net/ipv4/tcp_timer.c

   1 /*
   2  * INET         An implementation of the TCP/IP protocol suite for the LINUX
   3  *              operating system.  INET is implemented using the  BSD Socket
   4  *              interface as the means of communication with the user level.
   5  *
   6  *              Implementation of the Transmission Control Protocol(TCP).
   7  *
   8  * Authors:     Ross Biro
   9  *              Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
  10  *              Mark Evans, <evansmp@uhura.aston.ac.uk>
  11  *              Corey Minyard <wf-rch!minyard@relay.EU.net>
  12  *              Florian La Roche, <flla@stud.uni-sb.de>
  13  *              Charles Hedrick, <hedrick@klinzhai.rutgers.edu>
  14  *              Linus Torvalds, <torvalds@cs.helsinki.fi>
  15  *              Alan Cox, <gw4pts@gw4pts.ampr.org>
  16  *              Matthew Dillon, <dillon@apollo.west.oic.com>
  17  *              Arnt Gulbrandsen, <agulbra@nvg.unit.no>
  18  *              Jorge Cwik, <jorge@laser.satlink.net>
  19  */
  20
  21 #include <linux/module.h>
  22 #include <linux/gfp.h>
  23 #include <net/tcp.h>
  24
  25 int sysctl_tcp_syn_retries __read_mostly = TCP_SYN_RETRIES;
  26 int sysctl_tcp_synack_retries __read_mostly = TCP_SYNACK_RETRIES;
  27 int sysctl_tcp_keepalive_time __read_mostly = TCP_KEEPALIVE_TIME;
  28 int sysctl_tcp_keepalive_probes __read_mostly = TCP_KEEPALIVE_PROBES;
  29 int sysctl_tcp_keepalive_intvl __read_mostly = TCP_KEEPALIVE_INTVL;
  30 int sysctl_tcp_retries1 __read_mostly = TCP_RETR1;
  31 int sysctl_tcp_retries2 __read_mostly = TCP_RETR2;
  32 int sysctl_tcp_orphan_retries __read_mostly;
  33 int sysctl_tcp_thin_linear_timeouts __read_mostly;
  34
  35 static void tcp_write_err(struct sock *sk)
  36 {
  37         sk->sk_err = sk->sk_err_soft ? : ETIMEDOUT;
  38         sk->sk_error_report(sk);
  39
  40         tcp_done(sk);
  41         NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPABORTONTIMEOUT);
  42 }
  43
  44 /* Do not allow orphaned sockets to eat all our resources.
  45  * This is direct violation of TCP specs, but it is required
  46  * to prevent DoS attacks. It is called when a retransmission timeout
  47  * or zero probe timeout occurs on orphaned socket.
  48  *
  49  * Criteria is still not confirmed experimentally and may change.
  50  * We kill the socket, if:
  51  * 1. If number of orphaned sockets exceeds an administratively configured
  52  *    limit.
  53  * 2. If we have strong memory pressure.
  54  */
  55 static int tcp_out_of_resources(struct sock *sk, int do_reset)
  56 {
  57         struct tcp_sock *tp = tcp_sk(sk);
  58         int shift = 0;
  59
  60         /* If peer does not open window for long time, or did not transmit
  61          * anything for long time, penalize it. */
  62         if ((s32)(tcp_time_stamp - tp->lsndtime) > 2*TCP_RTO_MAX || !do_reset)
  63                 shift++;
  64
  65         /* If some dubious ICMP arrived, penalize even more. */
  66         if (sk->sk_err_soft)
  67                 shift++;
  68
  69         if (tcp_check_oom(sk, shift)) {
  70                 /* Catch exceptional cases, when connection requires reset.
  71                  *      1. Last segment was sent recently. */
  72                 if ((s32)(tcp_time_stamp - tp->lsndtime) <= TCP_TIMEWAIT_LEN ||
  73                     /*  2. Window is closed. */
  74                     (!tp->snd_wnd && !tp->packets_out))
  75                         do_reset = 1;
  76                 if (do_reset)
  77                         tcp_send_active_reset(sk, GFP_ATOMIC);
  78                 tcp_done(sk);
  79                 NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPABORTONMEMORY);
  80                 return 1;
  81         }
  82         return 0;
  83 }
  84
  85 /* Calculate maximal number or retries on an orphaned socket. */
  86 static int tcp_orphan_retries(struct sock *sk, int alive)
  87 {
  88         int retries = sysctl_tcp_orphan_retries; /* May be zero. */
  89
  90         /* We know from an ICMP that something is wrong. */
  91         if (sk->sk_err_soft && !alive)
  92                 retries = 0;
  93
  94         /* However, if socket sent something recently, select some safe
  95          * number of retries. 8 corresponds to >100 seconds with minimal
  96          * RTO of 200msec. */
  97         if (retries == 0 && alive)
  98                 retries = 8;
  99         return retries;
 100 }
 101
 102 static void tcp_mtu_probing(struct inet_connection_sock *icsk, struct sock *sk)
 103 {
 104         /* Black hole detection */
 105         if (sysctl_tcp_mtu_probing) {
 106                 if (!icsk->icsk_mtup.enabled) {
 107                         icsk->icsk_mtup.enabled = 1;
 108                         tcp_sync_mss(sk, icsk->icsk_pmtu_cookie);
 109                 } else {
 110                         struct tcp_sock *tp = tcp_sk(sk);
 111                         int mss;
 112
 113                         mss = tcp_mtu_to_mss(sk, icsk->icsk_mtup.search_low) >> 1;
 114                         mss = min(sysctl_tcp_base_mss, mss);
 115                         mss = max(mss, 68 - tp->tcp_header_len);
 116                         icsk->icsk_mtup.search_low = tcp_mss_to_mtu(sk, mss);
 117                         tcp_sync_mss(sk, icsk->icsk_pmtu_cookie);
 118                 }
 119         }
 120 }
 121
 122 /* This function calculates a "timeout" which is equivalent to the timeout of a
 123  * TCP connection after "boundary" unsuccessful, exponentially backed-off
 124  * retransmissions with an initial RTO of TCP_RTO_MIN or TCP_TIMEOUT_INIT if
 125  * syn_set flag is set.
 126  */
 127 static bool retransmits_timed_out(struct sock *sk,
 128                                   unsigned int boundary,
 129                                   unsigned int timeout,
 130                                   bool syn_set)
 131 {
 132         unsigned int linear_backoff_thresh, start_ts;
 133         unsigned int rto_base = syn_set ? TCP_TIMEOUT_INIT : TCP_RTO_MIN;
 134
 135         if (!inet_csk(sk)->icsk_retransmits)
 136                 return false;
 137
 138         if (unlikely(!tcp_sk(sk)->retrans_stamp))
 139                 start_ts = TCP_SKB_CB(tcp_write_queue_head(sk))->when;
 140         else
 141                 start_ts = tcp_sk(sk)->retrans_stamp;
 142
 143         if (likely(timeout == 0)) {
 144                 linear_backoff_thresh = ilog2(TCP_RTO_MAX/rto_base);
 145
 146                 if (boundary <= linear_backoff_thresh)
 147                         timeout = ((2 << boundary) - 1) * rto_base;
 148                 else
 149                         timeout = ((2 << linear_backoff_thresh) - 1) * rto_base +
 150                                 (boundary - linear_backoff_thresh) * TCP_RTO_MAX;
 151         }
 152         return (tcp_time_stamp - start_ts) >= timeout;
 153 }
 154
 155 /* A write timeout has occurred. Process the after effects. */
 156 static int tcp_write_timeout(struct sock *sk)
 157 {
 158         struct inet_connection_sock *icsk = inet_csk(sk);
 159         struct tcp_sock *tp = tcp_sk(sk);
 160         int retry_until;
 161         bool do_reset, syn_set = false;
 162
 163         if ((1 << sk->sk_state) & (TCPF_SYN_SENT | TCPF_SYN_RECV)) {
 164                 if (icsk->icsk_retransmits) {
 165                         dst_negative_advice(sk);
 166                         if (tp->syn_fastopen || tp->syn_data)
 167                                 tcp_fastopen_cache_set(sk, 0, NULL, true);
 168                         if (tp->syn_data)
 169                                 NET_INC_STATS_BH(sock_net(sk),
 170                                                  LINUX_MIB_TCPFASTOPENACTIVEFAIL);
 171                 }
 172                 retry_until = icsk->icsk_syn_retries ? : sysctl_tcp_syn_retries;
 173                 syn_set = true;
 174         } else {
 175                 if (retransmits_timed_out(sk, sysctl_tcp_retries1, 0, 0)) {
 176                         /* Black hole detection */
 177                         tcp_mtu_probing(icsk, sk);
 178
 179                         dst_negative_advice(sk);
 180                 }
 181
 182                 retry_until = sysctl_tcp_retries2;
 183                 if (sock_flag(sk, SOCK_DEAD)) {
 184                         const int alive = (icsk->icsk_rto < TCP_RTO_MAX);
 185
 186                         retry_until = tcp_orphan_retries(sk, alive);
 187                         do_reset = alive ||
 188                                 !retransmits_timed_out(sk, retry_until, 0, 0);
 189
 190                         if (tcp_out_of_resources(sk, do_reset))
 191                                 return 1;
 192                 }
 193         }
 194
 195         if (retransmits_timed_out(sk, retry_until,
 196                                   syn_set ? 0 : icsk->icsk_user_timeout, syn_set)) {
 197                 /* Has it gone just too far? */
 198                 tcp_write_err(sk);
 199                 return 1;
 200         }
 201         return 0;
 202 }
 203
 204 void tcp_delack_timer_handler(struct sock *sk)
 205 {
 206         struct tcp_sock *tp = tcp_sk(sk);
 207         struct inet_connection_sock *icsk = inet_csk(sk);
 208
 209         sk_mem_reclaim_partial(sk);
 210
 211         if (sk->sk_state == TCP_CLOSE || !(icsk->icsk_ack.pending & ICSK_ACK_TIMER))
 212                 goto out;
 213
 214         if (time_after(icsk->icsk_ack.timeout, jiffies)) {
 215                 sk_reset_timer(sk, &icsk->icsk_delack_timer, icsk->icsk_ack.timeout);
 216                 goto out;
 217         }
 218         icsk->icsk_ack.pending &= ~ICSK_ACK_TIMER;
 219
 220         if (!skb_queue_empty(&tp->ucopy.prequeue)) {
 221                 struct sk_buff *skb;
 222
 223                 NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPSCHEDULERFAILED);
 224
 225                 while ((skb = __skb_dequeue(&tp->ucopy.prequeue)) != NULL)
 226                         sk_backlog_rcv(sk, skb);
 227
 228                 tp->ucopy.memory = 0;
 229         }
 230
 231         if (inet_csk_ack_scheduled(sk)) {
 232                 if (!icsk->icsk_ack.pingpong) {
 233                         /* Delayed ACK missed: inflate ATO. */
 234                         icsk->icsk_ack.ato = min(icsk->icsk_ack.ato << 1, icsk->icsk_rto);
 235                 } else {
 236                         /* Delayed ACK missed: leave pingpong mode and
 237                          * deflate ATO.
 238                          */
 239                         icsk->icsk_ack.pingpong = 0;
 240                         icsk->icsk_ack.ato      = TCP_ATO_MIN;
 241                 }
 242                 tcp_send_ack(sk);
 243                 NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_DELAYEDACKS);
 244         }
 245
 246 out:
 247         if (sk_under_memory_pressure(sk))
 248                 sk_mem_reclaim(sk);
 249 }
 250
 251 static void tcp_delack_timer(unsigned long data)
 252 {
 253         struct sock *sk = (struct sock *)data;
 254
 255         bh_lock_sock(sk);
 256         if (!sock_owned_by_user(sk)) {
 257                 tcp_delack_timer_handler(sk);
 258         } else {
 259                 inet_csk(sk)->icsk_ack.blocked = 1;
 260                 NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_DELAYEDACKLOCKED);
 261                 /* deleguate our work to tcp_release_cb() */
 262                 if (!test_and_set_bit(TCP_DELACK_TIMER_DEFERRED, &tcp_sk(sk)->tsq_flags))
 263                         sock_hold(sk);
 264         }
 265         bh_unlock_sock(sk);
 266         sock_put(sk);
 267 }
 268
 269 static void tcp_probe_timer(struct sock *sk)
 270 {
 271         struct inet_connection_sock *icsk = inet_csk(sk);
 272         struct tcp_sock *tp = tcp_sk(sk);
 273         int max_probes;
 274
 275         if (tp->packets_out || !tcp_send_head(sk)) {
 276                 icsk->icsk_probes_out = 0;
 277                 return;
 278         }
 279
 280         /* *WARNING* RFC 1122 forbids this
 281          *
 282          * It doesn't AFAIK, because we kill the retransmit timer -AK
 283          *
 284          * FIXME: We ought not to do it, Solaris 2.5 actually has fixing
 285          * this behaviour in Solaris down as a bug fix. [AC]
 286          *
 287          * Let me to explain. icsk_probes_out is zeroed by incoming ACKs
 288          * even if they advertise zero window. Hence, connection is killed only
 289          * if we received no ACKs for normal connection timeout. It is not killed
 290          * only because window stays zero for some time, window may be zero
 291          * until armageddon and even later. We are in full accordance
 292          * with RFCs, only probe timer combines both retransmission timeout
 293          * and probe timeout in one bottle.                             --ANK
 294          */
 295         max_probes = sysctl_tcp_retries2;
 296
 297         if (sock_flag(sk, SOCK_DEAD)) {
 298                 const int alive = ((icsk->icsk_rto << icsk->icsk_backoff) < TCP_RTO_MAX);
 299
 300                 max_probes = tcp_orphan_retries(sk, alive);
 301
 302                 if (tcp_out_of_resources(sk, alive || icsk->icsk_probes_out <= max_probes))
 303                         return;
 304         }
 305
 306         if (icsk->icsk_probes_out > max_probes) {
 307                 tcp_write_err(sk);
 308         } else {
 309                 /* Only send another probe if we didn't close things up. */
 310                 tcp_send_probe0(sk);
 311         }
 312 }
 313
 314 /*
 315  *      Timer for Fast Open socket to retransmit SYNACK. Note that the
 316  *      sk here is the child socket, not the parent (listener) socket.
 317  */
 318 static void tcp_fastopen_synack_timer(struct sock *sk)
 319 {
 320         struct inet_connection_sock *icsk = inet_csk(sk);
 321         int max_retries = icsk->icsk_syn_retries ? :
 322             sysctl_tcp_synack_retries + 1; /* add one more retry for fastopen */
 323         struct request_sock *req;
 324
 325         req = tcp_sk(sk)->fastopen_rsk;
 326         req->rsk_ops->syn_ack_timeout(sk, req);
 327
 328         if (req->num_timeout >= max_retries) {
 329                 tcp_write_err(sk);
 330                 return;
 331         }
 332         /* XXX (TFO) - Unlike regular SYN-ACK retransmit, we ignore error
 333          * returned from rtx_syn_ack() to make it more persistent like
 334          * regular retransmit because if the child socket has been accepted
 335          * it's not good to give up too easily.
 336          */
 337         inet_rtx_syn_ack(sk, req);
 338         req->num_timeout++;
 339         inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS,
 340                           TCP_TIMEOUT_INIT << req->num_timeout, TCP_RTO_MAX);
 341 }
 342
 343 /*
 344  *      The TCP retransmit timer.
 345  */
 346
 347 void tcp_retransmit_timer(struct sock *sk)
 348 {
 349         struct tcp_sock *tp = tcp_sk(sk);
 350         struct inet_connection_sock *icsk = inet_csk(sk);
 351
 352         if (tp->fastopen_rsk) {
 353                 WARN_ON_ONCE(sk->sk_state != TCP_SYN_RECV &&
 354                              sk->sk_state != TCP_FIN_WAIT1);
 355                 tcp_fastopen_synack_timer(sk);
 356                 /* Before we receive ACK to our SYN-ACK don't retransmit
 357                  * anything else (e.g., data or FIN segments).
 358                  */
 359                 return;
 360         }
 361         if (!tp->packets_out)
 362                 goto out;
 363
 364         WARN_ON(tcp_write_queue_empty(sk));
 365
 366         tp->tlp_high_seq = 0;
 367
 368         if (!tp->snd_wnd && !sock_flag(sk, SOCK_DEAD) &&
 369             !((1 << sk->sk_state) & (TCPF_SYN_SENT | TCPF_SYN_RECV))) {
 370                 /* Receiver dastardly shrinks window. Our retransmits
 371                  * become zero probes, but we should not timeout this
 372                  * connection. If the socket is an orphan, time it out,
 373                  * we cannot allow such beasts to hang infinitely.
 374                  */
 375                 struct inet_sock *inet = inet_sk(sk);
 376                 if (sk->sk_family == AF_INET) {
 377                         LIMIT_NETDEBUG(KERN_DEBUG pr_fmt("Peer %pI4:%u/%u unexpectedly shrunk window %u:%u (repaired)\n"),
 378                                        &inet->inet_daddr,
 379                                        ntohs(inet->inet_dport), inet->inet_num,
 380                                        tp->snd_una, tp->snd_nxt);
 381                 }
 382 #if IS_ENABLED(CONFIG_IPV6)
 383                 else if (sk->sk_family == AF_INET6) {
 384                         LIMIT_NETDEBUG(KERN_DEBUG pr_fmt("Peer %pI6:%u/%u unexpectedly shrunk window %u:%u (repaired)\n"),
 385                                        &sk->sk_v6_daddr,
 386                                        ntohs(inet->inet_dport), inet->inet_num,
 387                                        tp->snd_una, tp->snd_nxt);
 388                 }
 389 #endif
 390                 if (tcp_time_stamp - tp->rcv_tstamp > TCP_RTO_MAX) {
 391                         tcp_write_err(sk);
 392                         goto out;
 393                 }
 394                 tcp_enter_loss(sk);
 395                 tcp_retransmit_skb(sk, tcp_write_queue_head(sk));
 396                 __sk_dst_reset(sk);
 397                 goto out_reset_timer;
 398         }
 399
 400         if (tcp_write_timeout(sk))
 401                 goto out;
 402
 403         if (icsk->icsk_retransmits == 0) {
 404                 int mib_idx;
 405
 406                 if (icsk->icsk_ca_state == TCP_CA_Recovery) {
 407                         if (tcp_is_sack(tp))
 408                                 mib_idx = LINUX_MIB_TCPSACKRECOVERYFAIL;
 409                         else
 410                                 mib_idx = LINUX_MIB_TCPRENORECOVERYFAIL;
 411                 } else if (icsk->icsk_ca_state == TCP_CA_Loss) {
 412                         mib_idx = LINUX_MIB_TCPLOSSFAILURES;
 413                 } else if ((icsk->icsk_ca_state == TCP_CA_Disorder) ||
 414                            tp->sacked_out) {
 415                         if (tcp_is_sack(tp))
 416                                 mib_idx = LINUX_MIB_TCPSACKFAILURES;
 417                         else
 418                                 mib_idx = LINUX_MIB_TCPRENOFAILURES;
 419                 } else {
 420                         mib_idx = LINUX_MIB_TCPTIMEOUTS;
 421                 }
 422                 NET_INC_STATS_BH(sock_net(sk), mib_idx);
 423         }
 424
 425         tcp_enter_loss(sk);
 426
 427         if (tcp_retransmit_skb(sk, tcp_write_queue_head(sk)) > 0) {
 428                 /* Retransmission failed because of local congestion,
 429                  * do not backoff.
 430                  */
 431                 if (!icsk->icsk_retransmits)
 432                         icsk->icsk_retransmits = 1;
 433                 inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS,
 434                                           min(icsk->icsk_rto, TCP_RESOURCE_PROBE_INTERVAL),
 435                                           TCP_RTO_MAX);
 436                 goto out;
 437         }
 438
 439         /* Increase the timeout each time we retransmit.  Note that
 440          * we do not increase the rtt estimate.  rto is initialized
 441          * from rtt, but increases here.  Jacobson (SIGCOMM 88) suggests
 442          * that doubling rto each time is the least we can get away with.
 443          * In KA9Q, Karn uses this for the first few times, and then
 444          * goes to quadratic.  netBSD doubles, but only goes up to *64,
 445          * and clamps at 1 to 64 sec afterwards.  Note that 120 sec is
 446          * defined in the protocol as the maximum possible RTT.  I guess
 447          * we'll have to use something other than TCP to talk to the
 448          * University of Mars.
 449          *
 450          * PAWS allows us longer timeouts and large windows, so once
 451          * implemented ftp to mars will work nicely. We will have to fix
 452          * the 120 second clamps though!
 453          */
 454         icsk->icsk_backoff++;
 455         icsk->icsk_retransmits++;
 456
 457 out_reset_timer:
 458         /* If stream is thin, use linear timeouts. Since 'icsk_backoff' is
 459          * used to reset timer, set to 0. Recalculate 'icsk_rto' as this
 460          * might be increased if the stream oscillates between thin and thick,
 461          * thus the old value might already be too high compared to the value
 462          * set by 'tcp_set_rto' in tcp_input.c which resets the rto without
 463          * backoff. Limit to TCP_THIN_LINEAR_RETRIES before initiating
 464          * exponential backoff behaviour to avoid continue hammering
 465          * linear-timeout retransmissions into a black hole
 466          */
 467         if (sk->sk_state == TCP_ESTABLISHED &&
 468             (tp->thin_lto || sysctl_tcp_thin_linear_timeouts) &&
 469             tcp_stream_is_thin(tp) &&
 470             icsk->icsk_retransmits <= TCP_THIN_LINEAR_RETRIES) {
 471                 icsk->icsk_backoff = 0;
 472                 icsk->icsk_rto = min(__tcp_set_rto(tp), TCP_RTO_MAX);
 473         } else {
 474                 /* Use normal (exponential) backoff */
 475                 icsk->icsk_rto = min(icsk->icsk_rto << 1, TCP_RTO_MAX);
 476         }
 477         inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS, icsk->icsk_rto, TCP_RTO_MAX);
 478         if (retransmits_timed_out(sk, sysctl_tcp_retries1 + 1, 0, 0))
 479                 __sk_dst_reset(sk);
 480
 481 out:;
 482 }
 483
 484 void tcp_write_timer_handler(struct sock *sk)
 485 {
 486         struct inet_connection_sock *icsk = inet_csk(sk);
 487         int event;
 488
 489         if (sk->sk_state == TCP_CLOSE || !icsk->icsk_pending)
 490                 goto out;
 491
 492         if (time_after(icsk->icsk_timeout, jiffies)) {
 493                 sk_reset_timer(sk, &icsk->icsk_retransmit_timer, icsk->icsk_timeout);
 494                 goto out;
 495         }
 496
 497         event = icsk->icsk_pending;
 498
 499         switch (event) {
 500         case ICSK_TIME_EARLY_RETRANS:
 501                 tcp_resume_early_retransmit(sk);
 502                 break;
 503         case ICSK_TIME_LOSS_PROBE:
 504                 tcp_send_loss_probe(sk);
 505                 break;
 506         case ICSK_TIME_RETRANS:
 507                 icsk->icsk_pending = 0;
 508                 tcp_retransmit_timer(sk);
 509                 break;
 510         case ICSK_TIME_PROBE0:
 511                 icsk->icsk_pending = 0;
 512                 tcp_probe_timer(sk);
 513                 break;
 514         }
 515
 516 out:
 517         sk_mem_reclaim(sk);
 518 }
 519
 520 static void tcp_write_timer(unsigned long data)
 521 {
 522         struct sock *sk = (struct sock *)data;
 523
 524         bh_lock_sock(sk);
 525         if (!sock_owned_by_user(sk)) {
 526                 tcp_write_timer_handler(sk);
 527         } else {
 528                 /* deleguate our work to tcp_release_cb() */
 529                 if (!test_and_set_bit(TCP_WRITE_TIMER_DEFERRED, &tcp_sk(sk)->tsq_flags))
 530                         sock_hold(sk);
 531         }
 532         bh_unlock_sock(sk);
 533         sock_put(sk);
 534 }
 535
 536 /*
 537  *      Timer for listening sockets
 538  */
 539
 540 static void tcp_synack_timer(struct sock *sk)
 541 {
 542         inet_csk_reqsk_queue_prune(sk, TCP_SYNQ_INTERVAL,
 543                                    TCP_TIMEOUT_INIT, TCP_RTO_MAX);
 544 }
 545
 546 void tcp_syn_ack_timeout(struct sock *sk, struct request_sock *req)
 547 {
 548         NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPTIMEOUTS);
 549 }
 550 EXPORT_SYMBOL(tcp_syn_ack_timeout);
 551
 552 void tcp_set_keepalive(struct sock *sk, int val)
 553 {
 554         if ((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN))
 555                 return;
 556
 557         if (val && !sock_flag(sk, SOCK_KEEPOPEN))
 558                 inet_csk_reset_keepalive_timer(sk, keepalive_time_when(tcp_sk(sk)));
 559         else if (!val)
 560                 inet_csk_delete_keepalive_timer(sk);
 561 }
 562
 563
 564 static void tcp_keepalive_timer (unsigned long data)
 565 {
 566         struct sock *sk = (struct sock *) data;
 567         struct inet_connection_sock *icsk = inet_csk(sk);
 568         struct tcp_sock *tp = tcp_sk(sk);
 569         u32 elapsed;
 570
 571         /* Only process if socket is not in use. */
 572         bh_lock_sock(sk);
 573         if (sock_owned_by_user(sk)) {
 574                 /* Try again later. */
 575                 inet_csk_reset_keepalive_timer (sk, HZ/20);
 576                 goto out;
 577         }
 578
 579         if (sk->sk_state == TCP_LISTEN) {
 580                 tcp_synack_timer(sk);
 581                 goto out;
 582         }
 583
 584         if (sk->sk_state == TCP_FIN_WAIT2 && sock_flag(sk, SOCK_DEAD)) {
 585                 if (tp->linger2 >= 0) {
 586                         const int tmo = tcp_fin_time(sk) - TCP_TIMEWAIT_LEN;
 587
 588                         if (tmo > 0) {
 589                                 tcp_time_wait(sk, TCP_FIN_WAIT2, tmo);
 590                                 goto out;
 591                         }
 592                 }
 593                 tcp_send_active_reset(sk, GFP_ATOMIC);
 594                 goto death;
 595         }
 596
 597         if (!sock_flag(sk, SOCK_KEEPOPEN) || sk->sk_state == TCP_CLOSE)
 598                 goto out;
 599
 600         elapsed = keepalive_time_when(tp);
 601
 602         /* It is alive without keepalive 8) */
 603         if (tp->packets_out || tcp_send_head(sk))
 604                 goto resched;
 605
 606         elapsed = keepalive_time_elapsed(tp);
 607
 608         if (elapsed >= keepalive_time_when(tp)) {
 609                 /* If the TCP_USER_TIMEOUT option is enabled, use that
 610                  * to determine when to timeout instead.
 611                  */
 612                 if ((icsk->icsk_user_timeout != 0 &&
 613                     elapsed >= icsk->icsk_user_timeout &&
 614                     icsk->icsk_probes_out > 0) ||
 615                     (icsk->icsk_user_timeout == 0 &&
 616                     icsk->icsk_probes_out >= keepalive_probes(tp))) {
 617                         tcp_send_active_reset(sk, GFP_ATOMIC);
 618                         tcp_write_err(sk);
 619                         goto out;
 620                 }
 621                 if (tcp_write_wakeup(sk) <= 0) {
 622                         icsk->icsk_probes_out++;
 623                         elapsed = keepalive_intvl_when(tp);
 624                 } else {
 625                         /* If keepalive was lost due to local congestion,
 626                          * try harder.
 627                          */
 628                         elapsed = TCP_RESOURCE_PROBE_INTERVAL;
 629                 }
 630         } else {
 631                 /* It is tp->rcv_tstamp + keepalive_time_when(tp) */
 632                 elapsed = keepalive_time_when(tp) - elapsed;
 633         }
 634
 635         sk_mem_reclaim(sk);
 636
 637 resched:
 638         inet_csk_reset_keepalive_timer (sk, elapsed);
 639         goto out;
 640
 641 death:
 642         tcp_done(sk);
 643
 644 out:
 645         bh_unlock_sock(sk);
 646         sock_put(sk);
 647 }
 648
 649 void tcp_init_xmit_timers(struct sock *sk)
 650 {
 651         inet_csk_init_xmit_timers(sk, &tcp_write_timer, &tcp_delack_timer,
 652                                   &tcp_keepalive_timer);
 653 }
 654 EXPORT_SYMBOL(tcp_init_xmit_timers);