1 /* SPDX-License-Identifier: GPL-2.0 */ 2 /* 3 * Operations on the network namespace 4 */ 5 #ifndef __NET_NET_NAMESPACE_H 6 #define __NET_NET_NAMESPACE_H 7 8 #include <linux/atomic.h> 9 #include <linux/refcount.h> 10 #include <linux/workqueue.h> 11 #include <linux/list.h> 12 #include <linux/sysctl.h> 13 #include <linux/uidgid.h> 14 15 #include <net/flow.h> 16 #include <net/netns/core.h> 17 #include <net/netns/mib.h> 18 #include <net/netns/unix.h> 19 #include <net/netns/packet.h> 20 #include <net/netns/ipv4.h> 21 #include <net/netns/ipv6.h> 22 #include <net/netns/nexthop.h> 23 #include <net/netns/ieee802154_6lowpan.h> 24 #include <net/netns/sctp.h> 25 #include <net/netns/netfilter.h> 26 #include <net/netns/x_tables.h> 27 #if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE) 28 #include <net/netns/conntrack.h> 29 #endif 30 #include <net/netns/nftables.h> 31 #include <net/netns/xfrm.h> 32 #include <net/netns/mpls.h> 33 #include <net/netns/can.h> 34 #include <net/netns/xdp.h> 35 #include <net/netns/smc.h> 36 #include <net/netns/bpf.h> 37 #include <linux/ns_common.h> 38 #include <linux/idr.h> 39 #include <linux/skbuff.h> 40 #include <linux/notifier.h> 41 42 struct user_namespace; 43 struct proc_dir_entry; 44 struct net_device; 45 struct sock; 46 struct ctl_table_header; 47 struct net_generic; 48 struct uevent_sock; 49 struct netns_ipvs; 50 struct bpf_prog; 51 52 53 #define NETDEV_HASHBITS 8 54 #define NETDEV_HASHENTRIES (1 << NETDEV_HASHBITS) 55 56 struct net { 57 /* First cache line can be often dirtied. 58 * Do not place here read-mostly fields. 59 */ 60 refcount_t passive; /* To decide when the network 61 * namespace should be freed. 62 */ 63 spinlock_t rules_mod_lock; 64 65 unsigned int dev_unreg_count; 66 67 unsigned int dev_base_seq; /* protected by rtnl_mutex */ 68 int ifindex; 69 70 spinlock_t nsid_lock; 71 atomic_t fnhe_genid; 72 73 struct list_head list; /* list of network namespaces */ 74 struct list_head exit_list; /* To linked to call pernet exit 75 * methods on dead net ( 76 * pernet_ops_rwsem read locked), 77 * or to unregister pernet ops 78 * (pernet_ops_rwsem write locked). 79 */ 80 struct llist_node cleanup_list; /* namespaces on death row */ 81 82 #ifdef CONFIG_KEYS 83 struct key_tag *key_domain; /* Key domain of operation tag */ 84 #endif 85 struct user_namespace *user_ns; /* Owning user namespace */ 86 struct ucounts *ucounts; 87 struct idr netns_ids; 88 89 struct ns_common ns; 90 91 struct list_head dev_base_head; 92 struct proc_dir_entry *proc_net; 93 struct proc_dir_entry *proc_net_stat; 94 95 #ifdef CONFIG_SYSCTL 96 struct ctl_table_set sysctls; 97 #endif 98 99 struct sock *rtnl; /* rtnetlink socket */ 100 struct sock *genl_sock; 101 102 struct uevent_sock *uevent_sock; /* uevent socket */ 103 104 struct hlist_head *dev_name_head; 105 struct hlist_head *dev_index_head; 106 struct raw_notifier_head netdev_chain; 107 108 /* Note that @hash_mix can be read millions times per second, 109 * it is critical that it is on a read_mostly cache line. 110 */ 111 u32 hash_mix; 112 113 struct net_device *loopback_dev; /* The loopback */ 114 115 /* core fib_rules */ 116 struct list_head rules_ops; 117 118 struct netns_core core; 119 struct netns_mib mib; 120 struct netns_packet packet; 121 struct netns_unix unx; 122 struct netns_nexthop nexthop; 123 struct netns_ipv4 ipv4; 124 #if IS_ENABLED(CONFIG_IPV6) 125 struct netns_ipv6 ipv6; 126 #endif 127 #if IS_ENABLED(CONFIG_IEEE802154_6LOWPAN) 128 struct netns_ieee802154_lowpan ieee802154_lowpan; 129 #endif 130 #if defined(CONFIG_IP_SCTP) || defined(CONFIG_IP_SCTP_MODULE) 131 struct netns_sctp sctp; 132 #endif 133 #ifdef CONFIG_NETFILTER 134 struct netns_nf nf; 135 struct netns_xt xt; 136 #if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE) 137 struct netns_ct ct; 138 #endif 139 #if defined(CONFIG_NF_TABLES) || defined(CONFIG_NF_TABLES_MODULE) 140 struct netns_nftables nft; 141 #endif 142 #endif 143 #ifdef CONFIG_WEXT_CORE 144 struct sk_buff_head wext_nlevents; 145 #endif 146 struct net_generic __rcu *gen; 147 148 /* Used to store attached BPF programs */ 149 struct netns_bpf bpf; 150 151 /* Note : following structs are cache line aligned */ 152 #ifdef CONFIG_XFRM 153 struct netns_xfrm xfrm; 154 #endif 155 156 u64 net_cookie; /* written once */ 157 158 #if IS_ENABLED(CONFIG_IP_VS) 159 struct netns_ipvs *ipvs; 160 #endif 161 #if IS_ENABLED(CONFIG_MPLS) 162 struct netns_mpls mpls; 163 #endif 164 #if IS_ENABLED(CONFIG_CAN) 165 struct netns_can can; 166 #endif 167 #ifdef CONFIG_XDP_SOCKETS 168 struct netns_xdp xdp; 169 #endif 170 #if IS_ENABLED(CONFIG_CRYPTO_USER) 171 struct sock *crypto_nlsk; 172 #endif 173 struct sock *diag_nlsk; 174 #if IS_ENABLED(CONFIG_SMC) 175 struct netns_smc smc; 176 #endif 177 } __randomize_layout; 178 179 #include <linux/seq_file_net.h> 180 181 /* Init's network namespace */ 182 extern struct net init_net; 183 184 #ifdef CONFIG_NET_NS 185 struct net *copy_net_ns(unsigned long flags, struct user_namespace *user_ns, 186 struct net *old_net); 187 188 void net_ns_get_ownership(const struct net *net, kuid_t *uid, kgid_t *gid); 189 190 void net_ns_barrier(void); 191 #else /* CONFIG_NET_NS */ 192 #include <linux/sched.h> 193 #include <linux/nsproxy.h> 194 static inline struct net *copy_net_ns(unsigned long flags, 195 struct user_namespace *user_ns, struct net *old_net) 196 { 197 if (flags & CLONE_NEWNET) 198 return ERR_PTR(-EINVAL); 199 return old_net; 200 } 201 202 static inline void net_ns_get_ownership(const struct net *net, 203 kuid_t *uid, kgid_t *gid) 204 { 205 *uid = GLOBAL_ROOT_UID; 206 *gid = GLOBAL_ROOT_GID; 207 } 208 209 static inline void net_ns_barrier(void) {} 210 #endif /* CONFIG_NET_NS */ 211 212 213 extern struct list_head net_namespace_list; 214 215 struct net *get_net_ns_by_pid(pid_t pid); 216 struct net *get_net_ns_by_fd(int fd); 217 218 #ifdef CONFIG_SYSCTL 219 void ipx_register_sysctl(void); 220 void ipx_unregister_sysctl(void); 221 #else 222 #define ipx_register_sysctl() 223 #define ipx_unregister_sysctl() 224 #endif 225 226 #ifdef CONFIG_NET_NS 227 void __put_net(struct net *net); 228 229 static inline struct net *get_net(struct net *net) 230 { 231 refcount_inc(&net->ns.count); 232 return net; 233 } 234 235 static inline struct net *maybe_get_net(struct net *net) 236 { 237 /* Used when we know struct net exists but we 238 * aren't guaranteed a previous reference count 239 * exists. If the reference count is zero this 240 * function fails and returns NULL. 241 */ 242 if (!refcount_inc_not_zero(&net->ns.count)) 243 net = NULL; 244 return net; 245 } 246 247 static inline void put_net(struct net *net) 248 { 249 if (refcount_dec_and_test(&net->ns.count)) 250 __put_net(net); 251 } 252 253 static inline 254 int net_eq(const struct net *net1, const struct net *net2) 255 { 256 return net1 == net2; 257 } 258 259 static inline int check_net(const struct net *net) 260 { 261 return refcount_read(&net->ns.count) != 0; 262 } 263 264 void net_drop_ns(void *); 265 266 #else 267 268 static inline struct net *get_net(struct net *net) 269 { 270 return net; 271 } 272 273 static inline void put_net(struct net *net) 274 { 275 } 276 277 static inline struct net *maybe_get_net(struct net *net) 278 { 279 return net; 280 } 281 282 static inline 283 int net_eq(const struct net *net1, const struct net *net2) 284 { 285 return 1; 286 } 287 288 static inline int check_net(const struct net *net) 289 { 290 return 1; 291 } 292 293 #define net_drop_ns NULL 294 #endif 295 296 297 typedef struct { 298 #ifdef CONFIG_NET_NS 299 struct net *net; 300 #endif 301 } possible_net_t; 302 303 static inline void write_pnet(possible_net_t *pnet, struct net *net) 304 { 305 #ifdef CONFIG_NET_NS 306 pnet->net = net; 307 #endif 308 } 309 310 static inline struct net *read_pnet(const possible_net_t *pnet) 311 { 312 #ifdef CONFIG_NET_NS 313 return pnet->net; 314 #else 315 return &init_net; 316 #endif 317 } 318 319 /* Protected by net_rwsem */ 320 #define for_each_net(VAR) \ 321 list_for_each_entry(VAR, &net_namespace_list, list) 322 #define for_each_net_continue_reverse(VAR) \ 323 list_for_each_entry_continue_reverse(VAR, &net_namespace_list, list) 324 #define for_each_net_rcu(VAR) \ 325 list_for_each_entry_rcu(VAR, &net_namespace_list, list) 326 327 #ifdef CONFIG_NET_NS 328 #define __net_init 329 #define __net_exit 330 #define __net_initdata 331 #define __net_initconst 332 #else 333 #define __net_init __init 334 #define __net_exit __ref 335 #define __net_initdata __initdata 336 #define __net_initconst __initconst 337 #endif 338 339 int peernet2id_alloc(struct net *net, struct net *peer, gfp_t gfp); 340 int peernet2id(const struct net *net, struct net *peer); 341 bool peernet_has_id(const struct net *net, struct net *peer); 342 struct net *get_net_ns_by_id(const struct net *net, int id); 343 344 struct pernet_operations { 345 struct list_head list; 346 /* 347 * Below methods are called without any exclusive locks. 348 * More than one net may be constructed and destructed 349 * in parallel on several cpus. Every pernet_operations 350 * have to keep in mind all other pernet_operations and 351 * to introduce a locking, if they share common resources. 352 * 353 * The only time they are called with exclusive lock is 354 * from register_pernet_subsys(), unregister_pernet_subsys() 355 * register_pernet_device() and unregister_pernet_device(). 356 * 357 * Exit methods using blocking RCU primitives, such as 358 * synchronize_rcu(), should be implemented via exit_batch. 359 * Then, destruction of a group of net requires single 360 * synchronize_rcu() related to these pernet_operations, 361 * instead of separate synchronize_rcu() for every net. 362 * Please, avoid synchronize_rcu() at all, where it's possible. 363 * 364 * Note that a combination of pre_exit() and exit() can 365 * be used, since a synchronize_rcu() is guaranteed between 366 * the calls. 367 */ 368 int (*init)(struct net *net); 369 void (*pre_exit)(struct net *net); 370 void (*exit)(struct net *net); 371 void (*exit_batch)(struct list_head *net_exit_list); 372 unsigned int *id; 373 size_t size; 374 }; 375 376 /* 377 * Use these carefully. If you implement a network device and it 378 * needs per network namespace operations use device pernet operations, 379 * otherwise use pernet subsys operations. 380 * 381 * Network interfaces need to be removed from a dying netns _before_ 382 * subsys notifiers can be called, as most of the network code cleanup 383 * (which is done from subsys notifiers) runs with the assumption that 384 * dev_remove_pack has been called so no new packets will arrive during 385 * and after the cleanup functions have been called. dev_remove_pack 386 * is not per namespace so instead the guarantee of no more packets 387 * arriving in a network namespace is provided by ensuring that all 388 * network devices and all sockets have left the network namespace 389 * before the cleanup methods are called. 390 * 391 * For the longest time the ipv4 icmp code was registered as a pernet 392 * device which caused kernel oops, and panics during network 393 * namespace cleanup. So please don't get this wrong. 394 */ 395 int register_pernet_subsys(struct pernet_operations *); 396 void unregister_pernet_subsys(struct pernet_operations *); 397 int register_pernet_device(struct pernet_operations *); 398 void unregister_pernet_device(struct pernet_operations *); 399 400 struct ctl_table; 401 402 #ifdef CONFIG_SYSCTL 403 int net_sysctl_init(void); 404 struct ctl_table_header *register_net_sysctl(struct net *net, const char *path, 405 struct ctl_table *table); 406 void unregister_net_sysctl_table(struct ctl_table_header *header); 407 #else 408 static inline int net_sysctl_init(void) { return 0; } 409 static inline struct ctl_table_header *register_net_sysctl(struct net *net, 410 const char *path, struct ctl_table *table) 411 { 412 return NULL; 413 } 414 static inline void unregister_net_sysctl_table(struct ctl_table_header *header) 415 { 416 } 417 #endif 418 419 static inline int rt_genid_ipv4(const struct net *net) 420 { 421 return atomic_read(&net->ipv4.rt_genid); 422 } 423 424 #if IS_ENABLED(CONFIG_IPV6) 425 static inline int rt_genid_ipv6(const struct net *net) 426 { 427 return atomic_read(&net->ipv6.fib6_sernum); 428 } 429 #endif 430 431 static inline void rt_genid_bump_ipv4(struct net *net) 432 { 433 atomic_inc(&net->ipv4.rt_genid); 434 } 435 436 extern void (*__fib6_flush_trees)(struct net *net); 437 static inline void rt_genid_bump_ipv6(struct net *net) 438 { 439 if (__fib6_flush_trees) 440 __fib6_flush_trees(net); 441 } 442 443 #if IS_ENABLED(CONFIG_IEEE802154_6LOWPAN) 444 static inline struct netns_ieee802154_lowpan * 445 net_ieee802154_lowpan(struct net *net) 446 { 447 return &net->ieee802154_lowpan; 448 } 449 #endif 450 451 /* For callers who don't really care about whether it's IPv4 or IPv6 */ 452 static inline void rt_genid_bump_all(struct net *net) 453 { 454 rt_genid_bump_ipv4(net); 455 rt_genid_bump_ipv6(net); 456 } 457 458 static inline int fnhe_genid(const struct net *net) 459 { 460 return atomic_read(&net->fnhe_genid); 461 } 462 463 static inline void fnhe_genid_bump(struct net *net) 464 { 465 atomic_inc(&net->fnhe_genid); 466 } 467 468 #endif /* __NET_NET_NAMESPACE_H */ 469