1 /* SPDX-License-Identifier: GPL-2.0-or-later */ 2 /* 3 * NET Generic infrastructure for INET connection oriented protocols. 4 * 5 * Definitions for inet_connection_sock 6 * 7 * Authors: Many people, see the TCP sources 8 * 9 * From code originally in TCP 10 */ 11 #ifndef _INET_CONNECTION_SOCK_H 12 #define _INET_CONNECTION_SOCK_H 13 14 #include <linux/compiler.h> 15 #include <linux/string.h> 16 #include <linux/timer.h> 17 #include <linux/poll.h> 18 #include <linux/kernel.h> 19 #include <linux/sockptr.h> 20 21 #include <net/inet_sock.h> 22 #include <net/request_sock.h> 23 24 /* Cancel timers, when they are not required. */ 25 #undef INET_CSK_CLEAR_TIMERS 26 27 struct inet_bind_bucket; 28 struct inet_bind2_bucket; 29 struct tcp_congestion_ops; 30 31 /* 32 * Pointers to address related TCP functions 33 * (i.e. things that depend on the address family) 34 */ 35 struct inet_connection_sock_af_ops { 36 int (*queue_xmit)(struct sock *sk, struct sk_buff *skb, struct flowi *fl); 37 void (*send_check)(struct sock *sk, struct sk_buff *skb); 38 int (*rebuild_header)(struct sock *sk); 39 void (*sk_rx_dst_set)(struct sock *sk, const struct sk_buff *skb); 40 int (*conn_request)(struct sock *sk, struct sk_buff *skb); 41 struct sock *(*syn_recv_sock)(const struct sock *sk, struct sk_buff *skb, 42 struct request_sock *req, 43 struct dst_entry *dst, 44 struct request_sock *req_unhash, 45 bool *own_req); 46 u16 net_header_len; 47 u16 net_frag_header_len; 48 u16 sockaddr_len; 49 int (*setsockopt)(struct sock *sk, int level, int optname, 50 sockptr_t optval, unsigned int optlen); 51 int (*getsockopt)(struct sock *sk, int level, int optname, 52 char __user *optval, int __user *optlen); 53 void (*addr2sockaddr)(struct sock *sk, struct sockaddr *); 54 void (*mtu_reduced)(struct sock *sk); 55 }; 56 57 /** inet_connection_sock - INET connection oriented sock 58 * 59 * @icsk_accept_queue: FIFO of established children 60 * @icsk_bind_hash: Bind node 61 * @icsk_bind2_hash: Bind node in the bhash2 table 62 * @icsk_timeout: Timeout 63 * @icsk_retransmit_timer: Resend (no ack) 64 * @icsk_rto: Retransmit timeout 65 * @icsk_pmtu_cookie Last pmtu seen by socket 66 * @icsk_ca_ops Pluggable congestion control hook 67 * @icsk_af_ops Operations which are AF_INET{4,6} specific 68 * @icsk_ulp_ops Pluggable ULP control hook 69 * @icsk_ulp_data ULP private data 70 * @icsk_clean_acked Clean acked data hook 71 * @icsk_ca_state: Congestion control state 72 * @icsk_retransmits: Number of unrecovered [RTO] timeouts 73 * @icsk_pending: Scheduled timer event 74 * @icsk_backoff: Backoff 75 * @icsk_syn_retries: Number of allowed SYN (or equivalent) retries 76 * @icsk_probes_out: unanswered 0 window probes 77 * @icsk_ext_hdr_len: Network protocol overhead (IP/IPv6 options) 78 * @icsk_ack: Delayed ACK control data 79 * @icsk_mtup; MTU probing control data 80 * @icsk_probes_tstamp: Probe timestamp (cleared by non-zero window ack) 81 * @icsk_user_timeout: TCP_USER_TIMEOUT value 82 */ 83 struct inet_connection_sock { 84 /* inet_sock has to be the first member! */ 85 struct inet_sock icsk_inet; 86 struct request_sock_queue icsk_accept_queue; 87 struct inet_bind_bucket *icsk_bind_hash; 88 struct inet_bind2_bucket *icsk_bind2_hash; 89 unsigned long icsk_timeout; 90 struct timer_list icsk_retransmit_timer; 91 struct timer_list icsk_delack_timer; 92 __u32 icsk_rto; 93 __u32 icsk_rto_min; 94 __u32 icsk_delack_max; 95 __u32 icsk_pmtu_cookie; 96 const struct tcp_congestion_ops *icsk_ca_ops; 97 const struct inet_connection_sock_af_ops *icsk_af_ops; 98 const struct tcp_ulp_ops *icsk_ulp_ops; 99 void __rcu *icsk_ulp_data; 100 void (*icsk_clean_acked)(struct sock *sk, u32 acked_seq); 101 unsigned int (*icsk_sync_mss)(struct sock *sk, u32 pmtu); 102 __u8 icsk_ca_state:5, 103 icsk_ca_initialized:1, 104 icsk_ca_setsockopt:1, 105 icsk_ca_dst_locked:1; 106 __u8 icsk_retransmits; 107 __u8 icsk_pending; 108 __u8 icsk_backoff; 109 __u8 icsk_syn_retries; 110 __u8 icsk_probes_out; 111 __u16 icsk_ext_hdr_len; 112 struct { 113 __u8 pending; /* ACK is pending */ 114 __u8 quick; /* Scheduled number of quick acks */ 115 __u8 pingpong; /* The session is interactive */ 116 __u8 retry; /* Number of attempts */ 117 __u32 ato; /* Predicted tick of soft clock */ 118 unsigned long timeout; /* Currently scheduled timeout */ 119 __u32 lrcvtime; /* timestamp of last received data packet */ 120 __u16 last_seg_size; /* Size of last incoming segment */ 121 __u16 rcv_mss; /* MSS used for delayed ACK decisions */ 122 } icsk_ack; 123 struct { 124 /* Range of MTUs to search */ 125 int search_high; 126 int search_low; 127 128 /* Information on the current probe. */ 129 u32 probe_size:31, 130 /* Is the MTUP feature enabled for this connection? */ 131 enabled:1; 132 133 u32 probe_timestamp; 134 } icsk_mtup; 135 u32 icsk_probes_tstamp; 136 u32 icsk_user_timeout; 137 138 u64 icsk_ca_priv[104 / sizeof(u64)]; 139 #define ICSK_CA_PRIV_SIZE sizeof_field(struct inet_connection_sock, icsk_ca_priv) 140 }; 141 142 #define ICSK_TIME_RETRANS 1 /* Retransmit timer */ 143 #define ICSK_TIME_DACK 2 /* Delayed ack timer */ 144 #define ICSK_TIME_PROBE0 3 /* Zero window probe timer */ 145 #define ICSK_TIME_LOSS_PROBE 5 /* Tail loss probe timer */ 146 #define ICSK_TIME_REO_TIMEOUT 6 /* Reordering timer */ 147 inet_csk(const struct sock * sk)148 static inline struct inet_connection_sock *inet_csk(const struct sock *sk) 149 { 150 return (struct inet_connection_sock *)sk; 151 } 152 inet_csk_ca(const struct sock * sk)153 static inline void *inet_csk_ca(const struct sock *sk) 154 { 155 return (void *)inet_csk(sk)->icsk_ca_priv; 156 } 157 158 struct sock *inet_csk_clone_lock(const struct sock *sk, 159 const struct request_sock *req, 160 const gfp_t priority); 161 162 enum inet_csk_ack_state_t { 163 ICSK_ACK_SCHED = 1, 164 ICSK_ACK_TIMER = 2, 165 ICSK_ACK_PUSHED = 4, 166 ICSK_ACK_PUSHED2 = 8, 167 ICSK_ACK_NOW = 16, /* Send the next ACK immediately (once) */ 168 ICSK_ACK_NOMEM = 32, 169 }; 170 171 void inet_csk_init_xmit_timers(struct sock *sk, 172 void (*retransmit_handler)(struct timer_list *), 173 void (*delack_handler)(struct timer_list *), 174 void (*keepalive_handler)(struct timer_list *)); 175 void inet_csk_clear_xmit_timers(struct sock *sk); 176 void inet_csk_clear_xmit_timers_sync(struct sock *sk); 177 inet_csk_schedule_ack(struct sock * sk)178 static inline void inet_csk_schedule_ack(struct sock *sk) 179 { 180 inet_csk(sk)->icsk_ack.pending |= ICSK_ACK_SCHED; 181 } 182 inet_csk_ack_scheduled(const struct sock * sk)183 static inline int inet_csk_ack_scheduled(const struct sock *sk) 184 { 185 return inet_csk(sk)->icsk_ack.pending & ICSK_ACK_SCHED; 186 } 187 inet_csk_delack_init(struct sock * sk)188 static inline void inet_csk_delack_init(struct sock *sk) 189 { 190 memset(&inet_csk(sk)->icsk_ack, 0, sizeof(inet_csk(sk)->icsk_ack)); 191 } 192 193 void inet_csk_delete_keepalive_timer(struct sock *sk); 194 void inet_csk_reset_keepalive_timer(struct sock *sk, unsigned long timeout); 195 inet_csk_clear_xmit_timer(struct sock * sk,const int what)196 static inline void inet_csk_clear_xmit_timer(struct sock *sk, const int what) 197 { 198 struct inet_connection_sock *icsk = inet_csk(sk); 199 200 if (what == ICSK_TIME_RETRANS || what == ICSK_TIME_PROBE0) { 201 icsk->icsk_pending = 0; 202 #ifdef INET_CSK_CLEAR_TIMERS 203 sk_stop_timer(sk, &icsk->icsk_retransmit_timer); 204 #endif 205 } else if (what == ICSK_TIME_DACK) { 206 icsk->icsk_ack.pending = 0; 207 icsk->icsk_ack.retry = 0; 208 #ifdef INET_CSK_CLEAR_TIMERS 209 sk_stop_timer(sk, &icsk->icsk_delack_timer); 210 #endif 211 } else { 212 pr_debug("inet_csk BUG: unknown timer value\n"); 213 } 214 } 215 216 /* 217 * Reset the retransmission timer 218 */ inet_csk_reset_xmit_timer(struct sock * sk,const int what,unsigned long when,const unsigned long max_when)219 static inline void inet_csk_reset_xmit_timer(struct sock *sk, const int what, 220 unsigned long when, 221 const unsigned long max_when) 222 { 223 struct inet_connection_sock *icsk = inet_csk(sk); 224 225 if (when > max_when) { 226 pr_debug("reset_xmit_timer: sk=%p %d when=0x%lx, caller=%p\n", 227 sk, what, when, (void *)_THIS_IP_); 228 when = max_when; 229 } 230 231 if (what == ICSK_TIME_RETRANS || what == ICSK_TIME_PROBE0 || 232 what == ICSK_TIME_LOSS_PROBE || what == ICSK_TIME_REO_TIMEOUT) { 233 icsk->icsk_pending = what; 234 icsk->icsk_timeout = jiffies + when; 235 sk_reset_timer(sk, &icsk->icsk_retransmit_timer, icsk->icsk_timeout); 236 } else if (what == ICSK_TIME_DACK) { 237 icsk->icsk_ack.pending |= ICSK_ACK_TIMER; 238 icsk->icsk_ack.timeout = jiffies + when; 239 sk_reset_timer(sk, &icsk->icsk_delack_timer, icsk->icsk_ack.timeout); 240 } else { 241 pr_debug("inet_csk BUG: unknown timer value\n"); 242 } 243 } 244 245 static inline unsigned long inet_csk_rto_backoff(const struct inet_connection_sock * icsk,unsigned long max_when)246 inet_csk_rto_backoff(const struct inet_connection_sock *icsk, 247 unsigned long max_when) 248 { 249 u64 when = (u64)icsk->icsk_rto << icsk->icsk_backoff; 250 251 return (unsigned long)min_t(u64, when, max_when); 252 } 253 254 struct sock *inet_csk_accept(struct sock *sk, int flags, int *err, bool kern); 255 256 int inet_csk_get_port(struct sock *sk, unsigned short snum); 257 258 struct dst_entry *inet_csk_route_req(const struct sock *sk, struct flowi4 *fl4, 259 const struct request_sock *req); 260 struct dst_entry *inet_csk_route_child_sock(const struct sock *sk, 261 struct sock *newsk, 262 const struct request_sock *req); 263 264 struct sock *inet_csk_reqsk_queue_add(struct sock *sk, 265 struct request_sock *req, 266 struct sock *child); 267 bool inet_csk_reqsk_queue_hash_add(struct sock *sk, struct request_sock *req, 268 unsigned long timeout); 269 struct sock *inet_csk_complete_hashdance(struct sock *sk, struct sock *child, 270 struct request_sock *req, 271 bool own_req); 272 inet_csk_reqsk_queue_added(struct sock * sk)273 static inline void inet_csk_reqsk_queue_added(struct sock *sk) 274 { 275 reqsk_queue_added(&inet_csk(sk)->icsk_accept_queue); 276 } 277 inet_csk_reqsk_queue_len(const struct sock * sk)278 static inline int inet_csk_reqsk_queue_len(const struct sock *sk) 279 { 280 return reqsk_queue_len(&inet_csk(sk)->icsk_accept_queue); 281 } 282 inet_csk_reqsk_queue_is_full(const struct sock * sk)283 static inline int inet_csk_reqsk_queue_is_full(const struct sock *sk) 284 { 285 return inet_csk_reqsk_queue_len(sk) > READ_ONCE(sk->sk_max_ack_backlog); 286 } 287 288 bool inet_csk_reqsk_queue_drop(struct sock *sk, struct request_sock *req); 289 void inet_csk_reqsk_queue_drop_and_put(struct sock *sk, struct request_sock *req); 290 291 static inline unsigned long reqsk_timeout(struct request_sock * req,unsigned long max_timeout)292 reqsk_timeout(struct request_sock *req, unsigned long max_timeout) 293 { 294 u64 timeout = (u64)req->timeout << req->num_timeout; 295 296 return (unsigned long)min_t(u64, timeout, max_timeout); 297 } 298 inet_csk_prepare_for_destroy_sock(struct sock * sk)299 static inline void inet_csk_prepare_for_destroy_sock(struct sock *sk) 300 { 301 /* The below has to be done to allow calling inet_csk_destroy_sock */ 302 sock_set_flag(sk, SOCK_DEAD); 303 this_cpu_inc(*sk->sk_prot->orphan_count); 304 } 305 306 void inet_csk_destroy_sock(struct sock *sk); 307 void inet_csk_prepare_forced_close(struct sock *sk); 308 309 /* 310 * LISTEN is a special case for poll.. 311 */ inet_csk_listen_poll(const struct sock * sk)312 static inline __poll_t inet_csk_listen_poll(const struct sock *sk) 313 { 314 return !reqsk_queue_empty(&inet_csk(sk)->icsk_accept_queue) ? 315 (EPOLLIN | EPOLLRDNORM) : 0; 316 } 317 318 int inet_csk_listen_start(struct sock *sk); 319 void inet_csk_listen_stop(struct sock *sk); 320 321 void inet_csk_addr2sockaddr(struct sock *sk, struct sockaddr *uaddr); 322 323 /* update the fast reuse flag when adding a socket */ 324 void inet_csk_update_fastreuse(struct inet_bind_bucket *tb, 325 struct sock *sk); 326 327 struct dst_entry *inet_csk_update_pmtu(struct sock *sk, u32 mtu); 328 329 #define TCP_PINGPONG_THRESH 1 330 inet_csk_enter_pingpong_mode(struct sock * sk)331 static inline void inet_csk_enter_pingpong_mode(struct sock *sk) 332 { 333 inet_csk(sk)->icsk_ack.pingpong = TCP_PINGPONG_THRESH; 334 } 335 inet_csk_exit_pingpong_mode(struct sock * sk)336 static inline void inet_csk_exit_pingpong_mode(struct sock *sk) 337 { 338 inet_csk(sk)->icsk_ack.pingpong = 0; 339 } 340 inet_csk_in_pingpong_mode(struct sock * sk)341 static inline bool inet_csk_in_pingpong_mode(struct sock *sk) 342 { 343 return inet_csk(sk)->icsk_ack.pingpong >= TCP_PINGPONG_THRESH; 344 } 345 inet_csk_has_ulp(const struct sock * sk)346 static inline bool inet_csk_has_ulp(const struct sock *sk) 347 { 348 return inet_test_bit(IS_ICSK, sk) && !!inet_csk(sk)->icsk_ulp_ops; 349 } 350 inet_init_csk_locks(struct sock * sk)351 static inline void inet_init_csk_locks(struct sock *sk) 352 { 353 struct inet_connection_sock *icsk = inet_csk(sk); 354 355 spin_lock_init(&icsk->icsk_accept_queue.rskq_lock); 356 spin_lock_init(&icsk->icsk_accept_queue.fastopenq.lock); 357 } 358 359 #endif /* _INET_CONNECTION_SOCK_H */ 360