1/* SPDX-License-Identifier: GPL-2.0 */
2#ifndef __NET_IP_TUNNELS_H
3#define __NET_IP_TUNNELS_H 1
4
5#include <linux/if_tunnel.h>
6#include <linux/netdevice.h>
7#include <linux/skbuff.h>
8#include <linux/socket.h>
9#include <linux/types.h>
10#include <linux/u64_stats_sync.h>
11#include <linux/bitops.h>
12
13#include <net/dsfield.h>
14#include <net/gro_cells.h>
15#include <net/inet_ecn.h>
16#include <net/netns/generic.h>
17#include <net/rtnetlink.h>
18#include <net/lwtunnel.h>
19#include <net/dst_cache.h>
20
21#if IS_ENABLED(CONFIG_IPV6)
22#include <net/ipv6.h>
23#include <net/ip6_fib.h>
24#include <net/ip6_route.h>
25#endif
26
27/* Keep error state on tunnel for 30 sec */
28#define IPTUNNEL_ERR_TIMEO	(30*HZ)
29
30/* Used to memset ip_tunnel padding. */
31#define IP_TUNNEL_KEY_SIZE	offsetofend(struct ip_tunnel_key, tp_dst)
32
33/* Used to memset ipv4 address padding. */
34#define IP_TUNNEL_KEY_IPV4_PAD	offsetofend(struct ip_tunnel_key, u.ipv4.dst)
35#define IP_TUNNEL_KEY_IPV4_PAD_LEN				\
36	(sizeof_field(struct ip_tunnel_key, u) -		\
37	 sizeof_field(struct ip_tunnel_key, u.ipv4))
38
39struct ip_tunnel_key {
40	__be64			tun_id;
41	union {
42		struct {
43			__be32	src;
44			__be32	dst;
45		} ipv4;
46		struct {
47			struct in6_addr src;
48			struct in6_addr dst;
49		} ipv6;
50	} u;
51	__be16			tun_flags;
52	u8			tos;		/* TOS for IPv4, TC for IPv6 */
53	u8			ttl;		/* TTL for IPv4, HL for IPv6 */
54	__be32			label;		/* Flow Label for IPv6 */
55	__be16			tp_src;
56	__be16			tp_dst;
57};
58
59/* Flags for ip_tunnel_info mode. */
60#define IP_TUNNEL_INFO_TX	0x01	/* represents tx tunnel parameters */
61#define IP_TUNNEL_INFO_IPV6	0x02	/* key contains IPv6 addresses */
62#define IP_TUNNEL_INFO_BRIDGE	0x04	/* represents a bridged tunnel id */
63
64/* Maximum tunnel options length. */
65#define IP_TUNNEL_OPTS_MAX					\
66	GENMASK((sizeof_field(struct ip_tunnel_info,		\
67			      options_len) * BITS_PER_BYTE) - 1, 0)
68
69struct ip_tunnel_info {
70	struct ip_tunnel_key	key;
71#ifdef CONFIG_DST_CACHE
72	struct dst_cache	dst_cache;
73#endif
74	u8			options_len;
75	u8			mode;
76};
77
78/* 6rd prefix/relay information */
79#ifdef CONFIG_IPV6_SIT_6RD
80struct ip_tunnel_6rd_parm {
81	struct in6_addr		prefix;
82	__be32			relay_prefix;
83	u16			prefixlen;
84	u16			relay_prefixlen;
85};
86#endif
87
88struct ip_tunnel_encap {
89	u16			type;
90	u16			flags;
91	__be16			sport;
92	__be16			dport;
93};
94
95struct ip_tunnel_prl_entry {
96	struct ip_tunnel_prl_entry __rcu *next;
97	__be32				addr;
98	u16				flags;
99	struct rcu_head			rcu_head;
100};
101
102struct metadata_dst;
103
104struct ip_tunnel {
105	struct ip_tunnel __rcu	*next;
106	struct hlist_node hash_node;
107	struct net_device	*dev;
108	struct net		*net;	/* netns for packet i/o */
109
110	unsigned long	err_time;	/* Time when the last ICMP error
111					 * arrived */
112	int		err_count;	/* Number of arrived ICMP errors */
113
114	/* These four fields used only by GRE */
115	u32		i_seqno;	/* The last seen seqno	*/
116	atomic_t	o_seqno;	/* The last output seqno */
117	int		tun_hlen;	/* Precalculated header length */
118
119	/* These four fields used only by ERSPAN */
120	u32		index;		/* ERSPAN type II index */
121	u8		erspan_ver;	/* ERSPAN version */
122	u8		dir;		/* ERSPAN direction */
123	u16		hwid;		/* ERSPAN hardware ID */
124
125	struct dst_cache dst_cache;
126
127	struct ip_tunnel_parm parms;
128
129	int		mlink;
130	int		encap_hlen;	/* Encap header length (FOU,GUE) */
131	int		hlen;		/* tun_hlen + encap_hlen */
132	struct ip_tunnel_encap encap;
133
134	/* for SIT */
135#ifdef CONFIG_IPV6_SIT_6RD
136	struct ip_tunnel_6rd_parm ip6rd;
137#endif
138	struct ip_tunnel_prl_entry __rcu *prl;	/* potential router list */
139	unsigned int		prl_count;	/* # of entries in PRL */
140	unsigned int		ip_tnl_net_id;
141	struct gro_cells	gro_cells;
142	__u32			fwmark;
143	bool			collect_md;
144	bool			ignore_df;
145};
146
147struct tnl_ptk_info {
148	__be16 flags;
149	__be16 proto;
150	__be32 key;
151	__be32 seq;
152	int hdr_len;
153};
154
155#define PACKET_RCVD	0
156#define PACKET_REJECT	1
157#define PACKET_NEXT	2
158
159#define IP_TNL_HASH_BITS   7
160#define IP_TNL_HASH_SIZE   (1 << IP_TNL_HASH_BITS)
161
162struct ip_tunnel_net {
163	struct net_device *fb_tunnel_dev;
164	struct rtnl_link_ops *rtnl_link_ops;
165	struct hlist_head tunnels[IP_TNL_HASH_SIZE];
166	struct ip_tunnel __rcu *collect_md_tun;
167	int type;
168};
169
170static inline void ip_tunnel_key_init(struct ip_tunnel_key *key,
171				      __be32 saddr, __be32 daddr,
172				      u8 tos, u8 ttl, __be32 label,
173				      __be16 tp_src, __be16 tp_dst,
174				      __be64 tun_id, __be16 tun_flags)
175{
176	key->tun_id = tun_id;
177	key->u.ipv4.src = saddr;
178	key->u.ipv4.dst = daddr;
179	memset((unsigned char *)key + IP_TUNNEL_KEY_IPV4_PAD,
180	       0, IP_TUNNEL_KEY_IPV4_PAD_LEN);
181	key->tos = tos;
182	key->ttl = ttl;
183	key->label = label;
184	key->tun_flags = tun_flags;
185
186	/* For the tunnel types on the top of IPsec, the tp_src and tp_dst of
187	 * the upper tunnel are used.
188	 * E.g: GRE over IPSEC, the tp_src and tp_port are zero.
189	 */
190	key->tp_src = tp_src;
191	key->tp_dst = tp_dst;
192
193	/* Clear struct padding. */
194	if (sizeof(*key) != IP_TUNNEL_KEY_SIZE)
195		memset((unsigned char *)key + IP_TUNNEL_KEY_SIZE,
196		       0, sizeof(*key) - IP_TUNNEL_KEY_SIZE);
197}
198
199static inline bool
200ip_tunnel_dst_cache_usable(const struct sk_buff *skb,
201			   const struct ip_tunnel_info *info)
202{
203	if (skb->mark)
204		return false;
205	if (!info)
206		return true;
207	if (info->key.tun_flags & TUNNEL_NOCACHE)
208		return false;
209
210	return true;
211}
212
213static inline unsigned short ip_tunnel_info_af(const struct ip_tunnel_info
214					       *tun_info)
215{
216	return tun_info->mode & IP_TUNNEL_INFO_IPV6 ? AF_INET6 : AF_INET;
217}
218
219static inline __be64 key32_to_tunnel_id(__be32 key)
220{
221#ifdef __BIG_ENDIAN
222	return (__force __be64)key;
223#else
224	return (__force __be64)((__force u64)key << 32);
225#endif
226}
227
228/* Returns the least-significant 32 bits of a __be64. */
229static inline __be32 tunnel_id_to_key32(__be64 tun_id)
230{
231#ifdef __BIG_ENDIAN
232	return (__force __be32)tun_id;
233#else
234	return (__force __be32)((__force u64)tun_id >> 32);
235#endif
236}
237
238#ifdef CONFIG_INET
239
240static inline void ip_tunnel_init_flow(struct flowi4 *fl4,
241				       int proto,
242				       __be32 daddr, __be32 saddr,
243				       __be32 key, __u8 tos, int oif,
244				       __u32 mark, __u32 tun_inner_hash)
245{
246	memset(fl4, 0, sizeof(*fl4));
247	fl4->flowi4_oif = oif;
248	fl4->daddr = daddr;
249	fl4->saddr = saddr;
250	fl4->flowi4_tos = tos;
251	fl4->flowi4_proto = proto;
252	fl4->fl4_gre_key = key;
253	fl4->flowi4_mark = mark;
254	fl4->flowi4_multipath_hash = tun_inner_hash;
255}
256
257int ip_tunnel_init(struct net_device *dev);
258void ip_tunnel_uninit(struct net_device *dev);
259void  ip_tunnel_dellink(struct net_device *dev, struct list_head *head);
260struct net *ip_tunnel_get_link_net(const struct net_device *dev);
261int ip_tunnel_get_iflink(const struct net_device *dev);
262int ip_tunnel_init_net(struct net *net, unsigned int ip_tnl_net_id,
263		       struct rtnl_link_ops *ops, char *devname);
264
265void ip_tunnel_delete_nets(struct list_head *list_net, unsigned int id,
266			   struct rtnl_link_ops *ops);
267
268void ip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev,
269		    const struct iphdr *tnl_params, const u8 protocol);
270void ip_md_tunnel_xmit(struct sk_buff *skb, struct net_device *dev,
271		       const u8 proto, int tunnel_hlen);
272int ip_tunnel_ctl(struct net_device *dev, struct ip_tunnel_parm *p, int cmd);
273int ip_tunnel_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd);
274int __ip_tunnel_change_mtu(struct net_device *dev, int new_mtu, bool strict);
275int ip_tunnel_change_mtu(struct net_device *dev, int new_mtu);
276
277void ip_tunnel_get_stats64(struct net_device *dev,
278			   struct rtnl_link_stats64 *tot);
279struct ip_tunnel *ip_tunnel_lookup(struct ip_tunnel_net *itn,
280				   int link, __be16 flags,
281				   __be32 remote, __be32 local,
282				   __be32 key);
283
284int ip_tunnel_rcv(struct ip_tunnel *tunnel, struct sk_buff *skb,
285		  const struct tnl_ptk_info *tpi, struct metadata_dst *tun_dst,
286		  bool log_ecn_error);
287int ip_tunnel_changelink(struct net_device *dev, struct nlattr *tb[],
288			 struct ip_tunnel_parm *p, __u32 fwmark);
289int ip_tunnel_newlink(struct net_device *dev, struct nlattr *tb[],
290		      struct ip_tunnel_parm *p, __u32 fwmark);
291void ip_tunnel_setup(struct net_device *dev, unsigned int net_id);
292
293extern const struct header_ops ip_tunnel_header_ops;
294__be16 ip_tunnel_parse_protocol(const struct sk_buff *skb);
295
296struct ip_tunnel_encap_ops {
297	size_t (*encap_hlen)(struct ip_tunnel_encap *e);
298	int (*build_header)(struct sk_buff *skb, struct ip_tunnel_encap *e,
299			    u8 *protocol, struct flowi4 *fl4);
300	int (*err_handler)(struct sk_buff *skb, u32 info);
301};
302
303#define MAX_IPTUN_ENCAP_OPS 8
304
305extern const struct ip_tunnel_encap_ops __rcu *
306		iptun_encaps[MAX_IPTUN_ENCAP_OPS];
307
308int ip_tunnel_encap_add_ops(const struct ip_tunnel_encap_ops *op,
309			    unsigned int num);
310int ip_tunnel_encap_del_ops(const struct ip_tunnel_encap_ops *op,
311			    unsigned int num);
312
313int ip_tunnel_encap_setup(struct ip_tunnel *t,
314			  struct ip_tunnel_encap *ipencap);
315
316static inline bool pskb_inet_may_pull(struct sk_buff *skb)
317{
318	int nhlen;
319
320	switch (skb->protocol) {
321#if IS_ENABLED(CONFIG_IPV6)
322	case htons(ETH_P_IPV6):
323		nhlen = sizeof(struct ipv6hdr);
324		break;
325#endif
326	case htons(ETH_P_IP):
327		nhlen = sizeof(struct iphdr);
328		break;
329	default:
330		nhlen = 0;
331	}
332
333	return pskb_network_may_pull(skb, nhlen);
334}
335
336static inline int ip_encap_hlen(struct ip_tunnel_encap *e)
337{
338	const struct ip_tunnel_encap_ops *ops;
339	int hlen = -EINVAL;
340
341	if (e->type == TUNNEL_ENCAP_NONE)
342		return 0;
343
344	if (e->type >= MAX_IPTUN_ENCAP_OPS)
345		return -EINVAL;
346
347	rcu_read_lock();
348	ops = rcu_dereference(iptun_encaps[e->type]);
349	if (likely(ops && ops->encap_hlen))
350		hlen = ops->encap_hlen(e);
351	rcu_read_unlock();
352
353	return hlen;
354}
355
356static inline int ip_tunnel_encap(struct sk_buff *skb, struct ip_tunnel *t,
357				  u8 *protocol, struct flowi4 *fl4)
358{
359	const struct ip_tunnel_encap_ops *ops;
360	int ret = -EINVAL;
361
362	if (t->encap.type == TUNNEL_ENCAP_NONE)
363		return 0;
364
365	if (t->encap.type >= MAX_IPTUN_ENCAP_OPS)
366		return -EINVAL;
367
368	rcu_read_lock();
369	ops = rcu_dereference(iptun_encaps[t->encap.type]);
370	if (likely(ops && ops->build_header))
371		ret = ops->build_header(skb, &t->encap, protocol, fl4);
372	rcu_read_unlock();
373
374	return ret;
375}
376
377/* Extract dsfield from inner protocol */
378static inline u8 ip_tunnel_get_dsfield(const struct iphdr *iph,
379				       const struct sk_buff *skb)
380{
381	__be16 payload_protocol = skb_protocol(skb, true);
382
383	if (payload_protocol == htons(ETH_P_IP))
384		return iph->tos;
385	else if (payload_protocol == htons(ETH_P_IPV6))
386		return ipv6_get_dsfield((const struct ipv6hdr *)iph);
387	else
388		return 0;
389}
390
391static inline u8 ip_tunnel_get_ttl(const struct iphdr *iph,
392				       const struct sk_buff *skb)
393{
394	__be16 payload_protocol = skb_protocol(skb, true);
395
396	if (payload_protocol == htons(ETH_P_IP))
397		return iph->ttl;
398	else if (payload_protocol == htons(ETH_P_IPV6))
399		return ((const struct ipv6hdr *)iph)->hop_limit;
400	else
401		return 0;
402}
403
404/* Propogate ECN bits out */
405static inline u8 ip_tunnel_ecn_encap(u8 tos, const struct iphdr *iph,
406				     const struct sk_buff *skb)
407{
408	u8 inner = ip_tunnel_get_dsfield(iph, skb);
409
410	return INET_ECN_encapsulate(tos, inner);
411}
412
413int __iptunnel_pull_header(struct sk_buff *skb, int hdr_len,
414			   __be16 inner_proto, bool raw_proto, bool xnet);
415
416static inline int iptunnel_pull_header(struct sk_buff *skb, int hdr_len,
417				       __be16 inner_proto, bool xnet)
418{
419	return __iptunnel_pull_header(skb, hdr_len, inner_proto, false, xnet);
420}
421
422void iptunnel_xmit(struct sock *sk, struct rtable *rt, struct sk_buff *skb,
423		   __be32 src, __be32 dst, u8 proto,
424		   u8 tos, u8 ttl, __be16 df, bool xnet);
425struct metadata_dst *iptunnel_metadata_reply(struct metadata_dst *md,
426					     gfp_t flags);
427int skb_tunnel_check_pmtu(struct sk_buff *skb, struct dst_entry *encap_dst,
428			  int headroom, bool reply);
429
430int iptunnel_handle_offloads(struct sk_buff *skb, int gso_type_mask);
431
432static inline int iptunnel_pull_offloads(struct sk_buff *skb)
433{
434	if (skb_is_gso(skb)) {
435		int err;
436
437		err = skb_unclone(skb, GFP_ATOMIC);
438		if (unlikely(err))
439			return err;
440		skb_shinfo(skb)->gso_type &= ~(NETIF_F_GSO_ENCAP_ALL >>
441					       NETIF_F_GSO_SHIFT);
442	}
443
444	skb->encapsulation = 0;
445	return 0;
446}
447
448static inline void iptunnel_xmit_stats(struct net_device *dev, int pkt_len)
449{
450	if (pkt_len > 0) {
451		struct pcpu_sw_netstats *tstats = get_cpu_ptr(dev->tstats);
452
453		u64_stats_update_begin(&tstats->syncp);
454		tstats->tx_bytes += pkt_len;
455		tstats->tx_packets++;
456		u64_stats_update_end(&tstats->syncp);
457		put_cpu_ptr(tstats);
458		return;
459	}
460
461	if (pkt_len < 0) {
462		DEV_STATS_INC(dev, tx_errors);
463		DEV_STATS_INC(dev, tx_aborted_errors);
464	} else {
465		DEV_STATS_INC(dev, tx_dropped);
466	}
467}
468
469static inline void *ip_tunnel_info_opts(struct ip_tunnel_info *info)
470{
471	return info + 1;
472}
473
474static inline void ip_tunnel_info_opts_get(void *to,
475					   const struct ip_tunnel_info *info)
476{
477	memcpy(to, info + 1, info->options_len);
478}
479
480static inline void ip_tunnel_info_opts_set(struct ip_tunnel_info *info,
481					   const void *from, int len,
482					   __be16 flags)
483{
484	info->options_len = len;
485	if (len > 0) {
486		memcpy(ip_tunnel_info_opts(info), from, len);
487		info->key.tun_flags |= flags;
488	}
489}
490
491static inline struct ip_tunnel_info *lwt_tun_info(struct lwtunnel_state *lwtstate)
492{
493	return (struct ip_tunnel_info *)lwtstate->data;
494}
495
496DECLARE_STATIC_KEY_FALSE(ip_tunnel_metadata_cnt);
497
498/* Returns > 0 if metadata should be collected */
499static inline int ip_tunnel_collect_metadata(void)
500{
501	return static_branch_unlikely(&ip_tunnel_metadata_cnt);
502}
503
504void __init ip_tunnel_core_init(void);
505
506void ip_tunnel_need_metadata(void);
507void ip_tunnel_unneed_metadata(void);
508
509#else /* CONFIG_INET */
510
511static inline struct ip_tunnel_info *lwt_tun_info(struct lwtunnel_state *lwtstate)
512{
513	return NULL;
514}
515
516static inline void ip_tunnel_need_metadata(void)
517{
518}
519
520static inline void ip_tunnel_unneed_metadata(void)
521{
522}
523
524static inline void ip_tunnel_info_opts_get(void *to,
525					   const struct ip_tunnel_info *info)
526{
527}
528
529static inline void ip_tunnel_info_opts_set(struct ip_tunnel_info *info,
530					   const void *from, int len,
531					   __be16 flags)
532{
533	info->options_len = 0;
534}
535
536#endif /* CONFIG_INET */
537
538#endif /* __NET_IP_TUNNELS_H */
539