Merge branch 'net-Geneve-options-support-for-TC-act_tunnel_key'

Jakub Kicinski says:

====================
net: Geneve options support for TC act_tunnel_key

Simon & Pieter say:

This set adds Geneve Options support to the TC tunnel key action.
It provides the plumbing required to configure Geneve variable length
options.  The options can be configured in the form CLASS:TYPE:DATA,
where CLASS is represented as a 16bit hexadecimal value, TYPE as an 8bit
hexadecimal value and DATA as a variable length hexadecimal value.
Additionally multiple options may be listed using a comma delimiter.

v2:
 - fix sparse warnings in patches 3 and 4 (first one reported by
   build bot).
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

drivers/net/geneve.c

@@ -236,7 +236,8 @@ static void geneve_rx(struct geneve_dev *geneve, struct geneve_sock *gs,
 		}
 		/* Update tunnel dst according to Geneve options. */
 		ip_tunnel_info_opts_set(&tun_dst->u.tun_info,
-					gnvh->options, gnvh->opt_len * 4);
+					gnvh->options, gnvh->opt_len * 4,
+					TUNNEL_GENEVE_OPT);
 	} else {
 		/* Drop packets w/ critical options,
 		 * since we don't support any...
@@ -675,7 +676,8 @@ static void geneve_build_header(struct genevehdr *geneveh,
 	geneveh->proto_type = htons(ETH_P_TEB);
 	geneveh->rsvd2 = 0;
 
-	ip_tunnel_info_opts_get(geneveh->options, info);
+	if (info->key.tun_flags & TUNNEL_GENEVE_OPT)
+		ip_tunnel_info_opts_get(geneveh->options, info);
 }
 
 static int geneve_build_skb(struct dst_entry *dst, struct sk_buff *skb,

drivers/net/vxlan.c

@@ -2122,7 +2122,8 @@ static void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 		vni = tunnel_id_to_key32(info->key.tun_id);
 		ifindex = 0;
 		dst_cache = &info->dst_cache;
-		if (info->options_len)
+		if (info->options_len &&
+		    info->key.tun_flags & TUNNEL_VXLAN_OPT)
 			md = ip_tunnel_info_opts(info);
 		ttl = info->key.ttl;
 		tos = info->key.tos;

include/net/ip_tunnels.h

@@ -466,10 +466,12 @@ static inline void ip_tunnel_info_opts_get(void *to,
 }
 
 static inline void ip_tunnel_info_opts_set(struct ip_tunnel_info *info,
-					   const void *from, int len)
+					   const void *from, int len,
+					   __be16 flags)
 {
 	memcpy(ip_tunnel_info_opts(info), from, len);
 	info->options_len = len;
+	info->key.tun_flags |= flags;
 }
 
 static inline struct ip_tunnel_info *lwt_tun_info(struct lwtunnel_state *lwtstate)
@@ -511,9 +513,11 @@ static inline void ip_tunnel_info_opts_get(void *to,
 }
 
 static inline void ip_tunnel_info_opts_set(struct ip_tunnel_info *info,
-					   const void *from, int len)
+					   const void *from, int len,
+					   __be16 flags)
 {
 	info->options_len = 0;
+	info->key.tun_flags |= flags;
 }
 
 #endif /* CONFIG_INET */

include/uapi/linux/tc_act/tc_tunnel_key.h

@@ -36,9 +36,35 @@ enum {
 	TCA_TUNNEL_KEY_PAD,
 	TCA_TUNNEL_KEY_ENC_DST_PORT,	/* be16 */
 	TCA_TUNNEL_KEY_NO_CSUM,		/* u8 */
+	TCA_TUNNEL_KEY_ENC_OPTS,	/* Nested TCA_TUNNEL_KEY_ENC_OPTS_
+					 * attributes
+					 */
 	__TCA_TUNNEL_KEY_MAX,
 };
 
 #define TCA_TUNNEL_KEY_MAX (__TCA_TUNNEL_KEY_MAX - 1)
 
+enum {
+	TCA_TUNNEL_KEY_ENC_OPTS_UNSPEC,
+	TCA_TUNNEL_KEY_ENC_OPTS_GENEVE,		/* Nested
+						 * TCA_TUNNEL_KEY_ENC_OPTS_
+						 * attributes
+						 */
+	__TCA_TUNNEL_KEY_ENC_OPTS_MAX,
+};
+
+#define TCA_TUNNEL_KEY_ENC_OPTS_MAX (__TCA_TUNNEL_KEY_ENC_OPTS_MAX - 1)
+
+enum {
+	TCA_TUNNEL_KEY_ENC_OPT_GENEVE_UNSPEC,
+	TCA_TUNNEL_KEY_ENC_OPT_GENEVE_CLASS,		/* be16 */
+	TCA_TUNNEL_KEY_ENC_OPT_GENEVE_TYPE,		/* u8 */
+	TCA_TUNNEL_KEY_ENC_OPT_GENEVE_DATA,		/* 4 to 128 bytes */
+
+	__TCA_TUNNEL_KEY_ENC_OPT_GENEVE_MAX,
+};
+
+#define TCA_TUNNEL_KEY_ENC_OPT_GENEVE_MAX \
+	(__TCA_TUNNEL_KEY_ENC_OPT_GENEVE_MAX - 1)
+
 #endif

net/core/filter.c

@@ -3582,7 +3582,7 @@ BPF_CALL_3(bpf_skb_set_tunnel_opt, struct sk_buff *, skb,
 	if (unlikely(size > IP_TUNNEL_OPTS_MAX))
 		return -ENOMEM;
 
-	ip_tunnel_info_opts_set(info, from, size);
+	ip_tunnel_info_opts_set(info, from, size, TUNNEL_OPTIONS_PRESENT);
 
 	return 0;
 }

net/ipv4/ip_gre.c

@@ -587,6 +587,8 @@ static void erspan_fb_xmit(struct sk_buff *skb, struct net_device *dev,
 		goto err_free_skb;
 
 	key = &tun_info->key;
+	if (!(tun_info->key.tun_flags & TUNNEL_ERSPAN_OPT))
+		goto err_free_rt;
 	md = ip_tunnel_info_opts(tun_info);
 	if (!md)
 		goto err_free_rt;

net/ipv6/ip6_gre.c

@@ -990,6 +990,8 @@ static netdev_tx_t ip6erspan_tunnel_xmit(struct sk_buff *skb,
 		fl6.flowi6_uid = sock_net_uid(dev_net(dev), NULL);
 
 		dsfield = key->tos;
+		if (!(tun_info->key.tun_flags & TUNNEL_ERSPAN_OPT))
+			goto tx_err;
 		md = ip_tunnel_info_opts(tun_info);
 		if (!md)
 			goto tx_err;

net/openvswitch/flow_netlink.c

@@ -2516,7 +2516,9 @@ static int validate_and_copy_set_tun(const struct nlattr *attr,
 	struct ovs_tunnel_info *ovs_tun;
 	struct nlattr *a;
 	int err = 0, start, opts_type;
+	__be16 dst_opt_type;
 
+	dst_opt_type = 0;
 	ovs_match_init(&match, &key, true, NULL);
 	opts_type = ip_tun_from_nlattr(nla_data(attr), &match, false, log);
 	if (opts_type < 0)
@@ -2528,10 +2530,13 @@ static int validate_and_copy_set_tun(const struct nlattr *attr,
 			err = validate_geneve_opts(&key);
 			if (err < 0)
 				return err;
+			dst_opt_type = TUNNEL_GENEVE_OPT;
 			break;
 		case OVS_TUNNEL_KEY_ATTR_VXLAN_OPTS:
+			dst_opt_type = TUNNEL_VXLAN_OPT;
 			break;
 		case OVS_TUNNEL_KEY_ATTR_ERSPAN_OPTS:
+			dst_opt_type = TUNNEL_ERSPAN_OPT;
 			break;
 		}
 	}
@@ -2574,7 +2579,7 @@ static int validate_and_copy_set_tun(const struct nlattr *attr,
 	 */
 	ip_tunnel_info_opts_set(tun_info,
 				TUN_METADATA_OPTS(&key, key.tun_opts_len),
-				key.tun_opts_len);
+				key.tun_opts_len, dst_opt_type);
 	add_nested_action_end(*sfa, start);
 
 	return err;

net/sched/act_tunnel_key.c

@@ -13,6 +13,7 @@
 #include <linux/kernel.h>
 #include <linux/skbuff.h>
 #include <linux/rtnetlink.h>
+#include <net/geneve.h>
 #include <net/netlink.h>
 #include <net/pkt_sched.h>
 #include <net/dst.h>
@@ -57,6 +58,135 @@ static int tunnel_key_act(struct sk_buff *skb, const struct tc_action *a,
 	return action;
 }
 
+static const struct nla_policy
+enc_opts_policy[TCA_TUNNEL_KEY_ENC_OPTS_MAX + 1] = {
+	[TCA_TUNNEL_KEY_ENC_OPTS_GENEVE]	= { .type = NLA_NESTED },
+};
+
+static const struct nla_policy
+geneve_opt_policy[TCA_TUNNEL_KEY_ENC_OPT_GENEVE_MAX + 1] = {
+	[TCA_TUNNEL_KEY_ENC_OPT_GENEVE_CLASS]	   = { .type = NLA_U16 },
+	[TCA_TUNNEL_KEY_ENC_OPT_GENEVE_TYPE]	   = { .type = NLA_U8 },
+	[TCA_TUNNEL_KEY_ENC_OPT_GENEVE_DATA]	   = { .type = NLA_BINARY,
+						       .len = 128 },
+};
+
+static int
+tunnel_key_copy_geneve_opt(const struct nlattr *nla, void *dst, int dst_len,
+			   struct netlink_ext_ack *extack)
+{
+	struct nlattr *tb[TCA_TUNNEL_KEY_ENC_OPT_GENEVE_MAX + 1];
+	int err, data_len, opt_len;
+	u8 *data;
+
+	err = nla_parse_nested(tb, TCA_TUNNEL_KEY_ENC_OPT_GENEVE_MAX,
+			       nla, geneve_opt_policy, extack);
+	if (err < 0)
+		return err;
+
+	if (!tb[TCA_TUNNEL_KEY_ENC_OPT_GENEVE_CLASS] ||
+	    !tb[TCA_TUNNEL_KEY_ENC_OPT_GENEVE_TYPE] ||
+	    !tb[TCA_TUNNEL_KEY_ENC_OPT_GENEVE_DATA]) {
+		NL_SET_ERR_MSG(extack, "Missing tunnel key geneve option class, type or data");
+		return -EINVAL;
+	}
+
+	data = nla_data(tb[TCA_TUNNEL_KEY_ENC_OPT_GENEVE_DATA]);
+	data_len = nla_len(tb[TCA_TUNNEL_KEY_ENC_OPT_GENEVE_DATA]);
+	if (data_len < 4) {
+		NL_SET_ERR_MSG(extack, "Tunnel key geneve option data is less than 4 bytes long");
+		return -ERANGE;
+	}
+	if (data_len % 4) {
+		NL_SET_ERR_MSG(extack, "Tunnel key geneve option data is not a multiple of 4 bytes long");
+		return -ERANGE;
+	}
+
+	opt_len = sizeof(struct geneve_opt) + data_len;
+	if (dst) {
+		struct geneve_opt *opt = dst;
+
+		WARN_ON(dst_len < opt_len);
+
+		opt->opt_class =
+			nla_get_be16(tb[TCA_TUNNEL_KEY_ENC_OPT_GENEVE_CLASS]);
+		opt->type = nla_get_u8(tb[TCA_TUNNEL_KEY_ENC_OPT_GENEVE_TYPE]);
+		opt->length = data_len / 4; /* length is in units of 4 bytes */
+		opt->r1 = 0;
+		opt->r2 = 0;
+		opt->r3 = 0;
+
+		memcpy(opt + 1, data, data_len);
+	}
+
+	return opt_len;
+}
+
+static int tunnel_key_copy_opts(const struct nlattr *nla, u8 *dst,
+				int dst_len, struct netlink_ext_ack *extack)
+{
+	int err, rem, opt_len, len = nla_len(nla), opts_len = 0;
+	const struct nlattr *attr, *head = nla_data(nla);
+
+	err = nla_validate(head, len, TCA_TUNNEL_KEY_ENC_OPTS_MAX,
+			   enc_opts_policy, extack);
+	if (err)
+		return err;
+
+	nla_for_each_attr(attr, head, len, rem) {
+		switch (nla_type(attr)) {
+		case TCA_TUNNEL_KEY_ENC_OPTS_GENEVE:
+			opt_len = tunnel_key_copy_geneve_opt(attr, dst,
+							     dst_len, extack);
+			if (opt_len < 0)
+				return opt_len;
+			opts_len += opt_len;
+			if (dst) {
+				dst_len -= opt_len;
+				dst += opt_len;
+			}
+			break;
+		}
+	}
+
+	if (!opts_len) {
+		NL_SET_ERR_MSG(extack, "Empty list of tunnel options");
+		return -EINVAL;
+	}
+
+	if (rem > 0) {
+		NL_SET_ERR_MSG(extack, "Trailing data after parsing tunnel key options attributes");
+		return -EINVAL;
+	}
+
+	return opts_len;
+}
+
+static int tunnel_key_get_opts_len(struct nlattr *nla,
+				   struct netlink_ext_ack *extack)
+{
+	return tunnel_key_copy_opts(nla, NULL, 0, extack);
+}
+
+static int tunnel_key_opts_set(struct nlattr *nla, struct ip_tunnel_info *info,
+			       int opts_len, struct netlink_ext_ack *extack)
+{
+	info->options_len = opts_len;
+	switch (nla_type(nla_data(nla))) {
+	case TCA_TUNNEL_KEY_ENC_OPTS_GENEVE:
+#if IS_ENABLED(CONFIG_INET)
+		info->key.tun_flags |= TUNNEL_GENEVE_OPT;
+		return tunnel_key_copy_opts(nla, ip_tunnel_info_opts(info),
+					    opts_len, extack);
+#else
+		return -EAFNOSUPPORT;
+#endif
+	default:
+		NL_SET_ERR_MSG(extack, "Cannot set tunnel options for unknown tunnel type");
+		return -EINVAL;
+	}
+}
+
 static const struct nla_policy tunnel_key_policy[TCA_TUNNEL_KEY_MAX + 1] = {
 	[TCA_TUNNEL_KEY_PARMS]	    = { .len = sizeof(struct tc_tunnel_key) },
 	[TCA_TUNNEL_KEY_ENC_IPV4_SRC] = { .type = NLA_U32 },
@@ -66,6 +196,7 @@ static const struct nla_policy tunnel_key_policy[TCA_TUNNEL_KEY_MAX + 1] = {
 	[TCA_TUNNEL_KEY_ENC_KEY_ID]   = { .type = NLA_U32 },
 	[TCA_TUNNEL_KEY_ENC_DST_PORT] = {.type = NLA_U16},
 	[TCA_TUNNEL_KEY_NO_CSUM]      = { .type = NLA_U8 },
+	[TCA_TUNNEL_KEY_ENC_OPTS]     = { .type = NLA_NESTED },
 };
 
 static int tunnel_key_init(struct net *net, struct nlattr *nla,
@@ -81,21 +212,28 @@ static int tunnel_key_init(struct net *net, struct nlattr *nla,
 	struct tcf_tunnel_key *t;
 	bool exists = false;
 	__be16 dst_port = 0;
+	int opts_len = 0;
 	__be64 key_id;
 	__be16 flags;
 	int ret = 0;
 	int err;
 
-	if (!nla)
+	if (!nla) {
+		NL_SET_ERR_MSG(extack, "Tunnel requires attributes to be passed");
 		return -EINVAL;
+	}
 
 	err = nla_parse_nested(tb, TCA_TUNNEL_KEY_MAX, nla, tunnel_key_policy,
-			       NULL);
-	if (err < 0)
+			       extack);
+	if (err < 0) {
+		NL_SET_ERR_MSG(extack, "Failed to parse nested tunnel key attributes");
 		return err;
+	}
 
-	if (!tb[TCA_TUNNEL_KEY_PARMS])
+	if (!tb[TCA_TUNNEL_KEY_PARMS]) {
+		NL_SET_ERR_MSG(extack, "Missing tunnel key parameters");
 		return -EINVAL;
+	}
 
 	parm = nla_data(tb[TCA_TUNNEL_KEY_PARMS]);
 	exists = tcf_idr_check(tn, parm->index, a, bind);
@@ -107,6 +245,7 @@ static int tunnel_key_init(struct net *net, struct nlattr *nla,
 		break;
 	case TCA_TUNNEL_KEY_ACT_SET:
 		if (!tb[TCA_TUNNEL_KEY_ENC_KEY_ID]) {
+			NL_SET_ERR_MSG(extack, "Missing tunnel key id");
 			ret = -EINVAL;
 			goto err_out;
 		}
@@ -121,6 +260,15 @@ static int tunnel_key_init(struct net *net, struct nlattr *nla,
 		if (tb[TCA_TUNNEL_KEY_ENC_DST_PORT])
 			dst_port = nla_get_be16(tb[TCA_TUNNEL_KEY_ENC_DST_PORT]);
 
+		if (tb[TCA_TUNNEL_KEY_ENC_OPTS]) {
+			opts_len = tunnel_key_get_opts_len(tb[TCA_TUNNEL_KEY_ENC_OPTS],
+							   extack);
+			if (opts_len < 0) {
+				ret = opts_len;
+				goto err_out;
+			}
+		}
+
 		if (tb[TCA_TUNNEL_KEY_ENC_IPV4_SRC] &&
 		    tb[TCA_TUNNEL_KEY_ENC_IPV4_DST]) {
 			__be32 saddr;
@@ -131,7 +279,7 @@ static int tunnel_key_init(struct net *net, struct nlattr *nla,
 
 			metadata = __ip_tun_set_dst(saddr, daddr, 0, 0,
 						    dst_port, flags,
-						    key_id, 0);
+						    key_id, opts_len);
 		} else if (tb[TCA_TUNNEL_KEY_ENC_IPV6_SRC] &&
 			   tb[TCA_TUNNEL_KEY_ENC_IPV6_DST]) {
 			struct in6_addr saddr;
@@ -143,16 +291,30 @@ static int tunnel_key_init(struct net *net, struct nlattr *nla,
 			metadata = __ipv6_tun_set_dst(&saddr, &daddr, 0, 0, dst_port,
 						      0, flags,
 						      key_id, 0);
+		} else {
+			NL_SET_ERR_MSG(extack, "Missing either ipv4 or ipv6 src and dst");
+			ret = -EINVAL;
+			goto err_out;
 		}
 
 		if (!metadata) {
-			ret = -EINVAL;
+			NL_SET_ERR_MSG(extack, "Cannot allocate tunnel metadata dst");
+			ret = -ENOMEM;
 			goto err_out;
 		}
 
+		if (opts_len) {
+			ret = tunnel_key_opts_set(tb[TCA_TUNNEL_KEY_ENC_OPTS],
+						  &metadata->u.tun_info,
+						  opts_len, extack);
+			if (ret < 0)
+				goto err_out;
+		}
+
 		metadata->u.tun_info.mode |= IP_TUNNEL_INFO_TX;
 		break;
 	default:
+		NL_SET_ERR_MSG(extack, "Unknown tunnel key action");
 		ret = -EINVAL;
 		goto err_out;
 	}
@@ -160,14 +322,18 @@ static int tunnel_key_init(struct net *net, struct nlattr *nla,
 	if (!exists) {
 		ret = tcf_idr_create(tn, parm->index, est, a,
 				     &act_tunnel_key_ops, bind, true);
-		if (ret)
+		if (ret) {
+			NL_SET_ERR_MSG(extack, "Cannot create TC IDR");
 			return ret;
+		}
 
 		ret = ACT_P_CREATED;
 	} else {
 		tcf_idr_release(*a, bind);
-		if (!ovr)
+		if (!ovr) {
+			NL_SET_ERR_MSG(extack, "TC IDR already exists");
 			return -EEXIST;
+		}
 	}
 
 	t = to_tunnel_key(*a);
@@ -177,6 +343,7 @@ static int tunnel_key_init(struct net *net, struct nlattr *nla,
 	if (unlikely(!params_new)) {
 		if (ret == ACT_P_CREATED)
 			tcf_idr_release(*a, bind);
+		NL_SET_ERR_MSG(extack, "Cannot allocate tunnel key parameters");
 		return -ENOMEM;
 	}
 
@@ -216,6 +383,61 @@ static void tunnel_key_release(struct tc_action *a)
 	}
 }
 
+static int tunnel_key_geneve_opts_dump(struct sk_buff *skb,
+				       const struct ip_tunnel_info *info)
+{
+	int len = info->options_len;
+	u8 *src = (u8 *)(info + 1);
+	struct nlattr *start;
+
+	start = nla_nest_start(skb, TCA_TUNNEL_KEY_ENC_OPTS_GENEVE);
+	if (!start)
+		return -EMSGSIZE;
+
+	while (len > 0) {
+		struct geneve_opt *opt = (struct geneve_opt *)src;
+
+		if (nla_put_be16(skb, TCA_TUNNEL_KEY_ENC_OPT_GENEVE_CLASS,
+				 opt->opt_class) ||
+		    nla_put_u8(skb, TCA_TUNNEL_KEY_ENC_OPT_GENEVE_TYPE,
+			       opt->type) ||
+		    nla_put(skb, TCA_TUNNEL_KEY_ENC_OPT_GENEVE_DATA,
+			    opt->length * 4, opt + 1))
+			return -EMSGSIZE;
+
+		len -= sizeof(struct geneve_opt) + opt->length * 4;
+		src += sizeof(struct geneve_opt) + opt->length * 4;
+	}
+
+	nla_nest_end(skb, start);
+	return 0;
+}
+
+static int tunnel_key_opts_dump(struct sk_buff *skb,
+				const struct ip_tunnel_info *info)
+{
+	struct nlattr *start;
+	int err;
+
+	if (!info->options_len)
+		return 0;
+
+	start = nla_nest_start(skb, TCA_TUNNEL_KEY_ENC_OPTS);
+	if (!start)
+		return -EMSGSIZE;
+
+	if (info->key.tun_flags & TUNNEL_GENEVE_OPT) {
+		err = tunnel_key_geneve_opts_dump(skb, info);
+		if (err)
+			return err;
+	} else {
+		return -EINVAL;
+	}
+
+	nla_nest_end(skb, start);
+	return 0;
+}
+
 static int tunnel_key_dump_addresses(struct sk_buff *skb,
 				     const struct ip_tunnel_info *info)
 {
@@ -266,8 +488,9 @@ static int tunnel_key_dump(struct sk_buff *skb, struct tc_action *a,
 		goto nla_put_failure;
 
 	if (params->tcft_action == TCA_TUNNEL_KEY_ACT_SET) {
-		struct ip_tunnel_key *key =
-			&params->tcft_enc_metadata->u.tun_info.key;
+		struct ip_tunnel_info *info =
+			&params->tcft_enc_metadata->u.tun_info;
+		struct ip_tunnel_key *key = &info->key;
 		__be32 key_id = tunnel_id_to_key32(key->tun_id);
 
 		if (nla_put_be32(skb, TCA_TUNNEL_KEY_ENC_KEY_ID, key_id) ||
@@ -275,7 +498,8 @@ static int tunnel_key_dump(struct sk_buff *skb, struct tc_action *a,
 					      &params->tcft_enc_metadata->u.tun_info) ||
 		    nla_put_be16(skb, TCA_TUNNEL_KEY_ENC_DST_PORT, key->tp_dst) ||
 		    nla_put_u8(skb, TCA_TUNNEL_KEY_NO_CSUM,
-			       !(key->tun_flags & TUNNEL_CSUM)))
+			       !(key->tun_flags & TUNNEL_CSUM)) ||
+		    tunnel_key_opts_dump(skb, info))
 			goto nla_put_failure;
 	}