Back to home page

OSCL-LXR

 
 

    


0001 /* SPDX-License-Identifier: GPL-2.0-or-later */
0002 /*
0003  * INET     An implementation of the TCP/IP protocol suite for the LINUX
0004  *      operating system.  INET is implemented using the  BSD Socket
0005  *      interface as the means of communication with the user level.
0006  *
0007  *      Definitions for the IP module.
0008  *
0009  * Version: @(#)ip.h    1.0.2   05/07/93
0010  *
0011  * Authors: Ross Biro
0012  *      Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
0013  *      Alan Cox, <gw4pts@gw4pts.ampr.org>
0014  *
0015  * Changes:
0016  *      Mike McLagan    :       Routing by source
0017  */
0018 #ifndef _IP_H
0019 #define _IP_H
0020 
0021 #include <linux/types.h>
0022 #include <linux/ip.h>
0023 #include <linux/in.h>
0024 #include <linux/skbuff.h>
0025 #include <linux/jhash.h>
0026 #include <linux/sockptr.h>
0027 #include <linux/static_key.h>
0028 
0029 #include <net/inet_sock.h>
0030 #include <net/route.h>
0031 #include <net/snmp.h>
0032 #include <net/flow.h>
0033 #include <net/flow_dissector.h>
0034 #include <net/netns/hash.h>
0035 #include <net/lwtunnel.h>
0036 
0037 #define IPV4_MAX_PMTU       65535U      /* RFC 2675, Section 5.1 */
0038 #define IPV4_MIN_MTU        68          /* RFC 791 */
0039 
0040 extern unsigned int sysctl_fib_sync_mem;
0041 extern unsigned int sysctl_fib_sync_mem_min;
0042 extern unsigned int sysctl_fib_sync_mem_max;
0043 
0044 struct sock;
0045 
0046 struct inet_skb_parm {
0047     int         iif;
0048     struct ip_options   opt;        /* Compiled IP options      */
0049     u16         flags;
0050 
0051 #define IPSKB_FORWARDED     BIT(0)
0052 #define IPSKB_XFRM_TUNNEL_SIZE  BIT(1)
0053 #define IPSKB_XFRM_TRANSFORMED  BIT(2)
0054 #define IPSKB_FRAG_COMPLETE BIT(3)
0055 #define IPSKB_REROUTED      BIT(4)
0056 #define IPSKB_DOREDIRECT    BIT(5)
0057 #define IPSKB_FRAG_PMTU     BIT(6)
0058 #define IPSKB_L3SLAVE       BIT(7)
0059 #define IPSKB_NOPOLICY      BIT(8)
0060 
0061     u16         frag_max_size;
0062 };
0063 
0064 static inline bool ipv4_l3mdev_skb(u16 flags)
0065 {
0066     return !!(flags & IPSKB_L3SLAVE);
0067 }
0068 
0069 static inline unsigned int ip_hdrlen(const struct sk_buff *skb)
0070 {
0071     return ip_hdr(skb)->ihl * 4;
0072 }
0073 
0074 struct ipcm_cookie {
0075     struct sockcm_cookie    sockc;
0076     __be32          addr;
0077     int         oif;
0078     struct ip_options_rcu   *opt;
0079     __u8            ttl;
0080     __s16           tos;
0081     char            priority;
0082     __u16           gso_size;
0083 };
0084 
0085 static inline void ipcm_init(struct ipcm_cookie *ipcm)
0086 {
0087     *ipcm = (struct ipcm_cookie) { .tos = -1 };
0088 }
0089 
0090 static inline void ipcm_init_sk(struct ipcm_cookie *ipcm,
0091                 const struct inet_sock *inet)
0092 {
0093     ipcm_init(ipcm);
0094 
0095     ipcm->sockc.mark = inet->sk.sk_mark;
0096     ipcm->sockc.tsflags = inet->sk.sk_tsflags;
0097     ipcm->oif = READ_ONCE(inet->sk.sk_bound_dev_if);
0098     ipcm->addr = inet->inet_saddr;
0099 }
0100 
0101 #define IPCB(skb) ((struct inet_skb_parm*)((skb)->cb))
0102 #define PKTINFO_SKB_CB(skb) ((struct in_pktinfo *)((skb)->cb))
0103 
0104 /* return enslaved device index if relevant */
0105 static inline int inet_sdif(const struct sk_buff *skb)
0106 {
0107 #if IS_ENABLED(CONFIG_NET_L3_MASTER_DEV)
0108     if (skb && ipv4_l3mdev_skb(IPCB(skb)->flags))
0109         return IPCB(skb)->iif;
0110 #endif
0111     return 0;
0112 }
0113 
0114 /* Special input handler for packets caught by router alert option.
0115    They are selected only by protocol field, and then processed likely
0116    local ones; but only if someone wants them! Otherwise, router
0117    not running rsvpd will kill RSVP.
0118 
0119    It is user level problem, what it will make with them.
0120    I have no idea, how it will masquearde or NAT them (it is joke, joke :-)),
0121    but receiver should be enough clever f.e. to forward mtrace requests,
0122    sent to multicast group to reach destination designated router.
0123  */
0124 
0125 struct ip_ra_chain {
0126     struct ip_ra_chain __rcu *next;
0127     struct sock     *sk;
0128     union {
0129         void            (*destructor)(struct sock *);
0130         struct sock     *saved_sk;
0131     };
0132     struct rcu_head     rcu;
0133 };
0134 
0135 /* IP flags. */
0136 #define IP_CE       0x8000      /* Flag: "Congestion"       */
0137 #define IP_DF       0x4000      /* Flag: "Don't Fragment"   */
0138 #define IP_MF       0x2000      /* Flag: "More Fragments"   */
0139 #define IP_OFFSET   0x1FFF      /* "Fragment Offset" part   */
0140 
0141 #define IP_FRAG_TIME    (30 * HZ)       /* fragment lifetime    */
0142 
0143 struct msghdr;
0144 struct net_device;
0145 struct packet_type;
0146 struct rtable;
0147 struct sockaddr;
0148 
0149 int igmp_mc_init(void);
0150 
0151 /*
0152  *  Functions provided by ip.c
0153  */
0154 
0155 int ip_build_and_send_pkt(struct sk_buff *skb, const struct sock *sk,
0156               __be32 saddr, __be32 daddr,
0157               struct ip_options_rcu *opt, u8 tos);
0158 int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt,
0159        struct net_device *orig_dev);
0160 void ip_list_rcv(struct list_head *head, struct packet_type *pt,
0161          struct net_device *orig_dev);
0162 int ip_local_deliver(struct sk_buff *skb);
0163 void ip_protocol_deliver_rcu(struct net *net, struct sk_buff *skb, int proto);
0164 int ip_mr_input(struct sk_buff *skb);
0165 int ip_output(struct net *net, struct sock *sk, struct sk_buff *skb);
0166 int ip_mc_output(struct net *net, struct sock *sk, struct sk_buff *skb);
0167 int ip_do_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
0168            int (*output)(struct net *, struct sock *, struct sk_buff *));
0169 
0170 struct ip_fraglist_iter {
0171     struct sk_buff  *frag;
0172     struct iphdr    *iph;
0173     int     offset;
0174     unsigned int    hlen;
0175 };
0176 
0177 void ip_fraglist_init(struct sk_buff *skb, struct iphdr *iph,
0178               unsigned int hlen, struct ip_fraglist_iter *iter);
0179 void ip_fraglist_prepare(struct sk_buff *skb, struct ip_fraglist_iter *iter);
0180 
0181 static inline struct sk_buff *ip_fraglist_next(struct ip_fraglist_iter *iter)
0182 {
0183     struct sk_buff *skb = iter->frag;
0184 
0185     iter->frag = skb->next;
0186     skb_mark_not_on_list(skb);
0187 
0188     return skb;
0189 }
0190 
0191 struct ip_frag_state {
0192     bool        DF;
0193     unsigned int    hlen;
0194     unsigned int    ll_rs;
0195     unsigned int    mtu;
0196     unsigned int    left;
0197     int     offset;
0198     int     ptr;
0199     __be16      not_last_frag;
0200 };
0201 
0202 void ip_frag_init(struct sk_buff *skb, unsigned int hlen, unsigned int ll_rs,
0203           unsigned int mtu, bool DF, struct ip_frag_state *state);
0204 struct sk_buff *ip_frag_next(struct sk_buff *skb,
0205                  struct ip_frag_state *state);
0206 
0207 void ip_send_check(struct iphdr *ip);
0208 int __ip_local_out(struct net *net, struct sock *sk, struct sk_buff *skb);
0209 int ip_local_out(struct net *net, struct sock *sk, struct sk_buff *skb);
0210 
0211 int __ip_queue_xmit(struct sock *sk, struct sk_buff *skb, struct flowi *fl,
0212             __u8 tos);
0213 void ip_init(void);
0214 int ip_append_data(struct sock *sk, struct flowi4 *fl4,
0215            int getfrag(void *from, char *to, int offset, int len,
0216                    int odd, struct sk_buff *skb),
0217            void *from, int len, int protolen,
0218            struct ipcm_cookie *ipc,
0219            struct rtable **rt,
0220            unsigned int flags);
0221 int ip_generic_getfrag(void *from, char *to, int offset, int len, int odd,
0222                struct sk_buff *skb);
0223 ssize_t ip_append_page(struct sock *sk, struct flowi4 *fl4, struct page *page,
0224                int offset, size_t size, int flags);
0225 struct sk_buff *__ip_make_skb(struct sock *sk, struct flowi4 *fl4,
0226                   struct sk_buff_head *queue,
0227                   struct inet_cork *cork);
0228 int ip_send_skb(struct net *net, struct sk_buff *skb);
0229 int ip_push_pending_frames(struct sock *sk, struct flowi4 *fl4);
0230 void ip_flush_pending_frames(struct sock *sk);
0231 struct sk_buff *ip_make_skb(struct sock *sk, struct flowi4 *fl4,
0232                 int getfrag(void *from, char *to, int offset,
0233                     int len, int odd, struct sk_buff *skb),
0234                 void *from, int length, int transhdrlen,
0235                 struct ipcm_cookie *ipc, struct rtable **rtp,
0236                 struct inet_cork *cork, unsigned int flags);
0237 
0238 int ip_queue_xmit(struct sock *sk, struct sk_buff *skb, struct flowi *fl);
0239 
0240 static inline struct sk_buff *ip_finish_skb(struct sock *sk, struct flowi4 *fl4)
0241 {
0242     return __ip_make_skb(sk, fl4, &sk->sk_write_queue, &inet_sk(sk)->cork.base);
0243 }
0244 
0245 static inline __u8 get_rttos(struct ipcm_cookie* ipc, struct inet_sock *inet)
0246 {
0247     return (ipc->tos != -1) ? RT_TOS(ipc->tos) : RT_TOS(inet->tos);
0248 }
0249 
0250 static inline __u8 get_rtconn_flags(struct ipcm_cookie* ipc, struct sock* sk)
0251 {
0252     return (ipc->tos != -1) ? RT_CONN_FLAGS_TOS(sk, ipc->tos) : RT_CONN_FLAGS(sk);
0253 }
0254 
0255 /* datagram.c */
0256 int __ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len);
0257 int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len);
0258 
0259 void ip4_datagram_release_cb(struct sock *sk);
0260 
0261 struct ip_reply_arg {
0262     struct kvec iov[1];
0263     int     flags;
0264     __wsum      csum;
0265     int     csumoffset; /* u16 offset of csum in iov[0].iov_base */
0266                 /* -1 if not needed */
0267     int     bound_dev_if;
0268     u8          tos;
0269     kuid_t      uid;
0270 };
0271 
0272 #define IP_REPLY_ARG_NOSRCCHECK 1
0273 
0274 static inline __u8 ip_reply_arg_flowi_flags(const struct ip_reply_arg *arg)
0275 {
0276     return (arg->flags & IP_REPLY_ARG_NOSRCCHECK) ? FLOWI_FLAG_ANYSRC : 0;
0277 }
0278 
0279 void ip_send_unicast_reply(struct sock *sk, struct sk_buff *skb,
0280                const struct ip_options *sopt,
0281                __be32 daddr, __be32 saddr,
0282                const struct ip_reply_arg *arg,
0283                unsigned int len, u64 transmit_time);
0284 
0285 #define IP_INC_STATS(net, field)    SNMP_INC_STATS64((net)->mib.ip_statistics, field)
0286 #define __IP_INC_STATS(net, field)  __SNMP_INC_STATS64((net)->mib.ip_statistics, field)
0287 #define IP_ADD_STATS(net, field, val)   SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val)
0288 #define __IP_ADD_STATS(net, field, val) __SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val)
0289 #define IP_UPD_PO_STATS(net, field, val) SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val)
0290 #define __IP_UPD_PO_STATS(net, field, val) __SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val)
0291 #define NET_INC_STATS(net, field)   SNMP_INC_STATS((net)->mib.net_statistics, field)
0292 #define __NET_INC_STATS(net, field) __SNMP_INC_STATS((net)->mib.net_statistics, field)
0293 #define NET_ADD_STATS(net, field, adnd) SNMP_ADD_STATS((net)->mib.net_statistics, field, adnd)
0294 #define __NET_ADD_STATS(net, field, adnd) __SNMP_ADD_STATS((net)->mib.net_statistics, field, adnd)
0295 
0296 static inline u64 snmp_get_cpu_field(void __percpu *mib, int cpu, int offt)
0297 {
0298     return  *(((unsigned long *)per_cpu_ptr(mib, cpu)) + offt);
0299 }
0300 
0301 unsigned long snmp_fold_field(void __percpu *mib, int offt);
0302 #if BITS_PER_LONG==32
0303 u64 snmp_get_cpu_field64(void __percpu *mib, int cpu, int offct,
0304              size_t syncp_offset);
0305 u64 snmp_fold_field64(void __percpu *mib, int offt, size_t sync_off);
0306 #else
0307 static inline u64  snmp_get_cpu_field64(void __percpu *mib, int cpu, int offct,
0308                     size_t syncp_offset)
0309 {
0310     return snmp_get_cpu_field(mib, cpu, offct);
0311 
0312 }
0313 
0314 static inline u64 snmp_fold_field64(void __percpu *mib, int offt, size_t syncp_off)
0315 {
0316     return snmp_fold_field(mib, offt);
0317 }
0318 #endif
0319 
0320 #define snmp_get_cpu_field64_batch(buff64, stats_list, mib_statistic, offset) \
0321 { \
0322     int i, c; \
0323     for_each_possible_cpu(c) { \
0324         for (i = 0; stats_list[i].name; i++) \
0325             buff64[i] += snmp_get_cpu_field64( \
0326                     mib_statistic, \
0327                     c, stats_list[i].entry, \
0328                     offset); \
0329     } \
0330 }
0331 
0332 #define snmp_get_cpu_field_batch(buff, stats_list, mib_statistic) \
0333 { \
0334     int i, c; \
0335     for_each_possible_cpu(c) { \
0336         for (i = 0; stats_list[i].name; i++) \
0337             buff[i] += snmp_get_cpu_field( \
0338                         mib_statistic, \
0339                         c, stats_list[i].entry); \
0340     } \
0341 }
0342 
0343 void inet_get_local_port_range(struct net *net, int *low, int *high);
0344 
0345 #ifdef CONFIG_SYSCTL
0346 static inline bool inet_is_local_reserved_port(struct net *net, unsigned short port)
0347 {
0348     if (!net->ipv4.sysctl_local_reserved_ports)
0349         return false;
0350     return test_bit(port, net->ipv4.sysctl_local_reserved_ports);
0351 }
0352 
0353 static inline bool sysctl_dev_name_is_allowed(const char *name)
0354 {
0355     return strcmp(name, "default") != 0  && strcmp(name, "all") != 0;
0356 }
0357 
0358 static inline bool inet_port_requires_bind_service(struct net *net, unsigned short port)
0359 {
0360     return port < READ_ONCE(net->ipv4.sysctl_ip_prot_sock);
0361 }
0362 
0363 #else
0364 static inline bool inet_is_local_reserved_port(struct net *net, unsigned short port)
0365 {
0366     return false;
0367 }
0368 
0369 static inline bool inet_port_requires_bind_service(struct net *net, unsigned short port)
0370 {
0371     return port < PROT_SOCK;
0372 }
0373 #endif
0374 
0375 __be32 inet_current_timestamp(void);
0376 
0377 /* From inetpeer.c */
0378 extern int inet_peer_threshold;
0379 extern int inet_peer_minttl;
0380 extern int inet_peer_maxttl;
0381 
0382 void ipfrag_init(void);
0383 
0384 void ip_static_sysctl_init(void);
0385 
0386 #define IP4_REPLY_MARK(net, mark) \
0387     (READ_ONCE((net)->ipv4.sysctl_fwmark_reflect) ? (mark) : 0)
0388 
0389 static inline bool ip_is_fragment(const struct iphdr *iph)
0390 {
0391     return (iph->frag_off & htons(IP_MF | IP_OFFSET)) != 0;
0392 }
0393 
0394 #ifdef CONFIG_INET
0395 #include <net/dst.h>
0396 
0397 /* The function in 2.2 was invalid, producing wrong result for
0398  * check=0xFEFF. It was noticed by Arthur Skawina _year_ ago. --ANK(000625) */
0399 static inline
0400 int ip_decrease_ttl(struct iphdr *iph)
0401 {
0402     u32 check = (__force u32)iph->check;
0403     check += (__force u32)htons(0x0100);
0404     iph->check = (__force __sum16)(check + (check>=0xFFFF));
0405     return --iph->ttl;
0406 }
0407 
0408 static inline int ip_mtu_locked(const struct dst_entry *dst)
0409 {
0410     const struct rtable *rt = (const struct rtable *)dst;
0411 
0412     return rt->rt_mtu_locked || dst_metric_locked(dst, RTAX_MTU);
0413 }
0414 
0415 static inline
0416 int ip_dont_fragment(const struct sock *sk, const struct dst_entry *dst)
0417 {
0418     u8 pmtudisc = READ_ONCE(inet_sk(sk)->pmtudisc);
0419 
0420     return  pmtudisc == IP_PMTUDISC_DO ||
0421         (pmtudisc == IP_PMTUDISC_WANT &&
0422          !ip_mtu_locked(dst));
0423 }
0424 
0425 static inline bool ip_sk_accept_pmtu(const struct sock *sk)
0426 {
0427     return inet_sk(sk)->pmtudisc != IP_PMTUDISC_INTERFACE &&
0428            inet_sk(sk)->pmtudisc != IP_PMTUDISC_OMIT;
0429 }
0430 
0431 static inline bool ip_sk_use_pmtu(const struct sock *sk)
0432 {
0433     return inet_sk(sk)->pmtudisc < IP_PMTUDISC_PROBE;
0434 }
0435 
0436 static inline bool ip_sk_ignore_df(const struct sock *sk)
0437 {
0438     return inet_sk(sk)->pmtudisc < IP_PMTUDISC_DO ||
0439            inet_sk(sk)->pmtudisc == IP_PMTUDISC_OMIT;
0440 }
0441 
0442 static inline unsigned int ip_dst_mtu_maybe_forward(const struct dst_entry *dst,
0443                             bool forwarding)
0444 {
0445     const struct rtable *rt = container_of(dst, struct rtable, dst);
0446     struct net *net = dev_net(dst->dev);
0447     unsigned int mtu;
0448 
0449     if (READ_ONCE(net->ipv4.sysctl_ip_fwd_use_pmtu) ||
0450         ip_mtu_locked(dst) ||
0451         !forwarding) {
0452         mtu = rt->rt_pmtu;
0453         if (mtu && time_before(jiffies, rt->dst.expires))
0454             goto out;
0455     }
0456 
0457     /* 'forwarding = true' case should always honour route mtu */
0458     mtu = dst_metric_raw(dst, RTAX_MTU);
0459     if (mtu)
0460         goto out;
0461 
0462     mtu = READ_ONCE(dst->dev->mtu);
0463 
0464     if (unlikely(ip_mtu_locked(dst))) {
0465         if (rt->rt_uses_gateway && mtu > 576)
0466             mtu = 576;
0467     }
0468 
0469 out:
0470     mtu = min_t(unsigned int, mtu, IP_MAX_MTU);
0471 
0472     return mtu - lwtunnel_headroom(dst->lwtstate, mtu);
0473 }
0474 
0475 static inline unsigned int ip_skb_dst_mtu(struct sock *sk,
0476                       const struct sk_buff *skb)
0477 {
0478     unsigned int mtu;
0479 
0480     if (!sk || !sk_fullsock(sk) || ip_sk_use_pmtu(sk)) {
0481         bool forwarding = IPCB(skb)->flags & IPSKB_FORWARDED;
0482 
0483         return ip_dst_mtu_maybe_forward(skb_dst(skb), forwarding);
0484     }
0485 
0486     mtu = min(READ_ONCE(skb_dst(skb)->dev->mtu), IP_MAX_MTU);
0487     return mtu - lwtunnel_headroom(skb_dst(skb)->lwtstate, mtu);
0488 }
0489 
0490 struct dst_metrics *ip_fib_metrics_init(struct net *net, struct nlattr *fc_mx,
0491                     int fc_mx_len,
0492                     struct netlink_ext_ack *extack);
0493 static inline void ip_fib_metrics_put(struct dst_metrics *fib_metrics)
0494 {
0495     if (fib_metrics != &dst_default_metrics &&
0496         refcount_dec_and_test(&fib_metrics->refcnt))
0497         kfree(fib_metrics);
0498 }
0499 
0500 /* ipv4 and ipv6 both use refcounted metrics if it is not the default */
0501 static inline
0502 void ip_dst_init_metrics(struct dst_entry *dst, struct dst_metrics *fib_metrics)
0503 {
0504     dst_init_metrics(dst, fib_metrics->metrics, true);
0505 
0506     if (fib_metrics != &dst_default_metrics) {
0507         dst->_metrics |= DST_METRICS_REFCOUNTED;
0508         refcount_inc(&fib_metrics->refcnt);
0509     }
0510 }
0511 
0512 static inline
0513 void ip_dst_metrics_put(struct dst_entry *dst)
0514 {
0515     struct dst_metrics *p = (struct dst_metrics *)DST_METRICS_PTR(dst);
0516 
0517     if (p != &dst_default_metrics && refcount_dec_and_test(&p->refcnt))
0518         kfree(p);
0519 }
0520 
0521 void __ip_select_ident(struct net *net, struct iphdr *iph, int segs);
0522 
0523 static inline void ip_select_ident_segs(struct net *net, struct sk_buff *skb,
0524                     struct sock *sk, int segs)
0525 {
0526     struct iphdr *iph = ip_hdr(skb);
0527 
0528     /* We had many attacks based on IPID, use the private
0529      * generator as much as we can.
0530      */
0531     if (sk && inet_sk(sk)->inet_daddr) {
0532         iph->id = htons(inet_sk(sk)->inet_id);
0533         inet_sk(sk)->inet_id += segs;
0534         return;
0535     }
0536     if ((iph->frag_off & htons(IP_DF)) && !skb->ignore_df) {
0537         iph->id = 0;
0538     } else {
0539         /* Unfortunately we need the big hammer to get a suitable IPID */
0540         __ip_select_ident(net, iph, segs);
0541     }
0542 }
0543 
0544 static inline void ip_select_ident(struct net *net, struct sk_buff *skb,
0545                    struct sock *sk)
0546 {
0547     ip_select_ident_segs(net, skb, sk, 1);
0548 }
0549 
0550 static inline __wsum inet_compute_pseudo(struct sk_buff *skb, int proto)
0551 {
0552     return csum_tcpudp_nofold(ip_hdr(skb)->saddr, ip_hdr(skb)->daddr,
0553                   skb->len, proto, 0);
0554 }
0555 
0556 /* copy IPv4 saddr & daddr to flow_keys, possibly using 64bit load/store
0557  * Equivalent to :  flow->v4addrs.src = iph->saddr;
0558  *          flow->v4addrs.dst = iph->daddr;
0559  */
0560 static inline void iph_to_flow_copy_v4addrs(struct flow_keys *flow,
0561                         const struct iphdr *iph)
0562 {
0563     BUILD_BUG_ON(offsetof(typeof(flow->addrs), v4addrs.dst) !=
0564              offsetof(typeof(flow->addrs), v4addrs.src) +
0565                   sizeof(flow->addrs.v4addrs.src));
0566     memcpy(&flow->addrs.v4addrs, &iph->saddr, sizeof(flow->addrs.v4addrs));
0567     flow->control.addr_type = FLOW_DISSECTOR_KEY_IPV4_ADDRS;
0568 }
0569 
0570 /*
0571  *  Map a multicast IP onto multicast MAC for type ethernet.
0572  */
0573 
0574 static inline void ip_eth_mc_map(__be32 naddr, char *buf)
0575 {
0576     __u32 addr=ntohl(naddr);
0577     buf[0]=0x01;
0578     buf[1]=0x00;
0579     buf[2]=0x5e;
0580     buf[5]=addr&0xFF;
0581     addr>>=8;
0582     buf[4]=addr&0xFF;
0583     addr>>=8;
0584     buf[3]=addr&0x7F;
0585 }
0586 
0587 /*
0588  *  Map a multicast IP onto multicast MAC for type IP-over-InfiniBand.
0589  *  Leave P_Key as 0 to be filled in by driver.
0590  */
0591 
0592 static inline void ip_ib_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
0593 {
0594     __u32 addr;
0595     unsigned char scope = broadcast[5] & 0xF;
0596 
0597     buf[0]  = 0;        /* Reserved */
0598     buf[1]  = 0xff;     /* Multicast QPN */
0599     buf[2]  = 0xff;
0600     buf[3]  = 0xff;
0601     addr    = ntohl(naddr);
0602     buf[4]  = 0xff;
0603     buf[5]  = 0x10 | scope; /* scope from broadcast address */
0604     buf[6]  = 0x40;     /* IPv4 signature */
0605     buf[7]  = 0x1b;
0606     buf[8]  = broadcast[8];     /* P_Key */
0607     buf[9]  = broadcast[9];
0608     buf[10] = 0;
0609     buf[11] = 0;
0610     buf[12] = 0;
0611     buf[13] = 0;
0612     buf[14] = 0;
0613     buf[15] = 0;
0614     buf[19] = addr & 0xff;
0615     addr  >>= 8;
0616     buf[18] = addr & 0xff;
0617     addr  >>= 8;
0618     buf[17] = addr & 0xff;
0619     addr  >>= 8;
0620     buf[16] = addr & 0x0f;
0621 }
0622 
0623 static inline void ip_ipgre_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
0624 {
0625     if ((broadcast[0] | broadcast[1] | broadcast[2] | broadcast[3]) != 0)
0626         memcpy(buf, broadcast, 4);
0627     else
0628         memcpy(buf, &naddr, sizeof(naddr));
0629 }
0630 
0631 #if IS_ENABLED(CONFIG_IPV6)
0632 #include <linux/ipv6.h>
0633 #endif
0634 
0635 static __inline__ void inet_reset_saddr(struct sock *sk)
0636 {
0637     inet_sk(sk)->inet_rcv_saddr = inet_sk(sk)->inet_saddr = 0;
0638 #if IS_ENABLED(CONFIG_IPV6)
0639     if (sk->sk_family == PF_INET6) {
0640         struct ipv6_pinfo *np = inet6_sk(sk);
0641 
0642         memset(&np->saddr, 0, sizeof(np->saddr));
0643         memset(&sk->sk_v6_rcv_saddr, 0, sizeof(sk->sk_v6_rcv_saddr));
0644     }
0645 #endif
0646 }
0647 
0648 #endif
0649 
0650 static inline unsigned int ipv4_addr_hash(__be32 ip)
0651 {
0652     return (__force unsigned int) ip;
0653 }
0654 
0655 static inline u32 ipv4_portaddr_hash(const struct net *net,
0656                      __be32 saddr,
0657                      unsigned int port)
0658 {
0659     return jhash_1word((__force u32)saddr, net_hash_mix(net)) ^ port;
0660 }
0661 
0662 bool ip_call_ra_chain(struct sk_buff *skb);
0663 
0664 /*
0665  *  Functions provided by ip_fragment.c
0666  */
0667 
0668 enum ip_defrag_users {
0669     IP_DEFRAG_LOCAL_DELIVER,
0670     IP_DEFRAG_CALL_RA_CHAIN,
0671     IP_DEFRAG_CONNTRACK_IN,
0672     __IP_DEFRAG_CONNTRACK_IN_END    = IP_DEFRAG_CONNTRACK_IN + USHRT_MAX,
0673     IP_DEFRAG_CONNTRACK_OUT,
0674     __IP_DEFRAG_CONNTRACK_OUT_END   = IP_DEFRAG_CONNTRACK_OUT + USHRT_MAX,
0675     IP_DEFRAG_CONNTRACK_BRIDGE_IN,
0676     __IP_DEFRAG_CONNTRACK_BRIDGE_IN = IP_DEFRAG_CONNTRACK_BRIDGE_IN + USHRT_MAX,
0677     IP_DEFRAG_VS_IN,
0678     IP_DEFRAG_VS_OUT,
0679     IP_DEFRAG_VS_FWD,
0680     IP_DEFRAG_AF_PACKET,
0681     IP_DEFRAG_MACVLAN,
0682 };
0683 
0684 /* Return true if the value of 'user' is between 'lower_bond'
0685  * and 'upper_bond' inclusively.
0686  */
0687 static inline bool ip_defrag_user_in_between(u32 user,
0688                          enum ip_defrag_users lower_bond,
0689                          enum ip_defrag_users upper_bond)
0690 {
0691     return user >= lower_bond && user <= upper_bond;
0692 }
0693 
0694 int ip_defrag(struct net *net, struct sk_buff *skb, u32 user);
0695 #ifdef CONFIG_INET
0696 struct sk_buff *ip_check_defrag(struct net *net, struct sk_buff *skb, u32 user);
0697 #else
0698 static inline struct sk_buff *ip_check_defrag(struct net *net, struct sk_buff *skb, u32 user)
0699 {
0700     return skb;
0701 }
0702 #endif
0703 
0704 /*
0705  *  Functions provided by ip_forward.c
0706  */
0707 
0708 int ip_forward(struct sk_buff *skb);
0709 
0710 /*
0711  *  Functions provided by ip_options.c
0712  */
0713 
0714 void ip_options_build(struct sk_buff *skb, struct ip_options *opt,
0715               __be32 daddr, struct rtable *rt);
0716 
0717 int __ip_options_echo(struct net *net, struct ip_options *dopt,
0718               struct sk_buff *skb, const struct ip_options *sopt);
0719 static inline int ip_options_echo(struct net *net, struct ip_options *dopt,
0720                   struct sk_buff *skb)
0721 {
0722     return __ip_options_echo(net, dopt, skb, &IPCB(skb)->opt);
0723 }
0724 
0725 void ip_options_fragment(struct sk_buff *skb);
0726 int __ip_options_compile(struct net *net, struct ip_options *opt,
0727              struct sk_buff *skb, __be32 *info);
0728 int ip_options_compile(struct net *net, struct ip_options *opt,
0729                struct sk_buff *skb);
0730 int ip_options_get(struct net *net, struct ip_options_rcu **optp,
0731            sockptr_t data, int optlen);
0732 void ip_options_undo(struct ip_options *opt);
0733 void ip_forward_options(struct sk_buff *skb);
0734 int ip_options_rcv_srr(struct sk_buff *skb, struct net_device *dev);
0735 
0736 /*
0737  *  Functions provided by ip_sockglue.c
0738  */
0739 
0740 void ipv4_pktinfo_prepare(const struct sock *sk, struct sk_buff *skb);
0741 void ip_cmsg_recv_offset(struct msghdr *msg, struct sock *sk,
0742              struct sk_buff *skb, int tlen, int offset);
0743 int ip_cmsg_send(struct sock *sk, struct msghdr *msg,
0744          struct ipcm_cookie *ipc, bool allow_ipv6);
0745 DECLARE_STATIC_KEY_FALSE(ip4_min_ttl);
0746 int ip_setsockopt(struct sock *sk, int level, int optname, sockptr_t optval,
0747           unsigned int optlen);
0748 int ip_getsockopt(struct sock *sk, int level, int optname, char __user *optval,
0749           int __user *optlen);
0750 int ip_ra_control(struct sock *sk, unsigned char on,
0751           void (*destructor)(struct sock *));
0752 
0753 int ip_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len);
0754 void ip_icmp_error(struct sock *sk, struct sk_buff *skb, int err, __be16 port,
0755            u32 info, u8 *payload);
0756 void ip_local_error(struct sock *sk, int err, __be32 daddr, __be16 dport,
0757             u32 info);
0758 
0759 static inline void ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb)
0760 {
0761     ip_cmsg_recv_offset(msg, skb->sk, skb, 0, 0);
0762 }
0763 
0764 bool icmp_global_allow(void);
0765 extern int sysctl_icmp_msgs_per_sec;
0766 extern int sysctl_icmp_msgs_burst;
0767 
0768 #ifdef CONFIG_PROC_FS
0769 int ip_misc_proc_init(void);
0770 #endif
0771 
0772 int rtm_getroute_parse_ip_proto(struct nlattr *attr, u8 *ip_proto, u8 family,
0773                 struct netlink_ext_ack *extack);
0774 
0775 static inline bool inetdev_valid_mtu(unsigned int mtu)
0776 {
0777     return likely(mtu >= IPV4_MIN_MTU);
0778 }
0779 
0780 void ip_sock_set_freebind(struct sock *sk);
0781 int ip_sock_set_mtu_discover(struct sock *sk, int val);
0782 void ip_sock_set_pktinfo(struct sock *sk);
0783 void ip_sock_set_recverr(struct sock *sk);
0784 void ip_sock_set_tos(struct sock *sk, int val);
0785 void  __ip_sock_set_tos(struct sock *sk, int val);
0786 
0787 #endif  /* _IP_H */