7 files changed, 643 insertions, 0 deletions

diff --git a/include/net/net_namespace.h b/include/net/net_namespace.h
index bcc4a8ed4450..da68c9a90ac5 100644
--- a/include/net/net_namespace.h
+++ b/include/net/net_namespace.h
@@ -22,6 +22,7 @@
 #if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE)
 #include <net/netns/conntrack.h>
 #endif
+#include <net/netns/nftables.h>
 #include <net/netns/xfrm.h>
 
 struct user_namespace;
@@ -101,6 +102,9 @@ struct net {
 #if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE)
 	struct netns_ct		ct;
 #endif
+#if defined(CONFIG_NF_TABLES) || defined(CONFIG_NF_TABLES_MODULE)
+	struct netns_nftables	nft;
+#endif
 #if IS_ENABLED(CONFIG_NF_DEFRAG_IPV6)
 	struct netns_nf_frag	nf_frag;
 #endif
diff --git a/include/net/netfilter/nf_nat.h b/include/net/netfilter/nf_nat.h
index c29b4e545f87..07eaaf604092 100644
--- a/include/net/netfilter/nf_nat.h
+++ b/include/net/netfilter/nf_nat.h
@@ -45,6 +45,9 @@ unsigned int nf_nat_setup_info(struct nf_conn *ct,
 			       const struct nf_nat_range *range,
 			       enum nf_nat_manip_type maniptype);
 
+extern unsigned int nf_nat_alloc_null_binding(struct nf_conn *ct,
+					      unsigned int hooknum);
+
 /* Is this tuple already taken? (not by us)*/
 int nf_nat_used_tuple(const struct nf_conntrack_tuple *tuple,
 		      const struct nf_conn *ignored_conntrack);
diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h
new file mode 100644
index 000000000000..54c4a5cafb64
--- /dev/null
+++ b/include/net/netfilter/nf_tables.h
@@ -0,0 +1,522 @@
+#ifndef _NET_NF_TABLES_H
+#define _NET_NF_TABLES_H
+
+#include <linux/list.h>
+#include <linux/netfilter.h>
+#include <linux/netfilter/x_tables.h>
+#include <linux/netfilter/nf_tables.h>
+#include <net/netlink.h>
+
+#define NFT_JUMP_STACK_SIZE	16
+
+struct nft_pktinfo {
+	struct sk_buff			*skb;
+	const struct net_device		*in;
+	const struct net_device		*out;
+	u8				hooknum;
+	u8				nhoff;
+	u8				thoff;
+	/* for x_tables compatibility */
+	struct xt_action_param		xt;
+};
+
+static inline void nft_set_pktinfo(struct nft_pktinfo *pkt,
+				   const struct nf_hook_ops *ops,
+				   struct sk_buff *skb,
+				   const struct net_device *in,
+				   const struct net_device *out)
+{
+	pkt->skb = skb;
+	pkt->in = pkt->xt.in = in;
+	pkt->out = pkt->xt.out = out;
+	pkt->hooknum = pkt->xt.hooknum = ops->hooknum;
+	pkt->xt.family = ops->pf;
+}
+
+struct nft_data {
+	union {
+		u32				data[4];
+		struct {
+			u32			verdict;
+			struct nft_chain	*chain;
+		};
+	};
+} __attribute__((aligned(__alignof__(u64))));
+
+static inline int nft_data_cmp(const struct nft_data *d1,
+			       const struct nft_data *d2,
+			       unsigned int len)
+{
+	return memcmp(d1->data, d2->data, len);
+}
+
+static inline void nft_data_copy(struct nft_data *dst,
+				 const struct nft_data *src)
+{
+	BUILD_BUG_ON(__alignof__(*dst) != __alignof__(u64));
+	*(u64 *)&dst->data[0] = *(u64 *)&src->data[0];
+	*(u64 *)&dst->data[2] = *(u64 *)&src->data[2];
+}
+
+static inline void nft_data_debug(const struct nft_data *data)
+{
+	pr_debug("data[0]=%x data[1]=%x data[2]=%x data[3]=%x\n",
+		 data->data[0], data->data[1],
+		 data->data[2], data->data[3]);
+}
+
+/**
+ *	struct nft_ctx - nf_tables rule/set context
+ *
+ *	@net: net namespace
+ * 	@skb: netlink skb
+ * 	@nlh: netlink message header
+ * 	@afi: address family info
+ * 	@table: the table the chain is contained in
+ * 	@chain: the chain the rule is contained in
+ *	@nla: netlink attributes
+ */
+struct nft_ctx {
+	struct net			*net;
+	const struct sk_buff		*skb;
+	const struct nlmsghdr		*nlh;
+	const struct nft_af_info	*afi;
+	const struct nft_table		*table;
+	const struct nft_chain		*chain;
+	const struct nlattr * const 	*nla;
+};
+
+struct nft_data_desc {
+	enum nft_data_types		type;
+	unsigned int			len;
+};
+
+extern int nft_data_init(const struct nft_ctx *ctx, struct nft_data *data,
+			 struct nft_data_desc *desc, const struct nlattr *nla);
+extern void nft_data_uninit(const struct nft_data *data,
+			    enum nft_data_types type);
+extern int nft_data_dump(struct sk_buff *skb, int attr,
+			 const struct nft_data *data,
+			 enum nft_data_types type, unsigned int len);
+
+static inline enum nft_data_types nft_dreg_to_type(enum nft_registers reg)
+{
+	return reg == NFT_REG_VERDICT ? NFT_DATA_VERDICT : NFT_DATA_VALUE;
+}
+
+static inline enum nft_registers nft_type_to_reg(enum nft_data_types type)
+{
+	return type == NFT_DATA_VERDICT ? NFT_REG_VERDICT : NFT_REG_1;
+}
+
+extern int nft_validate_input_register(enum nft_registers reg);
+extern int nft_validate_output_register(enum nft_registers reg);
+extern int nft_validate_data_load(const struct nft_ctx *ctx,
+				  enum nft_registers reg,
+				  const struct nft_data *data,
+				  enum nft_data_types type);
+
+/**
+ *	struct nft_set_elem - generic representation of set elements
+ *
+ *	@cookie: implementation specific element cookie
+ *	@key: element key
+ *	@data: element data (maps only)
+ *	@flags: element flags (end of interval)
+ *
+ *	The cookie can be used to store a handle to the element for subsequent
+ *	removal.
+ */
+struct nft_set_elem {
+	void			*cookie;
+	struct nft_data		key;
+	struct nft_data		data;
+	u32			flags;
+};
+
+struct nft_set;
+struct nft_set_iter {
+	unsigned int	count;
+	unsigned int	skip;
+	int		err;
+	int		(*fn)(const struct nft_ctx *ctx,
+			      const struct nft_set *set,
+			      const struct nft_set_iter *iter,
+			      const struct nft_set_elem *elem);
+};
+
+/**
+ *	struct nft_set_ops - nf_tables set operations
+ *
+ *	@lookup: look up an element within the set
+ *	@insert: insert new element into set
+ *	@remove: remove element from set
+ *	@walk: iterate over all set elemeennts
+ *	@privsize: function to return size of set private data
+ *	@init: initialize private data of new set instance
+ *	@destroy: destroy private data of set instance
+ *	@list: nf_tables_set_ops list node
+ *	@owner: module reference
+ *	@features: features supported by the implementation
+ */
+struct nft_set_ops {
+	bool				(*lookup)(const struct nft_set *set,
+						  const struct nft_data *key,
+						  struct nft_data *data);
+	int				(*get)(const struct nft_set *set,
+					       struct nft_set_elem *elem);
+	int				(*insert)(const struct nft_set *set,
+						  const struct nft_set_elem *elem);
+	void				(*remove)(const struct nft_set *set,
+						  const struct nft_set_elem *elem);
+	void				(*walk)(const struct nft_ctx *ctx,
+						const struct nft_set *set,
+						struct nft_set_iter *iter);
+
+	unsigned int			(*privsize)(const struct nlattr * const nla[]);
+	int				(*init)(const struct nft_set *set,
+						const struct nlattr * const nla[]);
+	void				(*destroy)(const struct nft_set *set);
+
+	struct list_head		list;
+	struct module			*owner;
+	u32				features;
+};
+
+extern int nft_register_set(struct nft_set_ops *ops);
+extern void nft_unregister_set(struct nft_set_ops *ops);
+
+/**
+ * 	struct nft_set - nf_tables set instance
+ *
+ *	@list: table set list node
+ *	@bindings: list of set bindings
+ * 	@name: name of the set
+ * 	@ktype: key type (numeric type defined by userspace, not used in the kernel)
+ * 	@dtype: data type (verdict or numeric type defined by userspace)
+ * 	@ops: set ops
+ * 	@flags: set flags
+ * 	@klen: key length
+ * 	@dlen: data length
+ * 	@data: private set data
+ */
+struct nft_set {
+	struct list_head		list;
+	struct list_head		bindings;
+	char				name[IFNAMSIZ];
+	u32				ktype;
+	u32				dtype;
+	/* runtime data below here */
+	const struct nft_set_ops	*ops ____cacheline_aligned;
+	u16				flags;
+	u8				klen;
+	u8				dlen;
+	unsigned char			data[]
+		__attribute__((aligned(__alignof__(u64))));
+};
+
+static inline void *nft_set_priv(const struct nft_set *set)
+{
+	return (void *)set->data;
+}
+
+extern struct nft_set *nf_tables_set_lookup(const struct nft_table *table,
+					    const struct nlattr *nla);
+
+/**
+ *	struct nft_set_binding - nf_tables set binding
+ *
+ *	@list: set bindings list node
+ *	@chain: chain containing the rule bound to the set
+ *
+ *	A set binding contains all information necessary for validation
+ *	of new elements added to a bound set.
+ */
+struct nft_set_binding {
+	struct list_head		list;
+	const struct nft_chain		*chain;
+};
+
+extern int nf_tables_bind_set(const struct nft_ctx *ctx, struct nft_set *set,
+			      struct nft_set_binding *binding);
+extern void nf_tables_unbind_set(const struct nft_ctx *ctx, struct nft_set *set,
+				 struct nft_set_binding *binding);
+
+
+/**
+ *	struct nft_expr_type - nf_tables expression type
+ *
+ *	@select_ops: function to select nft_expr_ops
+ *	@ops: default ops, used when no select_ops functions is present
+ *	@list: used internally
+ *	@name: Identifier
+ *	@owner: module reference
+ *	@policy: netlink attribute policy
+ *	@maxattr: highest netlink attribute number
+ */
+struct nft_expr_type {
+	const struct nft_expr_ops	*(*select_ops)(const struct nft_ctx *,
+						       const struct nlattr * const tb[]);
+	const struct nft_expr_ops	*ops;
+	struct list_head		list;
+	const char			*name;
+	struct module			*owner;
+	const struct nla_policy		*policy;
+	unsigned int			maxattr;
+};
+
+/**
+ *	struct nft_expr_ops - nf_tables expression operations
+ *
+ *	@eval: Expression evaluation function
+ *	@size: full expression size, including private data size
+ *	@init: initialization function
+ *	@destroy: destruction function
+ *	@dump: function to dump parameters
+ *	@type: expression type
+ *	@validate: validate expression, called during loop detection
+ *	@data: extra data to attach to this expression operation
+ */
+struct nft_expr;
+struct nft_expr_ops {
+	void				(*eval)(const struct nft_expr *expr,
+						struct nft_data data[NFT_REG_MAX + 1],
+						const struct nft_pktinfo *pkt);
+	unsigned int			size;
+
+	int				(*init)(const struct nft_ctx *ctx,
+						const struct nft_expr *expr,
+						const struct nlattr * const tb[]);
+	void				(*destroy)(const struct nft_expr *expr);
+	int				(*dump)(struct sk_buff *skb,
+						const struct nft_expr *expr);
+	int				(*validate)(const struct nft_ctx *ctx,
+						    const struct nft_expr *expr,
+						    const struct nft_data **data);
+	const struct nft_expr_type	*type;
+	void				*data;
+};
+
+#define NFT_EXPR_MAXATTR		16
+#define NFT_EXPR_SIZE(size)		(sizeof(struct nft_expr) + \
+					 ALIGN(size, __alignof__(struct nft_expr)))
+
+/**
+ *	struct nft_expr - nf_tables expression
+ *
+ *	@ops: expression ops
+ *	@data: expression private data
+ */
+struct nft_expr {
+	const struct nft_expr_ops	*ops;
+	unsigned char			data[];
+};
+
+static inline void *nft_expr_priv(const struct nft_expr *expr)
+{
+	return (void *)expr->data;
+}
+
+/**
+ *	struct nft_rule - nf_tables rule
+ *
+ *	@list: used internally
+ *	@rcu_head: used internally for rcu
+ *	@handle: rule handle
+ *	@genmask: generation mask
+ *	@dlen: length of expression data
+ *	@data: expression data
+ */
+struct nft_rule {
+	struct list_head		list;
+	struct rcu_head			rcu_head;
+	u64				handle:46,
+					genmask:2,
+					dlen:16;
+	unsigned char			data[]
+		__attribute__((aligned(__alignof__(struct nft_expr))));
+};
+
+/**
+ *	struct nft_rule_trans - nf_tables rule update in transaction
+ *
+ *	@list: used internally
+ *	@rule: rule that needs to be updated
+ *	@chain: chain that this rule belongs to
+ *	@table: table for which this chain applies
+ *	@nlh: netlink header of the message that contain this update
+ *	@family: family expressesed as AF_*
+ */
+struct nft_rule_trans {
+	struct list_head		list;
+	struct nft_rule			*rule;
+	const struct nft_chain		*chain;
+	const struct nft_table		*table;
+	const struct nlmsghdr		*nlh;
+	u8				family;
+};
+
+static inline struct nft_expr *nft_expr_first(const struct nft_rule *rule)
+{
+	return (struct nft_expr *)&rule->data[0];
+}
+
+static inline struct nft_expr *nft_expr_next(const struct nft_expr *expr)
+{
+	return ((void *)expr) + expr->ops->size;
+}
+
+static inline struct nft_expr *nft_expr_last(const struct nft_rule *rule)
+{
+	return (struct nft_expr *)&rule->data[rule->dlen];
+}
+
+/*
+ * The last pointer isn't really necessary, but the compiler isn't able to
+ * determine that the result of nft_expr_last() is always the same since it
+ * can't assume that the dlen value wasn't changed within calls in the loop.
+ */
+#define nft_rule_for_each_expr(expr, last, rule) \
+	for ((expr) = nft_expr_first(rule), (last) = nft_expr_last(rule); \
+	     (expr) != (last); \
+	     (expr) = nft_expr_next(expr))
+
+enum nft_chain_flags {
+	NFT_BASE_CHAIN			= 0x1,
+};
+
+/**
+ *	struct nft_chain - nf_tables chain
+ *
+ *	@rules: list of rules in the chain
+ *	@list: used internally
+ *	@rcu_head: used internally
+ *	@net: net namespace that this chain belongs to
+ *	@table: table that this chain belongs to
+ *	@handle: chain handle
+ *	@flags: bitmask of enum nft_chain_flags
+ *	@use: number of jump references to this chain
+ *	@level: length of longest path to this chain
+ *	@name: name of the chain
+ */
+struct nft_chain {
+	struct list_head		rules;
+	struct list_head		list;
+	struct rcu_head			rcu_head;
+	struct net			*net;
+	struct nft_table		*table;
+	u64				handle;
+	u8				flags;
+	u16				use;
+	u16				level;
+	char				name[NFT_CHAIN_MAXNAMELEN];
+};
+
+enum nft_chain_type {
+	NFT_CHAIN_T_DEFAULT = 0,
+	NFT_CHAIN_T_ROUTE,
+	NFT_CHAIN_T_NAT,
+	NFT_CHAIN_T_MAX
+};
+
+struct nft_stats {
+	u64 bytes;
+	u64 pkts;
+};
+
+/**
+ *	struct nft_base_chain - nf_tables base chain
+ *
+ *	@ops: netfilter hook ops
+ *	@type: chain type
+ *	@policy: default policy
+ *	@stats: per-cpu chain stats
+ *	@chain: the chain
+ */
+struct nft_base_chain {
+	struct nf_hook_ops		ops;
+	enum nft_chain_type		type;
+	u8				policy;
+	struct nft_stats __percpu	*stats;
+	struct nft_chain		chain;
+};
+
+static inline struct nft_base_chain *nft_base_chain(const struct nft_chain *chain)
+{
+	return container_of(chain, struct nft_base_chain, chain);
+}
+
+extern unsigned int nft_do_chain_pktinfo(struct nft_pktinfo *pkt,
+					 const struct nf_hook_ops *ops);
+
+/**
+ *	struct nft_table - nf_tables table
+ *
+ *	@list: used internally
+ *	@chains: chains in the table
+ *	@sets: sets in the table
+ *	@hgenerator: handle generator state
+ *	@use: number of chain references to this table
+ *	@flags: table flag (see enum nft_table_flags)
+ *	@name: name of the table
+ */
+struct nft_table {
+	struct list_head		list;
+	struct list_head		chains;
+	struct list_head		sets;
+	u64				hgenerator;
+	u32				use;
+	u16				flags;
+	char				name[];
+};
+
+/**
+ *	struct nft_af_info - nf_tables address family info
+ *
+ *	@list: used internally
+ *	@family: address family
+ *	@nhooks: number of hooks in this family
+ *	@owner: module owner
+ *	@tables: used internally
+ *	@hooks: hookfn overrides for packet validation
+ */
+struct nft_af_info {
+	struct list_head		list;
+	int				family;
+	unsigned int			nhooks;
+	struct module			*owner;
+	struct list_head		tables;
+	nf_hookfn			*hooks[NF_MAX_HOOKS];
+};
+
+extern int nft_register_afinfo(struct net *, struct nft_af_info *);
+extern void nft_unregister_afinfo(struct nft_af_info *);
+
+struct nf_chain_type {
+	unsigned int		hook_mask;
+	const char		*name;
+	enum nft_chain_type	type;
+	nf_hookfn		*fn[NF_MAX_HOOKS];
+	struct module		*me;
+	int			family;
+};
+
+extern int nft_register_chain_type(struct nf_chain_type *);
+extern void nft_unregister_chain_type(struct nf_chain_type *);
+
+extern int nft_register_expr(struct nft_expr_type *);
+extern void nft_unregister_expr(struct nft_expr_type *);
+
+#define MODULE_ALIAS_NFT_FAMILY(family)	\
+	MODULE_ALIAS("nft-afinfo-" __stringify(family))
+
+#define MODULE_ALIAS_NFT_CHAIN(family, name) \
+	MODULE_ALIAS("nft-chain-" __stringify(family) "-" name)
+
+#define MODULE_ALIAS_NFT_EXPR(name) \
+	MODULE_ALIAS("nft-expr-" name)
+
+#define MODULE_ALIAS_NFT_SET() \
+	MODULE_ALIAS("nft-set")
+
+#endif /* _NET_NF_TABLES_H */
diff --git a/include/net/netfilter/nf_tables_core.h b/include/net/netfilter/nf_tables_core.h
new file mode 100644
index 000000000000..fe7b16206a4e
--- /dev/null
+++ b/include/net/netfilter/nf_tables_core.h
@@ -0,0 +1,42 @@
+#ifndef _NET_NF_TABLES_CORE_H
+#define _NET_NF_TABLES_CORE_H
+
+extern int nf_tables_core_module_init(void);
+extern void nf_tables_core_module_exit(void);
+
+extern int nft_immediate_module_init(void);
+extern void nft_immediate_module_exit(void);
+
+struct nft_cmp_fast_expr {
+	u32			data;
+	enum nft_registers	sreg:8;
+	u8			len;
+};
+
+extern const struct nft_expr_ops nft_cmp_fast_ops;
+
+extern int nft_cmp_module_init(void);
+extern void nft_cmp_module_exit(void);
+
+extern int nft_lookup_module_init(void);
+extern void nft_lookup_module_exit(void);
+
+extern int nft_bitwise_module_init(void);
+extern void nft_bitwise_module_exit(void);
+
+extern int nft_byteorder_module_init(void);
+extern void nft_byteorder_module_exit(void);
+
+struct nft_payload {
+	enum nft_payload_bases	base:8;
+	u8			offset;
+	u8			len;
+	enum nft_registers	dreg:8;
+};
+
+extern const struct nft_expr_ops nft_payload_fast_ops;
+
+extern int nft_payload_module_init(void);
+extern void nft_payload_module_exit(void);
+
+#endif /* _NET_NF_TABLES_CORE_H */
diff --git a/include/net/netfilter/nf_tables_ipv4.h b/include/net/netfilter/nf_tables_ipv4.h
new file mode 100644
index 000000000000..1be1c2c197ee
--- /dev/null
+++ b/include/net/netfilter/nf_tables_ipv4.h
@@ -0,0 +1,23 @@
+#ifndef _NF_TABLES_IPV4_H_
+#define _NF_TABLES_IPV4_H_
+
+#include <net/netfilter/nf_tables.h>
+#include <net/ip.h>
+
+static inline void
+nft_set_pktinfo_ipv4(struct nft_pktinfo *pkt,
+		     const struct nf_hook_ops *ops,
+		     struct sk_buff *skb,
+		     const struct net_device *in,
+		     const struct net_device *out)
+{
+	struct iphdr *ip;
+
+	nft_set_pktinfo(pkt, ops, skb, in, out);
+
+	pkt->xt.thoff = ip_hdrlen(pkt->skb);
+	ip = ip_hdr(pkt->skb);
+	pkt->xt.fragoff = ntohs(ip->frag_off) & IP_OFFSET;
+}
+
+#endif
diff --git a/include/net/netfilter/nf_tables_ipv6.h b/include/net/netfilter/nf_tables_ipv6.h
new file mode 100644
index 000000000000..4a9b88a65963
--- /dev/null
+++ b/include/net/netfilter/nf_tables_ipv6.h
@@ -0,0 +1,30 @@
+#ifndef _NF_TABLES_IPV6_H_
+#define _NF_TABLES_IPV6_H_
+
+#include <linux/netfilter_ipv6/ip6_tables.h>
+#include <net/ipv6.h>
+
+static inline int
+nft_set_pktinfo_ipv6(struct nft_pktinfo *pkt,
+		     const struct nf_hook_ops *ops,
+		     struct sk_buff *skb,
+		     const struct net_device *in,
+		     const struct net_device *out)
+{
+	int protohdr, thoff = 0;
+	unsigned short frag_off;
+
+	nft_set_pktinfo(pkt, ops, skb, in, out);
+
+	protohdr = ipv6_find_hdr(pkt->skb, &thoff, -1, &frag_off, NULL);
+	/* If malformed, drop it */
+	if (protohdr < 0)
+		return -1;
+
+	pkt->xt.thoff = thoff;
+	pkt->xt.fragoff = frag_off;
+
+	return 0;
+}
+
+#endif
diff --git a/include/net/netns/nftables.h b/include/net/netns/nftables.h
new file mode 100644
index 000000000000..15d056d534e3
--- /dev/null
+++ b/include/net/netns/nftables.h
@@ -0,0 +1,19 @@
+#ifndef _NETNS_NFTABLES_H_
+#define _NETNS_NFTABLES_H_
+
+#include <linux/list.h>
+
+struct nft_af_info;
+
+struct netns_nftables {
+	struct list_head	af_info;
+	struct list_head	commit_list;
+	struct nft_af_info	*ipv4;
+	struct nft_af_info	*ipv6;
+	struct nft_af_info	*arp;
+	struct nft_af_info	*bridge;
+	u8			gencursor;
+	u8			genctr;
+};
+
+#endif