pfcp: always set pfcp metadata

In PFCP receive path set metadata needed by flower code to do correct
classification based on this metadata.

Signed-off-by: Michal Swiatkowski <[e-mail]>
Signed-off-by: Marcin Szycik <[e-mail]>
Reviewed-by: Simon Horman <[e-mail]>
Signed-off-by: Alexander Lobakin <[e-mail]>
Signed-off-by: David S. Miller <[e-mail]>

drivers/net/pfcp.c

···
 	struct socket		*sock;
 	struct net_device	*dev;
 	struct net		*net;
+
+	struct gro_cells	gro_cells;
 };
 
 static unsigned int pfcp_net_id __read_mostly;
···
 struct pfcp_net {
 	struct list_head	pfcp_dev_list;
 };
+
+static void
+pfcp_session_recv(struct pfcp_dev *pfcp, struct sk_buff *skb,
+		  struct pfcp_metadata *md)
+{
+	struct pfcphdr_session *unparsed = pfcp_hdr_session(skb);
+
+	md->seid = unparsed->seid;
+	md->type = PFCP_TYPE_SESSION;
+}
+
+static void
+pfcp_node_recv(struct pfcp_dev *pfcp, struct sk_buff *skb,
+	       struct pfcp_metadata *md)
+{
+	md->type = PFCP_TYPE_NODE;
+}
+
+static int pfcp_encap_recv(struct sock *sk, struct sk_buff *skb)
+{
+	IP_TUNNEL_DECLARE_FLAGS(flags) = { };
+	struct metadata_dst *tun_dst;
+	struct pfcp_metadata *md;
+	struct pfcphdr *unparsed;
+	struct pfcp_dev *pfcp;
+
+	if (unlikely(!pskb_may_pull(skb, PFCP_HLEN)))
+		goto drop;
+
+	pfcp = rcu_dereference_sk_user_data(sk);
+	if (unlikely(!pfcp))
+		goto drop;
+
+	unparsed = pfcp_hdr(skb);
+
+	ip_tunnel_flags_zero(flags);
+	tun_dst = udp_tun_rx_dst(skb, sk->sk_family, flags, 0,
+				 sizeof(*md));
+	if (unlikely(!tun_dst))
+		goto drop;
+
+	md = ip_tunnel_info_opts(&tun_dst->u.tun_info);
+	if (unlikely(!md))
+		goto drop;
+
+	if (unparsed->flags & PFCP_SEID_FLAG)
+		pfcp_session_recv(pfcp, skb, md);
+	else
+		pfcp_node_recv(pfcp, skb, md);
+
+	__set_bit(IP_TUNNEL_PFCP_OPT_BIT, flags);
+	ip_tunnel_info_opts_set(&tun_dst->u.tun_info, md, sizeof(*md),
+				flags);
+
+	if (unlikely(iptunnel_pull_header(skb, PFCP_HLEN, skb->protocol,
+					  !net_eq(sock_net(sk),
+					  dev_net(pfcp->dev)))))
+		goto drop;
+
+	skb_dst_set(skb, (struct dst_entry *)tun_dst);
+
+	skb_reset_network_header(skb);
+	skb_reset_mac_header(skb);
+	skb->dev = pfcp->dev;
+
+	gro_cells_receive(&pfcp->gro_cells, skb);
+
+	return 0;
+drop:
+	kfree_skb(skb);
+	return 0;
+}
 
 static void pfcp_del_sock(struct pfcp_dev *pfcp)
 {
···
 {
 	struct pfcp_dev *pfcp = netdev_priv(dev);
 
+	gro_cells_destroy(&pfcp->gro_cells);
 	pfcp_del_sock(pfcp);
 }
 
···
 
 	pfcp->dev = dev;
 
-	return 0;
+	return gro_cells_init(&pfcp->gro_cells, dev);
 }
 
 static const struct net_device_ops pfcp_netdev_ops = {
···
 	err = udp_sock_create(net, &udp_conf, &sock);
 	if (err)
 		return ERR_PTR(err);
+
+	tuncfg.sk_user_data = pfcp;
+	tuncfg.encap_rcv = pfcp_encap_recv;
+	tuncfg.encap_type = 1;
 
 	setup_udp_tunnel_sock(net, sock, &tuncfg);
 

include/net/ip_tunnels.h

···
 	__set_bit(IP_TUNNEL_VXLAN_OPT_BIT, present);
 	__set_bit(IP_TUNNEL_ERSPAN_OPT_BIT, present);
 	__set_bit(IP_TUNNEL_GTP_OPT_BIT, present);
+	__set_bit(IP_TUNNEL_PFCP_OPT_BIT, present);
 
 	ip_tunnel_flags_or(flags, flags, present);
 }
···
 	__set_bit(IP_TUNNEL_VXLAN_OPT_BIT, present);
 	__set_bit(IP_TUNNEL_ERSPAN_OPT_BIT, present);
 	__set_bit(IP_TUNNEL_GTP_OPT_BIT, present);
+	__set_bit(IP_TUNNEL_PFCP_OPT_BIT, present);
 
 	__ipt_flag_op(bitmap_andnot, flags, flags, present);
 }
···
 	__set_bit(IP_TUNNEL_VXLAN_OPT_BIT, present);
 	__set_bit(IP_TUNNEL_ERSPAN_OPT_BIT, present);
 	__set_bit(IP_TUNNEL_GTP_OPT_BIT, present);
+	__set_bit(IP_TUNNEL_PFCP_OPT_BIT, present);
 
 	return ip_tunnel_flags_intersect(flags, present);
 }

include/net/pfcp.h

···
 #ifndef _PFCP_H_
 #define _PFCP_H_
 
+#include <uapi/linux/if_ether.h>
+#include <net/dst_metadata.h>
 #include <linux/netdevice.h>
+#include <uapi/linux/ipv6.h>
+#include <net/udp_tunnel.h>
+#include <uapi/linux/udp.h>
+#include <uapi/linux/ip.h>
 #include <linux/string.h>
 #include <linux/types.h>
+#include <linux/bits.h>
 
 #define PFCP_PORT 8805
+
+/* PFCP protocol header */
+struct pfcphdr {
+	u8	flags;
+	u8	message_type;
+	__be16	message_length;
+};
+
+/* PFCP header flags */
+#define PFCP_SEID_FLAG		BIT(0)
+#define PFCP_MP_FLAG		BIT(1)
+
+#define PFCP_VERSION_MASK	GENMASK(4, 0)
+
+#define PFCP_HLEN (sizeof(struct udphdr) + sizeof(struct pfcphdr))
+
+/* PFCP node related messages */
+struct pfcphdr_node {
+	u8	seq_number[3];
+	u8	reserved;
+};
+
+/* PFCP session related messages */
+struct pfcphdr_session {
+	__be64	seid;
+	u8	seq_number[3];
+#ifdef __LITTLE_ENDIAN_BITFIELD
+	u8	message_priority:4,
+		reserved:4;
+#elif defined(__BIG_ENDIAN_BITFIELD)
+	u8	reserved:4,
+		message_priprity:4;
+#else
+#error "Please fix <asm/byteorder>"
+#endif
+};
+
+struct pfcp_metadata {
+	u8 type;
+	__be64 seid;
+} __packed;
+
+enum {
+	PFCP_TYPE_NODE		= 0,
+	PFCP_TYPE_SESSION	= 1,
+};
+
+#define PFCP_HEADROOM (sizeof(struct iphdr) + sizeof(struct udphdr) + \
+		       sizeof(struct pfcphdr) + sizeof(struct ethhdr))
+#define PFCP6_HEADROOM (sizeof(struct ipv6hdr) + sizeof(struct udphdr) + \
+			sizeof(struct pfcphdr) + sizeof(struct ethhdr))
+
+static inline struct pfcphdr *pfcp_hdr(struct sk_buff *skb)
+{
+	return (struct pfcphdr *)(udp_hdr(skb) + 1);
+}
+
+static inline struct pfcphdr_node *pfcp_hdr_node(struct sk_buff *skb)
+{
+	return (struct pfcphdr_node *)(pfcp_hdr(skb) + 1);
+}
+
+static inline struct pfcphdr_session *pfcp_hdr_session(struct sk_buff *skb)
+{
+	return (struct pfcphdr_session *)(pfcp_hdr(skb) + 1);
+}
 
 static inline bool netif_is_pfcp(const struct net_device *dev)
 {

include/uapi/linux/if_tunnel.h

···
 	IP_TUNNEL_VTI_BIT,
 	IP_TUNNEL_SIT_ISATAP_BIT	= IP_TUNNEL_VTI_BIT,
 
+	/* Flags starting from here are not available via the old UAPI */
+	IP_TUNNEL_PFCP_OPT_BIT,		/* OPTIONS_PRESENT */
+
 	__IP_TUNNEL_FLAG_NUM,
 };
 

include/uapi/linux/pkt_cls.h

···
 					 * TCA_FLOWER_KEY_ENC_OPT_GTP_
 					 * attributes
 					 */
+	TCA_FLOWER_KEY_ENC_OPTS_PFCP,	/* Nested
+					 * TCA_FLOWER_KEY_ENC_IPT_PFCP
+					 * attributes
+					 */
 	__TCA_FLOWER_KEY_ENC_OPTS_MAX,
 };
 
···
 
 #define TCA_FLOWER_KEY_ENC_OPT_GTP_MAX \
 		(__TCA_FLOWER_KEY_ENC_OPT_GTP_MAX - 1)
+
+enum {
+	TCA_FLOWER_KEY_ENC_OPT_PFCP_UNSPEC,
+	TCA_FLOWER_KEY_ENC_OPT_PFCP_TYPE,		/* u8 */
+	TCA_FLOWER_KEY_ENC_OPT_PFCP_SEID,		/* be64 */
+	__TCA_FLOWER_KEY_ENC_OPT_PFCP_MAX,
+};
+
+#define TCA_FLOWER_KEY_ENC_OPT_PFCP_MAX \
+		(__TCA_FLOWER_KEY_ENC_OPT_PFCP_MAX - 1)
 
 enum {
 	TCA_FLOWER_KEY_MPLS_OPTS_UNSPEC,

net/core/net_test.c

···
 			     ip_tunnel_flags_1),
 	IP_TUNNEL_FLAGS_TEST("conflict", ip_tunnel_flags_2_src, true,
 			     VTI_ISVTI, ip_tunnel_flags_2_exp),
-	IP_TUNNEL_FLAGS_TEST("new", ip_tunnel_flags_3_src,
-			     /* This must be set to ``false`` once
-			      * ``__IP_TUNNEL_FLAG_NUM`` goes above 17.
-			      */
-			     true, cpu_to_be16(BIT(IP_TUNNEL_VXLAN_OPT_BIT)),
+	IP_TUNNEL_FLAGS_TEST("new", ip_tunnel_flags_3_src, false,
+			     cpu_to_be16(BIT(IP_TUNNEL_VXLAN_OPT_BIT)),
 			     ip_tunnel_flags_3_exp),
 };
 

net/sched/cls_flower.c

···
 #include <net/vxlan.h>
 #include <net/erspan.h>
 #include <net/gtp.h>
+#include <net/pfcp.h>
 #include <net/tc_wrapper.h>
 
 #include <net/dst.h>
···
 	[TCA_FLOWER_KEY_ENC_OPTS_VXLAN]         = { .type = NLA_NESTED },
 	[TCA_FLOWER_KEY_ENC_OPTS_ERSPAN]        = { .type = NLA_NESTED },
 	[TCA_FLOWER_KEY_ENC_OPTS_GTP]		= { .type = NLA_NESTED },
+	[TCA_FLOWER_KEY_ENC_OPTS_PFCP]		= { .type = NLA_NESTED },
 };
 
 static const struct nla_policy
···
 gtp_opt_policy[TCA_FLOWER_KEY_ENC_OPT_GTP_MAX + 1] = {
 	[TCA_FLOWER_KEY_ENC_OPT_GTP_PDU_TYPE]	   = { .type = NLA_U8 },
 	[TCA_FLOWER_KEY_ENC_OPT_GTP_QFI]	   = { .type = NLA_U8 },
+};
+
+static const struct nla_policy
+pfcp_opt_policy[TCA_FLOWER_KEY_ENC_OPT_PFCP_MAX + 1] = {
+	[TCA_FLOWER_KEY_ENC_OPT_PFCP_TYPE]	   = { .type = NLA_U8 },
+	[TCA_FLOWER_KEY_ENC_OPT_PFCP_SEID]	   = { .type = NLA_U64 },
 };
 
 static const struct nla_policy
···
 	return sizeof(*sinfo);
 }
 
+static int fl_set_pfcp_opt(const struct nlattr *nla, struct fl_flow_key *key,
+			   int depth, int option_len,
+			   struct netlink_ext_ack *extack)
+{
+	struct nlattr *tb[TCA_FLOWER_KEY_ENC_OPT_PFCP_MAX + 1];
+	struct pfcp_metadata *md;
+	int err;
+
+	md = (struct pfcp_metadata *)&key->enc_opts.data[key->enc_opts.len];
+	memset(md, 0xff, sizeof(*md));
+
+	if (!depth)
+		return sizeof(*md);
+
+	if (nla_type(nla) != TCA_FLOWER_KEY_ENC_OPTS_PFCP) {
+		NL_SET_ERR_MSG_MOD(extack, "Non-pfcp option type for mask");
+		return -EINVAL;
+	}
+
+	err = nla_parse_nested(tb, TCA_FLOWER_KEY_ENC_OPT_PFCP_MAX, nla,
+			       pfcp_opt_policy, extack);
+	if (err < 0)
+		return err;
+
+	if (!option_len && !tb[TCA_FLOWER_KEY_ENC_OPT_PFCP_TYPE]) {
+		NL_SET_ERR_MSG_MOD(extack, "Missing tunnel key pfcp option type");
+		return -EINVAL;
+	}
+
+	if (tb[TCA_FLOWER_KEY_ENC_OPT_PFCP_TYPE])
+		md->type = nla_get_u8(tb[TCA_FLOWER_KEY_ENC_OPT_PFCP_TYPE]);
+
+	if (tb[TCA_FLOWER_KEY_ENC_OPT_PFCP_SEID])
+		md->seid = nla_get_be64(tb[TCA_FLOWER_KEY_ENC_OPT_PFCP_SEID]);
+
+	return sizeof(*md);
+}
+
 static int fl_set_enc_opt(struct nlattr **tb, struct fl_flow_key *key,
 			  struct fl_flow_key *mask,
 			  struct netlink_ext_ack *extack)
···
 			if (key->enc_opts.len != mask->enc_opts.len) {
 				NL_SET_ERR_MSG_MOD(extack,
 						   "Key and mask miss aligned");
+				return -EINVAL;
+			}
+			break;
+		case TCA_FLOWER_KEY_ENC_OPTS_PFCP:
+			if (key->enc_opts.dst_opt_type) {
+				NL_SET_ERR_MSG_MOD(extack, "Duplicate type for pfcp options");
+				return -EINVAL;
+			}
+			option_len = 0;
+			key->enc_opts.dst_opt_type = IP_TUNNEL_PFCP_OPT_BIT;
+			option_len = fl_set_pfcp_opt(nla_opt_key, key,
+						     key_depth, option_len,
+						     extack);
+			if (option_len < 0)
+				return option_len;
+
+			key->enc_opts.len += option_len;
+			/* At the same time we need to parse through the mask
+			 * in order to verify exact and mask attribute lengths.
+			 */
+			mask->enc_opts.dst_opt_type = IP_TUNNEL_PFCP_OPT_BIT;
+			option_len = fl_set_pfcp_opt(nla_opt_msk, mask,
+						     msk_depth, option_len,
+						     extack);
+			if (option_len < 0)
+				return option_len;
+
+			mask->enc_opts.len += option_len;
+			if (key->enc_opts.len != mask->enc_opts.len) {
+				NL_SET_ERR_MSG_MOD(extack, "Key and mask miss aligned");
 				return -EINVAL;
 			}
 			break;
···
 	return -EMSGSIZE;
 }
 
+static int fl_dump_key_pfcp_opt(struct sk_buff *skb,
+				struct flow_dissector_key_enc_opts *enc_opts)
+{
+	struct pfcp_metadata *md;
+	struct nlattr *nest;
+
+	nest = nla_nest_start_noflag(skb, TCA_FLOWER_KEY_ENC_OPTS_PFCP);
+	if (!nest)
+		goto nla_put_failure;
+
+	md = (struct pfcp_metadata *)&enc_opts->data[0];
+	if (nla_put_u8(skb, TCA_FLOWER_KEY_ENC_OPT_PFCP_TYPE, md->type))
+		goto nla_put_failure;
+
+	if (nla_put_be64(skb, TCA_FLOWER_KEY_ENC_OPT_PFCP_SEID,
+			 md->seid, 0))
+		goto nla_put_failure;
+
+	nla_nest_end(skb, nest);
+	return 0;
+
+nla_put_failure:
+	nla_nest_cancel(skb, nest);
+	return -EMSGSIZE;
+}
+
 static int fl_dump_key_ct(struct sk_buff *skb,
 			  struct flow_dissector_key_ct *key,
 			  struct flow_dissector_key_ct *mask)
···
 		break;
 	case IP_TUNNEL_GTP_OPT_BIT:
 		err = fl_dump_key_gtp_opt(skb, enc_opts);
+		if (err)
+			goto nla_put_failure;
+		break;
+	case IP_TUNNEL_PFCP_OPT_BIT:
+		err = fl_dump_key_pfcp_opt(skb, enc_opts);
 		if (err)
 			goto nla_put_failure;
 		break;