, &daddr, &udest->addr);
 
 	/* We use function that requires RCU lock */
@@ -2333,6 +2351,7 @@ static void ip_vs_copy_udest_compat(struct ip_vs_dest_user_kern *udest,
 	udest->u_threshold	= udest_compat->u_threshold;
 	udest->l_threshold	= udest_compat->l_threshold;
 	udest->af		= AF_INET;
+	udest->tun_type		= IP_VS_CONN_F_TUNNEL_TYPE_IPIP;
 }
 
 static int
@@ -2890,6 +2909,8 @@ static const struct nla_policy ip_vs_dest_policy[IPVS_DEST_ATTR_MAX + 1] = {
 	[IPVS_DEST_ATTR_PERSIST_CONNS]	= { .type = NLA_U32 },
 	[IPVS_DEST_ATTR_STATS]		= { .type = NLA_NESTED },
 	[IPVS_DEST_ATTR_ADDR_FAMILY]	= { .type = NLA_U16 },
+	[IPVS_DEST_ATTR_TUN_TYPE]	= { .type = NLA_U8 },
+	[IPVS_DEST_ATTR_TUN_PORT]	= { .type = NLA_U16 },
 };
 
 static int ip_vs_genl_fill_stats(struct sk_buff *skb, int container_type,
@@ -3193,6 +3214,10 @@ static int ip_vs_genl_fill_dest(struct sk_buff *skb, struct ip_vs_dest *dest)
 			 IP_VS_CONN_F_FWD_MASK)) ||
 	    nla_put_u32(skb, IPVS_DEST_ATTR_WEIGHT,
 			atomic_read(&dest->weight)) ||
+	    nla_put_u8(skb, IPVS_DEST_ATTR_TUN_TYPE,
+		       dest->tun_type) ||
+	    nla_put_be16(skb, IPVS_DEST_ATTR_TUN_PORT,
+			 dest->tun_port) ||
 	    nla_put_u32(skb, IPVS_DEST_ATTR_U_THRESH, dest->u_threshold) ||
 	    nla_put_u32(skb, IPVS_DEST_ATTR_L_THRESH, dest->l_threshold) ||
 	    nla_put_u32(skb, IPVS_DEST_ATTR_ACTIVE_CONNS,
@@ -3315,12 +3340,14 @@ static int ip_vs_genl_parse_dest(struct ip_vs_dest_user_kern *udest,
 	/* If a full entry was requested, check for the additional fields */
 	if (full_entry) {
 		struct nlattr *nla_fwd, *nla_weight, *nla_u_thresh,
-			      *nla_l_thresh;
+			      *nla_l_thresh, *nla_tun_type, *nla_tun_port;
 
 		nla_fwd		= attrs[IPVS_DEST_ATTR_FWD_METHOD];
 		nla_weight	= attrs[IPVS_DEST_ATTR_WEIGHT];
 		nla_u_thresh	= attrs[IPVS_DEST_ATTR_U_THRESH];
 		nla_l_thresh	= attrs[IPVS_DEST_ATTR_L_THRESH];
+		nla_tun_type	= attrs[IPVS_DEST_ATTR_TUN_TYPE];
+		nla_tun_port	= attrs[IPVS_DEST_ATTR_TUN_PORT];
 
 		if (!(nla_fwd && nla_weight && nla_u_thresh && nla_l_thresh))
 			return -EINVAL;
@@ -3330,6 +3357,12 @@ static int ip_vs_genl_parse_dest(struct ip_vs_dest_user_kern *udest,
 		udest->weight = nla_get_u32(nla_weight);
 		udest->u_threshold = nla_get_u32(nla_u_thresh);
 		udest->l_threshold = nla_get_u32(nla_l_thresh);
+
+		if (nla_tun_type)
+			udest->tun_type = nla_get_u8(nla_tun_type);
+
+		if (nla_tun_port)
+			udest->tun_port = nla_get_be16(nla_tun_port);
 	}
 
 	return 0;
diff --git a/net/netfilter/ipvs/ip_vs_xmit.c b/net/netfilter/ipvs/ip_vs_xmit.c
index 175349fcf91f..8d6f94b67772 100644
--- a/net/netfilter/ipvs/ip_vs_xmit.c
+++ b/net/netfilter/ipvs/ip_vs_xmit.c
@@ -32,6 +32,7 @@
 #include <linux/slab.h>
 #include <linux/tcp.h>                  /* for tcphdr */
 #include <net/ip.h>
+#include <net/gue.h>
 #include <net/tcp.h>                    /* for csum_tcpudp_magic */
 #include <net/udp.h>
 #include <net/icmp.h>                   /* for icmp_send */
@@ -382,6 +383,10 @@ __ip_vs_get_out_rt(struct netns_ipvs *ipvs, int skb_af, struct sk_buff *skb,
 		mtu = dst_mtu(&rt->dst);
 	} else {
 		mtu = dst_mtu(&rt->dst) - sizeof(struct iphdr);
+		if (!dest)
+			goto err_put;
+		if (dest->tun_type == IP_VS_CONN_F_TUNNEL_TYPE_GUE)
+			mtu -= sizeof(struct udphdr) + sizeof(struct guehdr);
 		if (mtu < 68) {
 			IP_VS_DBG_RL("%s(): mtu less than 68\n", __func__);
 			goto err_put;
@@ -533,6 +538,10 @@ __ip_vs_get_out_rt_v6(struct netns_ipvs *ipvs, int skb_af, struct sk_buff *skb,
 		mtu = dst_mtu(&rt->dst);
 	else {
 		mtu = dst_mtu(&rt->dst) - sizeof(struct ipv6hdr);
+		if (!dest)
+			goto err_put;
+		if (dest->tun_type == IP_VS_CONN_F_TUNNEL_TYPE_GUE)
+			mtu -= sizeof(struct udphdr) + sizeof(struct guehdr);
 		if (mtu < IPV6_MIN_MTU) {
 			IP_VS_DBG_RL("%s(): mtu less than %d\n", __func__,
 				     IPV6_MIN_MTU);
@@ -989,6 +998,41 @@ static inline int __tun_gso_type_mask(int encaps_af, int orig_af)
 	}
 }
 
+static int
+ipvs_gue_encap(struct net *net, struct sk_buff *skb,
+	       struct ip_vs_conn *cp, __u8 *next_protocol)
+{
+	__be16 dport;
+	__be16 sport = udp_flow_src_port(net, skb, 0, 0, false);
+	struct udphdr  *udph;	/* Our new UDP header */
+	struct guehdr  *gueh;	/* Our new GUE header */
+
+	skb_push(skb, sizeof(struct guehdr));
+
+	gueh = (struct guehdr *)skb->data;
+
+	gueh->control = 0;
+	gueh->version = 0;
+	gueh->hlen = 0;
+	gueh->flags = 0;
+	gueh->proto_ctype = *next_protocol;
+
+	skb_push(skb, sizeof(struct udphdr));
+	skb_reset_transport_header(skb);
+
+	udph = udp_hdr(skb);
+
+	dport = cp->dest->tun_port;
+	udph->dest = dport;
+	udph->source = sport;
+	udph->len = htons(skb->len);
+	udph->check = 0;
+
+	*next_protocol = IPPROTO_UDP;
+
+	return 0;
+}
+
 /*
  *   IP Tunneling transmitter
  *
@@ -1025,6 +1069,7 @@ ip_vs_tunnel_xmit(struct sk_buff *skb, struct ip_vs_conn *cp,
 	struct iphdr  *iph;			/* Our new IP header */
 	unsigned int max_headroom;		/* The extra header space needed */
 	int ret, local;
+	int tun_type, gso_type;
 
 	EnterFunction(10);
 
@@ -1046,6 +1091,11 @@ ip_vs_tunnel_xmit(struct sk_buff *skb, struct ip_vs_conn *cp,
 	 */
 	max_headroom = LL_RESERVED_SPACE(tdev) + sizeof(struct iphdr);
 
+	tun_type = cp->dest->tun_type;
+
+	if (tun_type == IP_VS_CONN_F_TUNNEL_TYPE_GUE)
+		max_headroom += sizeof(struct udphdr) + sizeof(struct guehdr);
+
 	/* We only care about the df field if sysctl_pmtu_disc(ipvs) is set */
 	dfp = sysctl_pmtu_disc(ipvs) ? &df : NULL;
 	skb = ip_vs_prepare_tunneled_skb(skb, cp->af, max_headroom,
@@ -1054,11 +1104,20 @@ ip_vs_tunnel_xmit(struct sk_buff *skb, struct ip_vs_conn *cp,
 	if (IS_ERR(skb))
 		goto tx_error;
 
-	if (iptunnel_handle_offloads(skb, __tun_gso_type_mask(AF_INET, cp->af)))
+	gso_type = __tun_gso_type_mask(AF_INET, cp->af);
+	if (tun_type == IP_VS_CONN_F_TUNNEL_TYPE_GUE)
+		gso_type |= SKB_GSO_UDP_TUNNEL;
+
+	if (iptunnel_handle_offloads(skb, gso_type))
 		goto tx_error;
 
 	skb->transport_header = skb->network_header;
 
+	skb_set_inner_ipproto(skb, next_protocol);
+
+	if (tun_type == IP_VS_CONN_F_TUNNEL_TYPE_GUE)
+		ipvs_gue_encap(net, skb, cp, &next_protocol);
+
 	skb_push(skb, sizeof(struct iphdr));
 	skb_reset_network_header(skb);
 	memset(&(IPCB(skb)->opt), 0, sizeof(IPCB(skb)->opt));
@@ -1102,6 +1161,8 @@ int
 ip_vs_tunnel_xmit_v6(struct sk_buff *skb, struct ip_vs_conn *cp,
 		     struct ip_vs_protocol *pp, struct ip_vs_iphdr *ipvsh)
 {
+	struct netns_ipvs *ipvs = cp->ipvs;
+	struct net *net = ipvs->net;
 	struct rt6_info *rt;		/* Route to the other host */
 	struct in6_addr saddr;		/* Source for tunnel */
 	struct net_device *tdev;	/* Device to other host */
@@ -1112,10 +1173,11 @@ ip_vs_tunnel_xmit_v6(struct sk_buff *skb, struct ip_vs_conn *cp,
 	struct ipv6hdr  *iph;		/* Our new IP header */
 	unsigned int max_headroom;	/* The extra header space needed */
 	int ret, local;
+	int tun_type, gso_type;
 
 	EnterFunction(10);
 
-	local = __ip_vs_get_out_rt_v6(cp->ipvs, cp->af, skb, cp->dest,
+	local = __ip_vs_get_out_rt_v6(ipvs, cp->af, skb, cp->dest,
 				      &cp->daddr.in6,
 				      &saddr, ipvsh, 1,
 				      IP_VS_RT_MODE_LOCAL |
@@ -1134,17 +1196,31 @@ ip_vs_tunnel_xmit_v6(struct sk_buff *skb, struct ip_vs_conn *cp,
 	 */
 	max_headroom = LL_RESERVED_SPACE(tdev) + sizeof(struct ipv6hdr);
 
+	tun_type = cp->dest->tun_type;
+
+	if (tun_type == IP_VS_CONN_F_TUNNEL_TYPE_GUE)
+		max_headroom += sizeof(struct udphdr) + sizeof(struct guehdr);
+
 	skb = ip_vs_prepare_tunneled_skb(skb, cp->af, max_headroom,
 					 &next_protocol, &payload_len,
 					 &dsfield, &ttl, NULL);
 	if (IS_ERR(skb))
 		goto tx_error;
 
-	if (iptunnel_handle_offloads(skb, __tun_gso_type_mask(AF_INET6, cp->af)))
+	gso_type = __tun_gso_type_mask(AF_INET6, cp->af);
+	if (tun_type == IP_VS_CONN_F_TUNNEL_TYPE_GUE)
+		gso_type |= SKB_GSO_UDP_TUNNEL;
+
+	if (iptunnel_handle_offloads(skb, gso_type))
 		goto tx_error;
 
 	skb->transport_header = skb->network_header;
 
+	skb_set_inner_ipproto(skb, next_protocol);
+
+	if (tun_type == IP_VS_CONN_F_TUNNEL_TYPE_GUE)
+		ipvs_gue_encap(net, skb, cp, &next_protocol);
+
 	skb_push(skb, sizeof(struct ipv6hdr));
 	skb_reset_network_header(skb);
 	memset(&(IPCB(skb)->opt), 0, sizeof(IPCB(skb)->opt));
@@ -1167,7 +1243,7 @@ ip_vs_tunnel_xmit_v6(struct sk_buff *skb, struct ip_vs_conn *cp,
 
 	ret = ip_vs_tunnel_xmit_prepare(skb, cp);
 	if (ret == NF_ACCEPT)
-		ip6_local_out(cp->ipvs->net, skb->sk, skb);
+		ip6_local_out(net, skb->sk, skb);
 	else if (ret == NF_DROP)
 		kfree_skb(skb);
 
-- 
2.21.0