]>
Commit | Line | Data |
---|---|---|
b2441318 | 1 | /* SPDX-License-Identifier: GPL-2.0 */ |
5f256bec EB |
2 | /* |
3 | * Operations on the network namespace | |
4 | */ | |
5 | #ifndef __NET_NET_NAMESPACE_H | |
6 | #define __NET_NET_NAMESPACE_H | |
7 | ||
60063497 | 8 | #include <linux/atomic.h> |
c122e14d | 9 | #include <linux/refcount.h> |
5f256bec EB |
10 | #include <linux/workqueue.h> |
11 | #include <linux/list.h> | |
bee95250 | 12 | #include <linux/sysctl.h> |
fbdeaed4 | 13 | #include <linux/uidgid.h> |
5f256bec | 14 | |
6a662719 | 15 | #include <net/flow.h> |
8efa6e93 | 16 | #include <net/netns/core.h> |
852566f5 | 17 | #include <net/netns/mib.h> |
a0a53c8b | 18 | #include <net/netns/unix.h> |
2aaef4e4 | 19 | #include <net/netns/packet.h> |
8afd351c | 20 | #include <net/netns/ipv4.h> |
b0f159db | 21 | #include <net/netns/ipv6.h> |
ab84be7e | 22 | #include <net/netns/nexthop.h> |
633fc86f | 23 | #include <net/netns/ieee802154_6lowpan.h> |
4db67e80 | 24 | #include <net/netns/sctp.h> |
67019cc9 | 25 | #include <net/netns/dccp.h> |
f3c1a44a | 26 | #include <net/netns/netfilter.h> |
8d870052 | 27 | #include <net/netns/x_tables.h> |
dfdb8d79 AD |
28 | #if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE) |
29 | #include <net/netns/conntrack.h> | |
30 | #endif | |
99633ab2 | 31 | #include <net/netns/nftables.h> |
d62ddc21 | 32 | #include <net/netns/xfrm.h> |
0189197f | 33 | #include <net/netns/mpls.h> |
8e8cda6d | 34 | #include <net/netns/can.h> |
1d0dc069 | 35 | #include <net/netns/xdp.h> |
a3fd7cee | 36 | #include <net/netns/bpf.h> |
435d5f4b | 37 | #include <linux/ns_common.h> |
04c52dec PNA |
38 | #include <linux/idr.h> |
39 | #include <linux/skbuff.h> | |
a30c7b42 | 40 | #include <linux/notifier.h> |
a0a53c8b | 41 | |
038e7332 | 42 | struct user_namespace; |
457c4cbc | 43 | struct proc_dir_entry; |
2774c7ab | 44 | struct net_device; |
97c53cac | 45 | struct sock; |
1597fbc0 | 46 | struct ctl_table_header; |
dec827d1 | 47 | struct net_generic; |
94e5e308 | 48 | struct uevent_sock; |
2553d064 | 49 | struct netns_ipvs; |
d58e468b | 50 | struct bpf_prog; |
1597fbc0 | 51 | |
7c28bd0b ED |
52 | |
53 | #define NETDEV_HASHBITS 8 | |
54 | #define NETDEV_HASHENTRIES (1 << NETDEV_HASHBITS) | |
55 | ||
5f256bec | 56 | struct net { |
2a06b898 ED |
57 | /* First cache line can be often dirtied. |
58 | * Do not place here read-mostly fields. | |
59 | */ | |
aad12c23 | 60 | refcount_t passive; /* To decide when the network |
a685e089 AV |
61 | * namespace should be freed. |
62 | */ | |
8e602ce2 ED |
63 | spinlock_t rules_mod_lock; |
64 | ||
2a06b898 ED |
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; | |
33cf7c90 | 72 | |
5f256bec | 73 | struct list_head list; /* list of network namespaces */ |
19efbd93 | 74 | struct list_head exit_list; /* To linked to call pernet exit |
4420bf21 KT |
75 | * methods on dead net ( |
76 | * pernet_ops_rwsem read locked), | |
77 | * or to unregister pernet ops | |
78 | * (pernet_ops_rwsem write locked). | |
19efbd93 | 79 | */ |
65b7b5b9 KT |
80 | struct llist_node cleanup_list; /* namespaces on death row */ |
81 | ||
9b242610 DH |
82 | #ifdef CONFIG_KEYS |
83 | struct key_tag *key_domain; /* Key domain of operation tag */ | |
84 | #endif | |
038e7332 | 85 | struct user_namespace *user_ns; /* Owning user namespace */ |
70328660 | 86 | struct ucounts *ucounts; |
0c7aecd4 | 87 | struct idr netns_ids; |
038e7332 | 88 | |
435d5f4b | 89 | struct ns_common ns; |
98f842e6 | 90 | |
2a06b898 | 91 | struct list_head dev_base_head; |
457c4cbc EB |
92 | struct proc_dir_entry *proc_net; |
93 | struct proc_dir_entry *proc_net_stat; | |
881d966b | 94 | |
73455092 AV |
95 | #ifdef CONFIG_SYSCTL |
96 | struct ctl_table_set sysctls; | |
97 | #endif | |
95bdfccb | 98 | |
8e602ce2 ED |
99 | struct sock *rtnl; /* rtnetlink socket */ |
100 | struct sock *genl_sock; | |
2774c7ab | 101 | |
94e5e308 CB |
102 | struct uevent_sock *uevent_sock; /* uevent socket */ |
103 | ||
881d966b EB |
104 | struct hlist_head *dev_name_head; |
105 | struct hlist_head *dev_index_head; | |
a30c7b42 JP |
106 | struct raw_notifier_head netdev_chain; |
107 | ||
2a06b898 ED |
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 */ | |
97c53cac | 114 | |
5fd30ee7 DL |
115 | /* core fib_rules */ |
116 | struct list_head rules_ops; | |
5fd30ee7 | 117 | |
8efa6e93 | 118 | struct netns_core core; |
852566f5 | 119 | struct netns_mib mib; |
2aaef4e4 | 120 | struct netns_packet packet; |
a0a53c8b | 121 | struct netns_unix unx; |
ab84be7e | 122 | struct netns_nexthop nexthop; |
8afd351c | 123 | struct netns_ipv4 ipv4; |
dfd56b8b | 124 | #if IS_ENABLED(CONFIG_IPV6) |
b0f159db DL |
125 | struct netns_ipv6 ipv6; |
126 | #endif | |
633fc86f AA |
127 | #if IS_ENABLED(CONFIG_IEEE802154_6LOWPAN) |
128 | struct netns_ieee802154_lowpan ieee802154_lowpan; | |
129 | #endif | |
4db67e80 EB |
130 | #if defined(CONFIG_IP_SCTP) || defined(CONFIG_IP_SCTP_MODULE) |
131 | struct netns_sctp sctp; | |
132 | #endif | |
67019cc9 PE |
133 | #if defined(CONFIG_IP_DCCP) || defined(CONFIG_IP_DCCP_MODULE) |
134 | struct netns_dccp dccp; | |
135 | #endif | |
8d870052 | 136 | #ifdef CONFIG_NETFILTER |
f3c1a44a | 137 | struct netns_nf nf; |
8d870052 | 138 | struct netns_xt xt; |
dfdb8d79 AD |
139 | #if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE) |
140 | struct netns_ct ct; | |
c038a767 | 141 | #endif |
99633ab2 PNA |
142 | #if defined(CONFIG_NF_TABLES) || defined(CONFIG_NF_TABLES_MODULE) |
143 | struct netns_nftables nft; | |
144 | #endif | |
c038a767 AW |
145 | #if IS_ENABLED(CONFIG_NF_DEFRAG_IPV6) |
146 | struct netns_nf_frag nf_frag; | |
9ce7bc03 | 147 | struct ctl_table_header *nf_frag_frags_hdr; |
dfdb8d79 | 148 | #endif |
cd8c20b6 AD |
149 | struct sock *nfnl; |
150 | struct sock *nfnl_stash; | |
19576c94 PN |
151 | #if IS_ENABLED(CONFIG_NF_CT_NETLINK_TIMEOUT) |
152 | struct list_head nfct_timeout_list; | |
153 | #endif | |
d62ddc21 | 154 | #endif |
3d23e349 | 155 | #ifdef CONFIG_WEXT_CORE |
b333b3d2 | 156 | struct sk_buff_head wext_nlevents; |
8d870052 | 157 | #endif |
1c87733d | 158 | struct net_generic __rcu *gen; |
8e602ce2 | 159 | |
a3fd7cee JS |
160 | /* Used to store attached BPF programs */ |
161 | struct netns_bpf bpf; | |
d58e468b | 162 | |
8e602ce2 ED |
163 | /* Note : following structs are cache line aligned */ |
164 | #ifdef CONFIG_XFRM | |
165 | struct netns_xfrm xfrm; | |
166 | #endif | |
f318903c DB |
167 | |
168 | atomic64_t net_cookie; /* written once */ | |
169 | ||
8b4d14d8 | 170 | #if IS_ENABLED(CONFIG_IP_VS) |
61b1ab45 | 171 | struct netns_ipvs *ipvs; |
0189197f EB |
172 | #endif |
173 | #if IS_ENABLED(CONFIG_MPLS) | |
174 | struct netns_mpls mpls; | |
8e8cda6d MK |
175 | #endif |
176 | #if IS_ENABLED(CONFIG_CAN) | |
177 | struct netns_can can; | |
1d0dc069 BT |
178 | #endif |
179 | #ifdef CONFIG_XDP_SOCKETS | |
180 | struct netns_xdp xdp; | |
91b05a7e OM |
181 | #endif |
182 | #if IS_ENABLED(CONFIG_CRYPTO_USER) | |
183 | struct sock *crypto_nlsk; | |
8b4d14d8 | 184 | #endif |
51d7cccf | 185 | struct sock *diag_nlsk; |
3859a271 | 186 | } __randomize_layout; |
5f256bec | 187 | |
c0f39322 DL |
188 | #include <linux/seq_file_net.h> |
189 | ||
4fabcd71 | 190 | /* Init's network namespace */ |
5f256bec | 191 | extern struct net init_net; |
a4aa834a | 192 | |
d727abcb | 193 | #ifdef CONFIG_NET_NS |
e67e16ea JP |
194 | struct net *copy_net_ns(unsigned long flags, struct user_namespace *user_ns, |
195 | struct net *old_net); | |
225c0a01 | 196 | |
fbdeaed4 TH |
197 | void net_ns_get_ownership(const struct net *net, kuid_t *uid, kgid_t *gid); |
198 | ||
7866cc57 | 199 | void net_ns_barrier(void); |
d727abcb EB |
200 | #else /* CONFIG_NET_NS */ |
201 | #include <linux/sched.h> | |
202 | #include <linux/nsproxy.h> | |
038e7332 EB |
203 | static inline struct net *copy_net_ns(unsigned long flags, |
204 | struct user_namespace *user_ns, struct net *old_net) | |
9dd776b6 | 205 | { |
d727abcb EB |
206 | if (flags & CLONE_NEWNET) |
207 | return ERR_PTR(-EINVAL); | |
208 | return old_net; | |
9dd776b6 | 209 | } |
7866cc57 | 210 | |
fbdeaed4 TH |
211 | static inline void net_ns_get_ownership(const struct net *net, |
212 | kuid_t *uid, kgid_t *gid) | |
213 | { | |
214 | *uid = GLOBAL_ROOT_UID; | |
215 | *gid = GLOBAL_ROOT_GID; | |
216 | } | |
217 | ||
7866cc57 | 218 | static inline void net_ns_barrier(void) {} |
d727abcb | 219 | #endif /* CONFIG_NET_NS */ |
225c0a01 DL |
220 | |
221 | ||
222 | extern struct list_head net_namespace_list; | |
9dd776b6 | 223 | |
e67e16ea | 224 | struct net *get_net_ns_by_pid(pid_t pid); |
0f5258cd | 225 | struct net *get_net_ns_by_fd(int fd); |
30ffee84 | 226 | |
92acdc58 | 227 | u64 __net_gen_cookie(struct net *net); |
5a95cbb8 | 228 | |
535d3ae9 RK |
229 | #ifdef CONFIG_SYSCTL |
230 | void ipx_register_sysctl(void); | |
231 | void ipx_unregister_sysctl(void); | |
232 | #else | |
233 | #define ipx_register_sysctl() | |
234 | #define ipx_unregister_sysctl() | |
235 | #endif | |
236 | ||
d4655795 | 237 | #ifdef CONFIG_NET_NS |
e67e16ea | 238 | void __put_net(struct net *net); |
5f256bec EB |
239 | |
240 | static inline struct net *get_net(struct net *net) | |
241 | { | |
8b8f3e66 | 242 | refcount_inc(&net->ns.count); |
5f256bec EB |
243 | return net; |
244 | } | |
245 | ||
077130c0 EB |
246 | static inline struct net *maybe_get_net(struct net *net) |
247 | { | |
248 | /* Used when we know struct net exists but we | |
249 | * aren't guaranteed a previous reference count | |
250 | * exists. If the reference count is zero this | |
251 | * function fails and returns NULL. | |
252 | */ | |
8b8f3e66 | 253 | if (!refcount_inc_not_zero(&net->ns.count)) |
077130c0 EB |
254 | net = NULL; |
255 | return net; | |
256 | } | |
257 | ||
5f256bec EB |
258 | static inline void put_net(struct net *net) |
259 | { | |
8b8f3e66 | 260 | if (refcount_dec_and_test(&net->ns.count)) |
5f256bec EB |
261 | __put_net(net); |
262 | } | |
263 | ||
878628fb YH |
264 | static inline |
265 | int net_eq(const struct net *net1, const struct net *net2) | |
266 | { | |
267 | return net1 == net2; | |
268 | } | |
a685e089 | 269 | |
4ee806d5 DS |
270 | static inline int check_net(const struct net *net) |
271 | { | |
8b8f3e66 | 272 | return refcount_read(&net->ns.count) != 0; |
4ee806d5 DS |
273 | } |
274 | ||
e67e16ea | 275 | void net_drop_ns(void *); |
a685e089 | 276 | |
d4655795 | 277 | #else |
b9f75f45 | 278 | |
d4655795 PE |
279 | static inline struct net *get_net(struct net *net) |
280 | { | |
281 | return net; | |
282 | } | |
283 | ||
284 | static inline void put_net(struct net *net) | |
285 | { | |
286 | } | |
287 | ||
5d1e4468 DL |
288 | static inline struct net *maybe_get_net(struct net *net) |
289 | { | |
290 | return net; | |
291 | } | |
292 | ||
293 | static inline | |
294 | int net_eq(const struct net *net1, const struct net *net2) | |
295 | { | |
296 | return 1; | |
297 | } | |
a685e089 | 298 | |
4ee806d5 DS |
299 | static inline int check_net(const struct net *net) |
300 | { | |
301 | return 1; | |
302 | } | |
303 | ||
a685e089 | 304 | #define net_drop_ns NULL |
5d1e4468 DL |
305 | #endif |
306 | ||
307 | ||
0c5c9fb5 | 308 | typedef struct { |
8f424b5f | 309 | #ifdef CONFIG_NET_NS |
0c5c9fb5 EB |
310 | struct net *net; |
311 | #endif | |
312 | } possible_net_t; | |
8f424b5f | 313 | |
0c5c9fb5 | 314 | static inline void write_pnet(possible_net_t *pnet, struct net *net) |
8f424b5f | 315 | { |
0c5c9fb5 EB |
316 | #ifdef CONFIG_NET_NS |
317 | pnet->net = net; | |
318 | #endif | |
8f424b5f ED |
319 | } |
320 | ||
0c5c9fb5 | 321 | static inline struct net *read_pnet(const possible_net_t *pnet) |
8f424b5f | 322 | { |
0c5c9fb5 EB |
323 | #ifdef CONFIG_NET_NS |
324 | return pnet->net; | |
8f424b5f | 325 | #else |
0c5c9fb5 | 326 | return &init_net; |
8f424b5f | 327 | #endif |
0c5c9fb5 | 328 | } |
5d1e4468 | 329 | |
f0b07bb1 | 330 | /* Protected by net_rwsem */ |
5f256bec EB |
331 | #define for_each_net(VAR) \ |
332 | list_for_each_entry(VAR, &net_namespace_list, list) | |
afa0df59 JP |
333 | #define for_each_net_continue_reverse(VAR) \ |
334 | list_for_each_entry_continue_reverse(VAR, &net_namespace_list, list) | |
11a28d37 JB |
335 | #define for_each_net_rcu(VAR) \ |
336 | list_for_each_entry_rcu(VAR, &net_namespace_list, list) | |
337 | ||
4665079c PE |
338 | #ifdef CONFIG_NET_NS |
339 | #define __net_init | |
340 | #define __net_exit | |
022cbae6 | 341 | #define __net_initdata |
04a6f82c | 342 | #define __net_initconst |
4665079c PE |
343 | #else |
344 | #define __net_init __init | |
bd721ea7 | 345 | #define __net_exit __ref |
022cbae6 | 346 | #define __net_initdata __initdata |
04a6f82c | 347 | #define __net_initconst __initconst |
4665079c | 348 | #endif |
5f256bec | 349 | |
d4e4fdf9 | 350 | int peernet2id_alloc(struct net *net, struct net *peer, gfp_t gfp); |
56f200c7 GN |
351 | int peernet2id(const struct net *net, struct net *peer); |
352 | bool peernet_has_id(const struct net *net, struct net *peer); | |
353 | struct net *get_net_ns_by_id(const struct net *net, int id); | |
0c7aecd4 | 354 | |
5f256bec EB |
355 | struct pernet_operations { |
356 | struct list_head list; | |
6056415d KT |
357 | /* |
358 | * Below methods are called without any exclusive locks. | |
359 | * More than one net may be constructed and destructed | |
360 | * in parallel on several cpus. Every pernet_operations | |
361 | * have to keep in mind all other pernet_operations and | |
362 | * to introduce a locking, if they share common resources. | |
363 | * | |
8518e9bb KT |
364 | * The only time they are called with exclusive lock is |
365 | * from register_pernet_subsys(), unregister_pernet_subsys() | |
366 | * register_pernet_device() and unregister_pernet_device(). | |
367 | * | |
6056415d KT |
368 | * Exit methods using blocking RCU primitives, such as |
369 | * synchronize_rcu(), should be implemented via exit_batch. | |
370 | * Then, destruction of a group of net requires single | |
371 | * synchronize_rcu() related to these pernet_operations, | |
372 | * instead of separate synchronize_rcu() for every net. | |
373 | * Please, avoid synchronize_rcu() at all, where it's possible. | |
d7d99872 ED |
374 | * |
375 | * Note that a combination of pre_exit() and exit() can | |
376 | * be used, since a synchronize_rcu() is guaranteed between | |
377 | * the calls. | |
6056415d | 378 | */ |
5f256bec | 379 | int (*init)(struct net *net); |
d7d99872 | 380 | void (*pre_exit)(struct net *net); |
5f256bec | 381 | void (*exit)(struct net *net); |
72ad937a | 382 | void (*exit_batch)(struct list_head *net_exit_list); |
c7d03a00 | 383 | unsigned int *id; |
f875bae0 | 384 | size_t size; |
5f256bec EB |
385 | }; |
386 | ||
17edde52 EB |
387 | /* |
388 | * Use these carefully. If you implement a network device and it | |
389 | * needs per network namespace operations use device pernet operations, | |
390 | * otherwise use pernet subsys operations. | |
391 | * | |
4edf547b JB |
392 | * Network interfaces need to be removed from a dying netns _before_ |
393 | * subsys notifiers can be called, as most of the network code cleanup | |
394 | * (which is done from subsys notifiers) runs with the assumption that | |
395 | * dev_remove_pack has been called so no new packets will arrive during | |
396 | * and after the cleanup functions have been called. dev_remove_pack | |
397 | * is not per namespace so instead the guarantee of no more packets | |
398 | * arriving in a network namespace is provided by ensuring that all | |
399 | * network devices and all sockets have left the network namespace | |
400 | * before the cleanup methods are called. | |
17edde52 EB |
401 | * |
402 | * For the longest time the ipv4 icmp code was registered as a pernet | |
403 | * device which caused kernel oops, and panics during network | |
404 | * namespace cleanup. So please don't get this wrong. | |
405 | */ | |
e67e16ea JP |
406 | int register_pernet_subsys(struct pernet_operations *); |
407 | void unregister_pernet_subsys(struct pernet_operations *); | |
408 | int register_pernet_device(struct pernet_operations *); | |
409 | void unregister_pernet_device(struct pernet_operations *); | |
f875bae0 | 410 | |
95bdfccb EB |
411 | struct ctl_table; |
412 | struct ctl_table_header; | |
d62c612e | 413 | |
2ca794e5 | 414 | #ifdef CONFIG_SYSCTL |
e67e16ea JP |
415 | int net_sysctl_init(void); |
416 | struct ctl_table_header *register_net_sysctl(struct net *net, const char *path, | |
417 | struct ctl_table *table); | |
418 | void unregister_net_sysctl_table(struct ctl_table_header *header); | |
48c74958 EB |
419 | #else |
420 | static inline int net_sysctl_init(void) { return 0; } | |
421 | static inline struct ctl_table_header *register_net_sysctl(struct net *net, | |
422 | const char *path, struct ctl_table *table) | |
423 | { | |
424 | return NULL; | |
425 | } | |
426 | static inline void unregister_net_sysctl_table(struct ctl_table_header *header) | |
427 | { | |
428 | } | |
429 | #endif | |
430 | ||
56f200c7 | 431 | static inline int rt_genid_ipv4(const struct net *net) |
b42664f8 | 432 | { |
ca4c3fc2 | 433 | return atomic_read(&net->ipv4.rt_genid); |
b42664f8 ND |
434 | } |
435 | ||
8f34e53b DA |
436 | #if IS_ENABLED(CONFIG_IPV6) |
437 | static inline int rt_genid_ipv6(const struct net *net) | |
438 | { | |
439 | return atomic_read(&net->ipv6.fib6_sernum); | |
440 | } | |
441 | #endif | |
442 | ||
ca4c3fc2 | 443 | static inline void rt_genid_bump_ipv4(struct net *net) |
b42664f8 | 444 | { |
ca4c3fc2 | 445 | atomic_inc(&net->ipv4.rt_genid); |
446 | } | |
447 | ||
705f1c86 | 448 | extern void (*__fib6_flush_trees)(struct net *net); |
ca4c3fc2 | 449 | static inline void rt_genid_bump_ipv6(struct net *net) |
450 | { | |
705f1c86 HFS |
451 | if (__fib6_flush_trees) |
452 | __fib6_flush_trees(net); | |
ca4c3fc2 | 453 | } |
ca4c3fc2 | 454 | |
599018a7 LR |
455 | #if IS_ENABLED(CONFIG_IEEE802154_6LOWPAN) |
456 | static inline struct netns_ieee802154_lowpan * | |
457 | net_ieee802154_lowpan(struct net *net) | |
458 | { | |
459 | return &net->ieee802154_lowpan; | |
460 | } | |
599018a7 LR |
461 | #endif |
462 | ||
ca4c3fc2 | 463 | /* For callers who don't really care about whether it's IPv4 or IPv6 */ |
464 | static inline void rt_genid_bump_all(struct net *net) | |
465 | { | |
466 | rt_genid_bump_ipv4(net); | |
467 | rt_genid_bump_ipv6(net); | |
b42664f8 | 468 | } |
95bdfccb | 469 | |
56f200c7 | 470 | static inline int fnhe_genid(const struct net *net) |
5aad1de5 TT |
471 | { |
472 | return atomic_read(&net->fnhe_genid); | |
473 | } | |
474 | ||
475 | static inline void fnhe_genid_bump(struct net *net) | |
476 | { | |
477 | atomic_inc(&net->fnhe_genid); | |
478 | } | |
479 | ||
5f256bec | 480 | #endif /* __NET_NET_NAMESPACE_H */ |