Blame view
net/mptcp/protocol.h
11.5 KB
f870fa0b5
|
1 2 3 4 5 6 7 8 |
/* SPDX-License-Identifier: GPL-2.0 */ /* Multipath TCP * * Copyright (c) 2017 - 2019, Intel Corporation. */ #ifndef __MPTCP_PROTOCOL_H #define __MPTCP_PROTOCOL_H |
79c0949e9
|
9 10 11 |
#include <linux/random.h> #include <net/tcp.h> #include <net/inet_connection_sock.h> |
cc7972ea1
|
12 |
#define MPTCP_SUPPORTED_VERSION 1 |
eda7acddf
|
13 14 15 16 17 |
/* MPTCP option bits */ #define OPTION_MPTCP_MPC_SYN BIT(0) #define OPTION_MPTCP_MPC_SYNACK BIT(1) #define OPTION_MPTCP_MPC_ACK BIT(2) |
f296234c9
|
18 19 20 |
#define OPTION_MPTCP_MPJ_SYN BIT(3) #define OPTION_MPTCP_MPJ_SYNACK BIT(4) #define OPTION_MPTCP_MPJ_ACK BIT(5) |
3df523ab5
|
21 22 23 |
#define OPTION_MPTCP_ADD_ADDR BIT(6) #define OPTION_MPTCP_ADD_ADDR6 BIT(7) #define OPTION_MPTCP_RM_ADDR BIT(8) |
eda7acddf
|
24 25 26 27 28 29 30 31 32 33 34 35 |
/* MPTCP option subtypes */ #define MPTCPOPT_MP_CAPABLE 0 #define MPTCPOPT_MP_JOIN 1 #define MPTCPOPT_DSS 2 #define MPTCPOPT_ADD_ADDR 3 #define MPTCPOPT_RM_ADDR 4 #define MPTCPOPT_MP_PRIO 5 #define MPTCPOPT_MP_FAIL 6 #define MPTCPOPT_MP_FASTCLOSE 7 /* MPTCP suboption lengths */ |
cc7972ea1
|
36 |
#define TCPOLEN_MPTCP_MPC_SYN 4 |
eda7acddf
|
37 38 |
#define TCPOLEN_MPTCP_MPC_SYNACK 12 #define TCPOLEN_MPTCP_MPC_ACK 20 |
cc7972ea1
|
39 |
#define TCPOLEN_MPTCP_MPC_ACK_DATA 22 |
f296234c9
|
40 41 42 |
#define TCPOLEN_MPTCP_MPJ_SYN 12 #define TCPOLEN_MPTCP_MPJ_SYNACK 16 #define TCPOLEN_MPTCP_MPJ_ACK 24 |
6d0060f60
|
43 |
#define TCPOLEN_MPTCP_DSS_BASE 4 |
648ef4b88
|
44 |
#define TCPOLEN_MPTCP_DSS_ACK32 4 |
6d0060f60
|
45 |
#define TCPOLEN_MPTCP_DSS_ACK64 8 |
648ef4b88
|
46 |
#define TCPOLEN_MPTCP_DSS_MAP32 10 |
6d0060f60
|
47 48 |
#define TCPOLEN_MPTCP_DSS_MAP64 14 #define TCPOLEN_MPTCP_DSS_CHECKSUM 2 |
3df523ab5
|
49 50 51 52 53 54 55 56 57 58 |
#define TCPOLEN_MPTCP_ADD_ADDR 16 #define TCPOLEN_MPTCP_ADD_ADDR_PORT 18 #define TCPOLEN_MPTCP_ADD_ADDR_BASE 8 #define TCPOLEN_MPTCP_ADD_ADDR_BASE_PORT 10 #define TCPOLEN_MPTCP_ADD_ADDR6 28 #define TCPOLEN_MPTCP_ADD_ADDR6_PORT 30 #define TCPOLEN_MPTCP_ADD_ADDR6_BASE 20 #define TCPOLEN_MPTCP_ADD_ADDR6_BASE_PORT 22 #define TCPOLEN_MPTCP_PORT_LEN 2 #define TCPOLEN_MPTCP_RM_ADDR_BASE 4 |
eda7acddf
|
59 |
|
ec3edaa7c
|
60 |
/* MPTCP MP_JOIN flags */ |
f296234c9
|
61 62 |
#define MPTCPOPT_BACKUP BIT(0) #define MPTCPOPT_HMAC_LEN 20 |
ec3edaa7c
|
63 |
#define MPTCPOPT_THMAC_LEN 8 |
f296234c9
|
64 |
|
eda7acddf
|
65 66 67 68 |
/* MPTCP MP_CAPABLE flags */ #define MPTCP_VERSION_MASK (0x0F) #define MPTCP_CAP_CHECKSUM_REQD BIT(7) #define MPTCP_CAP_EXTENSIBILITY BIT(6) |
65492c5a6
|
69 |
#define MPTCP_CAP_HMAC_SHA256 BIT(0) |
eda7acddf
|
70 |
#define MPTCP_CAP_FLAG_MASK (0x3F) |
6d0060f60
|
71 72 73 74 75 76 |
/* MPTCP DSS flags */ #define MPTCP_DSS_DATA_FIN BIT(4) #define MPTCP_DSS_DSN64 BIT(3) #define MPTCP_DSS_HAS_MAP BIT(2) #define MPTCP_DSS_ACK64 BIT(1) #define MPTCP_DSS_HAS_ACK BIT(0) |
648ef4b88
|
77 |
#define MPTCP_DSS_FLAG_MASK (0x1F) |
3df523ab5
|
78 79 80 81 82 |
/* MPTCP ADD_ADDR flags */ #define MPTCP_ADDR_ECHO BIT(0) #define MPTCP_ADDR_HMAC_LEN 20 #define MPTCP_ADDR_IPVERSION_4 4 #define MPTCP_ADDR_IPVERSION_6 6 |
648ef4b88
|
83 |
/* MPTCP socket flags */ |
d99bfed58
|
84 85 |
#define MPTCP_DATA_READY 0 #define MPTCP_SEND_SPACE 1 |
3b1d6210a
|
86 |
#define MPTCP_WORK_RTX 2 |
59832e246
|
87 |
#define MPTCP_WORK_EOF 3 |
6d0060f60
|
88 |
|
3df523ab5
|
89 90 91 92 93 |
static inline __be32 mptcp_option(u8 subopt, u8 len, u8 nib, u8 field) { return htonl((TCPOPT_MPTCP << 24) | (len << 16) | (subopt << 12) | ((nib & 0xF) << 8) | field); } |
1b1c7a0ef
|
94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 |
#define MPTCP_PM_MAX_ADDR 4 struct mptcp_addr_info { sa_family_t family; __be16 port; u8 id; union { struct in_addr addr; #if IS_ENABLED(CONFIG_MPTCP_IPV6) struct in6_addr addr6; #endif }; }; enum mptcp_pm_status { MPTCP_PM_ADD_ADDR_RECEIVED, MPTCP_PM_ESTABLISHED, MPTCP_PM_SUBFLOW_ESTABLISHED, }; struct mptcp_pm_data { struct mptcp_addr_info local; struct mptcp_addr_info remote; spinlock_t lock; /*protects the whole PM data */ bool addr_signal; bool server_side; bool work_pending; bool accept_addr; bool accept_subflow; u8 add_addr_signaled; u8 add_addr_accepted; u8 local_addr_used; u8 subflows; u8 add_addr_signal_max; u8 add_addr_accept_max; u8 local_addr_max; u8 subflows_max; u8 status; struct work_struct work; }; |
18b683bff
|
137 138 139 140 141 142 143 144 |
struct mptcp_data_frag { struct list_head list; u64 data_seq; int data_len; int offset; int overhead; struct page *page; }; |
f870fa0b5
|
145 146 147 148 |
/* MPTCP connection sock */ struct mptcp_sock { /* inet_connection_sock must be the first member */ struct inet_connection_sock sk; |
cec37a6e4
|
149 150 |
u64 local_key; u64 remote_key; |
6d0060f60
|
151 152 |
u64 write_seq; u64 ack_seq; |
cc9d25669
|
153 |
atomic64_t snd_una; |
b51f9b80c
|
154 |
unsigned long timer_ival; |
79c0949e9
|
155 |
u32 token; |
648ef4b88
|
156 |
unsigned long flags; |
d22f4988f
|
157 |
bool can_ack; |
ec3edaa7c
|
158 |
spinlock_t join_list_lock; |
809920171
|
159 |
struct work_struct work; |
cec37a6e4
|
160 |
struct list_head conn_list; |
18b683bff
|
161 |
struct list_head rtx_queue; |
ec3edaa7c
|
162 |
struct list_head join_list; |
6d0060f60
|
163 |
struct skb_ext *cached_ext; /* for the next sendmsg */ |
f870fa0b5
|
164 |
struct socket *subflow; /* outgoing connect/listener/!mp_capable */ |
8ab183deb
|
165 |
struct sock *first; |
1b1c7a0ef
|
166 |
struct mptcp_pm_data pm; |
f870fa0b5
|
167 |
}; |
cec37a6e4
|
168 169 |
#define mptcp_for_each_subflow(__msk, __subflow) \ list_for_each_entry(__subflow, &((__msk)->conn_list), node) |
f870fa0b5
|
170 171 172 173 |
static inline struct mptcp_sock *mptcp_sk(const struct sock *sk) { return (struct mptcp_sock *)sk; } |
18b683bff
|
174 175 176 177 178 179 180 181 182 |
static inline struct mptcp_data_frag *mptcp_rtx_tail(const struct sock *sk) { struct mptcp_sock *msk = mptcp_sk(sk); if (list_empty(&msk->rtx_queue)) return NULL; return list_last_entry(&msk->rtx_queue, struct mptcp_data_frag, list); } |
7948f6cc9
|
183 184 185 186 187 188 189 190 191 |
static inline struct mptcp_data_frag *mptcp_rtx_head(const struct sock *sk) { struct mptcp_sock *msk = mptcp_sk(sk); if (list_empty(&msk->rtx_queue)) return NULL; return list_first_entry(&msk->rtx_queue, struct mptcp_data_frag, list); } |
cec37a6e4
|
192 193 |
struct mptcp_subflow_request_sock { struct tcp_request_sock sk; |
d22f4988f
|
194 |
u16 mp_capable : 1, |
cec37a6e4
|
195 |
mp_join : 1, |
d22f4988f
|
196 197 |
backup : 1, remote_key_valid : 1; |
1b1c7a0ef
|
198 |
u8 local_id; |
f296234c9
|
199 |
u8 remote_id; |
cec37a6e4
|
200 201 |
u64 local_key; u64 remote_key; |
79c0949e9
|
202 203 |
u64 idsn; u32 token; |
648ef4b88
|
204 |
u32 ssn_offset; |
f296234c9
|
205 206 207 |
u64 thmac; u32 local_nonce; u32 remote_nonce; |
cec37a6e4
|
208 209 210 211 212 213 214 |
}; static inline struct mptcp_subflow_request_sock * mptcp_subflow_rsk(const struct request_sock *rsk) { return (struct mptcp_subflow_request_sock *)rsk; } |
2303f994b
|
215 216 |
/* MPTCP subflow context */ struct mptcp_subflow_context { |
cec37a6e4
|
217 218 219 |
struct list_head node;/* conn_list of subflows */ u64 local_key; u64 remote_key; |
79c0949e9
|
220 |
u64 idsn; |
648ef4b88
|
221 |
u64 map_seq; |
cc7972ea1
|
222 |
u32 snd_isn; |
79c0949e9
|
223 |
u32 token; |
6d0060f60
|
224 |
u32 rel_write_seq; |
648ef4b88
|
225 226 227 |
u32 map_subflow_seq; u32 ssn_offset; u32 map_data_len; |
cec37a6e4
|
228 |
u32 request_mptcp : 1, /* send MP_CAPABLE */ |
ec3edaa7c
|
229 230 |
request_join : 1, /* send MP_JOIN */ request_bkup : 1, |
cec37a6e4
|
231 |
mp_capable : 1, /* remote is MPTCP capable */ |
f296234c9
|
232 |
mp_join : 1, /* remote is JOINing */ |
0be534f5c
|
233 |
fully_established : 1, /* path validated */ |
f296234c9
|
234 |
pm_notified : 1, /* PM hook called for established status */ |
648ef4b88
|
235 236 |
conn_finished : 1, map_valid : 1, |
d22f4988f
|
237 |
mpc_map : 1, |
f296234c9
|
238 |
backup : 1, |
648ef4b88
|
239 |
data_avail : 1, |
d22f4988f
|
240 |
rx_eof : 1, |
76c42a29c
|
241 |
data_fin_tx_enable : 1, |
d22f4988f
|
242 |
can_ack : 1; /* only after processing the remote a key */ |
76c42a29c
|
243 |
u64 data_fin_tx_seq; |
f296234c9
|
244 245 246 |
u32 remote_nonce; u64 thmac; u32 local_nonce; |
ec3edaa7c
|
247 248 |
u32 remote_token; u8 hmac[MPTCPOPT_HMAC_LEN]; |
f296234c9
|
249 250 |
u8 local_id; u8 remote_id; |
648ef4b88
|
251 |
|
2303f994b
|
252 253 |
struct sock *tcp_sock; /* tcp sk backpointer */ struct sock *conn; /* parent mptcp_sock */ |
cec37a6e4
|
254 |
const struct inet_connection_sock_af_ops *icsk_af_ops; |
648ef4b88
|
255 256 257 |
void (*tcp_data_ready)(struct sock *sk); void (*tcp_state_change)(struct sock *sk); void (*tcp_write_space)(struct sock *sk); |
2303f994b
|
258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 |
struct rcu_head rcu; }; static inline struct mptcp_subflow_context * mptcp_subflow_ctx(const struct sock *sk) { struct inet_connection_sock *icsk = inet_csk(sk); /* Use RCU on icsk_ulp_data only for sock diag code */ return (__force struct mptcp_subflow_context *)icsk->icsk_ulp_data; } static inline struct sock * mptcp_subflow_tcp_sock(const struct mptcp_subflow_context *subflow) { return subflow->tcp_sock; } |
648ef4b88
|
275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 |
static inline u64 mptcp_subflow_get_map_offset(const struct mptcp_subflow_context *subflow) { return tcp_sk(mptcp_subflow_tcp_sock(subflow))->copied_seq - subflow->ssn_offset - subflow->map_subflow_seq; } static inline u64 mptcp_subflow_get_mapped_dsn(const struct mptcp_subflow_context *subflow) { return subflow->map_seq + mptcp_subflow_get_map_offset(subflow); } int mptcp_is_enabled(struct net *net); bool mptcp_subflow_data_available(struct sock *sk); |
2303f994b
|
291 |
void mptcp_subflow_init(void); |
ec3edaa7c
|
292 293 294 295 296 |
/* called with sk socket lock held */ int __mptcp_subflow_connect(struct sock *sk, int ifindex, const struct mptcp_addr_info *loc, const struct mptcp_addr_info *remote); |
2303f994b
|
297 |
int mptcp_subflow_create_socket(struct sock *sk, struct socket **new_sock); |
648ef4b88
|
298 299 300 301 302 303 304 305 306 |
static inline void mptcp_subflow_tcp_fallback(struct sock *sk, struct mptcp_subflow_context *ctx) { sk->sk_data_ready = ctx->tcp_data_ready; sk->sk_state_change = ctx->tcp_state_change; sk->sk_write_space = ctx->tcp_write_space; inet_csk(sk)->icsk_af_ops = ctx->icsk_af_ops; } |
cec37a6e4
|
307 308 309 310 |
extern const struct inet_connection_sock_af_ops ipv4_specific; #if IS_ENABLED(CONFIG_MPTCP_IPV6) extern const struct inet_connection_sock_af_ops ipv6_specific; #endif |
648ef4b88
|
311 |
void mptcp_proto_init(void); |
784325e9f
|
312 313 314 |
#if IS_ENABLED(CONFIG_MPTCP_IPV6) int mptcp_proto_v6_init(void); #endif |
648ef4b88
|
315 |
|
58b099196
|
316 |
struct sock *mptcp_sk_clone(const struct sock *sk, struct request_sock *req); |
cec37a6e4
|
317 318 319 320 |
void mptcp_get_options(const struct sk_buff *skb, struct tcp_options_received *opt_rx); void mptcp_finish_connect(struct sock *sk); |
2e52213c7
|
321 |
void mptcp_data_ready(struct sock *sk, struct sock *ssk); |
f296234c9
|
322 |
bool mptcp_finish_join(struct sock *sk); |
b51f9b80c
|
323 |
void mptcp_data_acked(struct sock *sk); |
59832e246
|
324 |
void mptcp_subflow_eof(struct sock *sk); |
cec37a6e4
|
325 |
|
79c0949e9
|
326 327 328 |
int mptcp_token_new_request(struct request_sock *req); void mptcp_token_destroy_request(u32 token); int mptcp_token_new_connect(struct sock *sk); |
58b099196
|
329 |
int mptcp_token_new_accept(u32 token, struct sock *conn); |
f296234c9
|
330 |
struct mptcp_sock *mptcp_token_get_sock(u32 token); |
79c0949e9
|
331 332 333 334 335 336 337 338 339 340 341 342 343 344 |
void mptcp_token_destroy(u32 token); void mptcp_crypto_key_sha(u64 key, u32 *token, u64 *idsn); static inline void mptcp_crypto_key_gen_sha(u64 *key, u32 *token, u64 *idsn) { /* we might consider a faster version that computes the key as a * hash of some information available in the MPTCP socket. Use * random data at the moment, as it's probably the safest option * in case multiple sockets are opened in different namespaces at * the same time. */ get_random_bytes(key, sizeof(u64)); mptcp_crypto_key_sha(*key, token, idsn); } |
3df523ab5
|
345 |
void mptcp_crypto_hmac_sha(u64 key1, u64 key2, u8 *msg, int len, void *hmac); |
79c0949e9
|
346 |
|
1b1c7a0ef
|
347 348 |
void mptcp_pm_init(void); void mptcp_pm_data_init(struct mptcp_sock *msk); |
926bdeab5
|
349 |
void mptcp_pm_close(struct mptcp_sock *msk); |
1b1c7a0ef
|
350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 |
void mptcp_pm_new_connection(struct mptcp_sock *msk, int server_side); void mptcp_pm_fully_established(struct mptcp_sock *msk); bool mptcp_pm_allow_new_subflow(struct mptcp_sock *msk); void mptcp_pm_connection_closed(struct mptcp_sock *msk); void mptcp_pm_subflow_established(struct mptcp_sock *msk, struct mptcp_subflow_context *subflow); void mptcp_pm_subflow_closed(struct mptcp_sock *msk, u8 id); void mptcp_pm_add_addr_received(struct mptcp_sock *msk, const struct mptcp_addr_info *addr); int mptcp_pm_announce_addr(struct mptcp_sock *msk, const struct mptcp_addr_info *addr); int mptcp_pm_remove_addr(struct mptcp_sock *msk, u8 local_id); int mptcp_pm_remove_subflow(struct mptcp_sock *msk, u8 remote_id); static inline bool mptcp_pm_should_signal(struct mptcp_sock *msk) { return READ_ONCE(msk->pm.addr_signal); } static inline unsigned int mptcp_add_addr_len(int family) { if (family == AF_INET) return TCPOLEN_MPTCP_ADD_ADDR; return TCPOLEN_MPTCP_ADD_ADDR6; } bool mptcp_pm_addr_signal(struct mptcp_sock *msk, unsigned int remaining, struct mptcp_addr_info *saddr); int mptcp_pm_get_local_id(struct mptcp_sock *msk, struct sock_common *skc); |
01cacb00b
|
380 381 382 383 384 385 |
void mptcp_pm_nl_init(void); void mptcp_pm_nl_data_init(struct mptcp_sock *msk); void mptcp_pm_nl_fully_established(struct mptcp_sock *msk); void mptcp_pm_nl_subflow_established(struct mptcp_sock *msk); void mptcp_pm_nl_add_addr_received(struct mptcp_sock *msk); int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk, struct sock_common *skc); |
6d0060f60
|
386 387 388 389 |
static inline struct mptcp_ext *mptcp_get_ext(struct sk_buff *skb) { return (struct mptcp_ext *)skb_ext_find(skb, SKB_EXT_MPTCP); } |
648ef4b88
|
390 391 392 393 394 395 |
static inline bool before64(__u64 seq1, __u64 seq2) { return (__s64)(seq1 - seq2) < 0; } #define after64(seq2, seq1) before64(seq1, seq2) |
5147dfb50
|
396 |
void mptcp_diag_subflow_init(struct tcp_ulp_ops *ops); |
f870fa0b5
|
397 |
#endif /* __MPTCP_PROTOCOL_H */ |