net/ceph/messenger.c

   1 #include <linux/ceph/ceph_debug.h>
   2
   3 #include <linux/crc32c.h>
   4 #include <linux/ctype.h>
   5 #include <linux/highmem.h>
   6 #include <linux/inet.h>
   7 #include <linux/kthread.h>
   8 #include <linux/net.h>
   9 #include <linux/slab.h>
  10 #include <linux/socket.h>
  11 #include <linux/string.h>
  12 #include <linux/bio.h>
  13 #include <linux/blkdev.h>
  14 #include <linux/dns_resolver.h>
  15 #include <net/tcp.h>
  16
  17 #include <linux/ceph/libceph.h>
  18 #include <linux/ceph/messenger.h>
  19 #include <linux/ceph/decode.h>
  20 #include <linux/ceph/pagelist.h>
  21 #include <linux/export.h>
  22
  23 /*
  24  * Ceph uses the messenger to exchange ceph_msg messages with other
  25  * hosts in the system.  The messenger provides ordered and reliable
  26  * delivery.  We tolerate TCP disconnects by reconnecting (with
  27  * exponential backoff) in the case of a fault (disconnection, bad
  28  * crc, protocol error).  Acks allow sent messages to be discarded by
  29  * the sender.
  30  */
  31
  32 /* static tag bytes (protocol control messages) */
  33 static char tag_msg = CEPH_MSGR_TAG_MSG;
  34 static char tag_ack = CEPH_MSGR_TAG_ACK;
  35 static char tag_keepalive = CEPH_MSGR_TAG_KEEPALIVE;
  36
  37 #ifdef CONFIG_LOCKDEP
  38 static struct lock_class_key socket_class;
  39 #endif
  40
  41
  42 static void queue_con(struct ceph_connection *con);
  43 static void con_work(struct work_struct *);
  44 static void ceph_fault(struct ceph_connection *con);
  45
  46 /*
  47  * nicely render a sockaddr as a string.
  48  */
  49 #define MAX_ADDR_STR 20
  50 #define MAX_ADDR_STR_LEN 60
  51 static char addr_str[MAX_ADDR_STR][MAX_ADDR_STR_LEN];
  52 static DEFINE_SPINLOCK(addr_str_lock);
  53 static int last_addr_str;
  54
  55 const char *ceph_pr_addr(const struct sockaddr_storage *ss)
  56 {
  57         int i;
  58         char *s;
  59         struct sockaddr_in *in4 = (void *)ss;
  60         struct sockaddr_in6 *in6 = (void *)ss;
  61
  62         spin_lock(&addr_str_lock);
  63         i = last_addr_str++;
  64         if (last_addr_str == MAX_ADDR_STR)
  65                 last_addr_str = 0;
  66         spin_unlock(&addr_str_lock);
  67         s = addr_str[i];
  68
  69         switch (ss->ss_family) {
  70         case AF_INET:
  71                 snprintf(s, MAX_ADDR_STR_LEN, "%pI4:%u", &in4->sin_addr,
  72                          (unsigned int)ntohs(in4->sin_port));
  73                 break;
  74
  75         case AF_INET6:
  76                 snprintf(s, MAX_ADDR_STR_LEN, "[%pI6c]:%u", &in6->sin6_addr,
  77                          (unsigned int)ntohs(in6->sin6_port));
  78                 break;
  79
  80         default:
  81                 snprintf(s, MAX_ADDR_STR_LEN, "(unknown sockaddr family %d)",
  82                          (int)ss->ss_family);
  83         }
  84
  85         return s;
  86 }
  87 EXPORT_SYMBOL(ceph_pr_addr);
  88
  89 static void encode_my_addr(struct ceph_messenger *msgr)
  90 {
  91         memcpy(&msgr->my_enc_addr, &msgr->inst.addr, sizeof(msgr->my_enc_addr));
  92         ceph_encode_addr(&msgr->my_enc_addr);
  93 }
  94
  95 /*
  96  * work queue for all reading and writing to/from the socket.
  97  */
  98 struct workqueue_struct *ceph_msgr_wq;
  99
 100 int ceph_msgr_init(void)
 101 {
 102         ceph_msgr_wq = alloc_workqueue("ceph-msgr", WQ_NON_REENTRANT, 0);
 103         if (!ceph_msgr_wq) {
 104                 pr_err("msgr_init failed to create workqueue\n");
 105                 return -ENOMEM;
 106         }
 107         return 0;
 108 }
 109 EXPORT_SYMBOL(ceph_msgr_init);
 110
 111 void ceph_msgr_exit(void)
 112 {
 113         destroy_workqueue(ceph_msgr_wq);
 114 }
 115 EXPORT_SYMBOL(ceph_msgr_exit);
 116
 117 void ceph_msgr_flush(void)
 118 {
 119         flush_workqueue(ceph_msgr_wq);
 120 }
 121 EXPORT_SYMBOL(ceph_msgr_flush);
 122
 123
 124 /*
 125  * socket callback functions
 126  */
 127
 128 /* data available on socket, or listen socket received a connect */
 129 static void ceph_data_ready(struct sock *sk, int count_unused)
 130 {
 131         struct ceph_connection *con =
 132                 (struct ceph_connection *)sk->sk_user_data;
 133         if (sk->sk_state != TCP_CLOSE_WAIT) {
 134                 dout("ceph_data_ready on %p state = %lu, queueing work\n",
 135                      con, con->state);
 136                 queue_con(con);
 137         }
 138 }
 139
 140 /* socket has buffer space for writing */
 141 static void ceph_write_space(struct sock *sk)
 142 {
 143         struct ceph_connection *con =
 144                 (struct ceph_connection *)sk->sk_user_data;
 145
 146         /* only queue to workqueue if there is data we want to write. */
 147         if (test_bit(WRITE_PENDING, &con->state)) {
 148                 dout("ceph_write_space %p queueing write work\n", con);
 149                 queue_con(con);
 150         } else {
 151                 dout("ceph_write_space %p nothing to write\n", con);
 152         }
 153
 154         /* since we have our own write_space, clear the SOCK_NOSPACE flag */
 155         clear_bit(SOCK_NOSPACE, &sk->sk_socket->flags);
 156 }
 157
 158 /* socket's state has changed */
 159 static void ceph_state_change(struct sock *sk)
 160 {
 161         struct ceph_connection *con =
 162                 (struct ceph_connection *)sk->sk_user_data;
 163
 164         dout("ceph_state_change %p state = %lu sk_state = %u\n",
 165              con, con->state, sk->sk_state);
 166
 167         if (test_bit(CLOSED, &con->state))
 168                 return;
 169
 170         switch (sk->sk_state) {
 171         case TCP_CLOSE:
 172                 dout("ceph_state_change TCP_CLOSE\n");
 173         case TCP_CLOSE_WAIT:
 174                 dout("ceph_state_change TCP_CLOSE_WAIT\n");
 175                 if (test_and_set_bit(SOCK_CLOSED, &con->state) == 0) {
 176                         if (test_bit(CONNECTING, &con->state))
 177                                 con->error_msg = "connection failed";
 178                         else
 179                                 con->error_msg = "socket closed";
 180                         queue_con(con);
 181                 }
 182                 break;
 183         case TCP_ESTABLISHED:
 184                 dout("ceph_state_change TCP_ESTABLISHED\n");
 185                 queue_con(con);
 186                 break;
 187         }
 188 }
 189
 190 /*
 191  * set up socket callbacks
 192  */
 193 static void set_sock_callbacks(struct socket *sock,
 194                                struct ceph_connection *con)
 195 {
 196         struct sock *sk = sock->sk;
 197         sk->sk_user_data = (void *)con;
 198         sk->sk_data_ready = ceph_data_ready;
 199         sk->sk_write_space = ceph_write_space;
 200         sk->sk_state_change = ceph_state_change;
 201 }
 202
 203
 204 /*
 205  * socket helpers
 206  */
 207
 208 /*
 209  * initiate connection to a remote socket.
 210  */
 211 static struct socket *ceph_tcp_connect(struct ceph_connection *con)
 212 {
 213         struct sockaddr_storage *paddr = &con->peer_addr.in_addr;
 214         struct socket *sock;
 215         int ret;
 216
 217         BUG_ON(con->sock);
 218         ret = sock_create_kern(con->peer_addr.in_addr.ss_family, SOCK_STREAM,
 219                                IPPROTO_TCP, &sock);
 220         if (ret)
 221                 return ERR_PTR(ret);
 222         con->sock = sock;
 223         sock->sk->sk_allocation = GFP_NOFS;
 224
 225 #ifdef CONFIG_LOCKDEP
 226         lockdep_set_class(&sock->sk->sk_lock, &socket_class);
 227 #endif
 228
 229         set_sock_callbacks(sock, con);
 230
 231         dout("connect %s\n", ceph_pr_addr(&con->peer_addr.in_addr));
 232
 233         ret = sock->ops->connect(sock, (struct sockaddr *)paddr, sizeof(*paddr),
 234                                  O_NONBLOCK);
 235         if (ret == -EINPROGRESS) {
 236                 dout("connect %s EINPROGRESS sk_state = %u\n",
 237                      ceph_pr_addr(&con->peer_addr.in_addr),
 238                      sock->sk->sk_state);
 239                 ret = 0;
 240         }
 241         if (ret < 0) {
 242                 pr_err("connect %s error %d\n",
 243                        ceph_pr_addr(&con->peer_addr.in_addr), ret);
 244                 sock_release(sock);
 245                 con->sock = NULL;
 246                 con->error_msg = "connect error";
 247         }
 248
 249         if (ret < 0)
 250                 return ERR_PTR(ret);
 251         return sock;
 252 }
 253
 254 static int ceph_tcp_recvmsg(struct socket *sock, void *buf, size_t len)
 255 {
 256         struct kvec iov = {buf, len};
 257         struct msghdr msg = { .msg_flags = MSG_DONTWAIT | MSG_NOSIGNAL };
 258         int r;
 259
 260         r = kernel_recvmsg(sock, &msg, &iov, 1, len, msg.msg_flags);
 261         if (r == -EAGAIN)
 262                 r = 0;
 263         return r;
 264 }
 265
 266 /*
 267  * write something.  @more is true if caller will be sending more data
 268  * shortly.
 269  */
 270 static int ceph_tcp_sendmsg(struct socket *sock, struct kvec *iov,
 271                      size_t kvlen, size_t len, int more)
 272 {
 273         struct msghdr msg = { .msg_flags = MSG_DONTWAIT | MSG_NOSIGNAL };
 274         int r;
 275
 276         if (more)
 277                 msg.msg_flags |= MSG_MORE;
 278         else
 279                 msg.msg_flags |= MSG_EOR;  /* superfluous, but what the hell */
 280
 281         r = kernel_sendmsg(sock, &msg, iov, kvlen, len);
 282         if (r == -EAGAIN)
 283                 r = 0;
 284         return r;
 285 }
 286
 287
 288 /*
 289  * Shutdown/close the socket for the given connection.
 290  */
 291 static int con_close_socket(struct ceph_connection *con)
 292 {
 293         int rc;
 294
 295         dout("con_close_socket on %p sock %p\n", con, con->sock);
 296         if (!con->sock)
 297                 return 0;
 298         set_bit(SOCK_CLOSED, &con->state);
 299         rc = con->sock->ops->shutdown(con->sock, SHUT_RDWR);
 300         sock_release(con->sock);
 301         con->sock = NULL;
 302         clear_bit(SOCK_CLOSED, &con->state);
 303         return rc;
 304 }
 305
 306 /*
 307  * Reset a connection.  Discard all incoming and outgoing messages
 308  * and clear *_seq state.
 309  */
 310 static void ceph_msg_remove(struct ceph_msg *msg)
 311 {
 312         list_del_init(&msg->list_head);
 313         ceph_msg_put(msg);
 314 }
 315 static void ceph_msg_remove_list(struct list_head *head)
 316 {
 317         while (!list_empty(head)) {
 318                 struct ceph_msg *msg = list_first_entry(head, struct ceph_msg,
 319                                                         list_head);
 320                 ceph_msg_remove(msg);
 321         }
 322 }
 323
 324 static void reset_connection(struct ceph_connection *con)
 325 {
 326         /* reset connection, out_queue, msg_ and connect_seq */
 327         /* discard existing out_queue and msg_seq */
 328         ceph_msg_remove_list(&con->out_queue);
 329         ceph_msg_remove_list(&con->out_sent);
 330
 331         if (con->in_msg) {
 332                 ceph_msg_put(con->in_msg);
 333                 con->in_msg = NULL;
 334         }
 335
 336         con->connect_seq = 0;
 337         con->out_seq = 0;
 338         if (con->out_msg) {
 339                 ceph_msg_put(con->out_msg);
 340                 con->out_msg = NULL;
 341         }
 342         con->in_seq = 0;
 343         con->in_seq_acked = 0;
 344 }
 345
 346 /*
 347  * mark a peer down.  drop any open connections.
 348  */
 349 void ceph_con_close(struct ceph_connection *con)
 350 {
 351         dout("con_close %p peer %s\n", con,
 352              ceph_pr_addr(&con->peer_addr.in_addr));
 353         set_bit(CLOSED, &con->state);  /* in case there's queued work */
 354         clear_bit(STANDBY, &con->state);  /* avoid connect_seq bump */
 355         clear_bit(LOSSYTX, &con->state);  /* so we retry next connect */
 356         clear_bit(KEEPALIVE_PENDING, &con->state);
 357         clear_bit(WRITE_PENDING, &con->state);
 358         mutex_lock(&con->mutex);
 359         reset_connection(con);
 360         con->peer_global_seq = 0;
 361         cancel_delayed_work(&con->work);
 362         mutex_unlock(&con->mutex);
 363         queue_con(con);
 364 }
 365 EXPORT_SYMBOL(ceph_con_close);
 366
 367 /*
 368  * Reopen a closed connection, with a new peer address.
 369  */
 370 void ceph_con_open(struct ceph_connection *con, struct ceph_entity_addr *addr)
 371 {
 372         dout("con_open %p %s\n", con, ceph_pr_addr(&addr->in_addr));
 373         set_bit(OPENING, &con->state);
 374         clear_bit(CLOSED, &con->state);
 375         memcpy(&con->peer_addr, addr, sizeof(*addr));
 376         con->delay = 0;      /* reset backoff memory */
 377         queue_con(con);
 378 }
 379 EXPORT_SYMBOL(ceph_con_open);
 380
 381 /*
 382  * return true if this connection ever successfully opened
 383  */
 384 bool ceph_con_opened(struct ceph_connection *con)
 385 {
 386         return con->connect_seq > 0;
 387 }
 388
 389 /*
 390  * generic get/put
 391  */
 392 struct ceph_connection *ceph_con_get(struct ceph_connection *con)
 393 {
 394         dout("con_get %p nref = %d -> %d\n", con,
 395              atomic_read(&con->nref), atomic_read(&con->nref) + 1);
 396         if (atomic_inc_not_zero(&con->nref))
 397                 return con;
 398         return NULL;
 399 }
 400
 401 void ceph_con_put(struct ceph_connection *con)
 402 {
 403         dout("con_put %p nref = %d -> %d\n", con,
 404              atomic_read(&con->nref), atomic_read(&con->nref) - 1);
 405         BUG_ON(atomic_read(&con->nref) == 0);
 406         if (atomic_dec_and_test(&con->nref)) {
 407                 BUG_ON(con->sock);
 408                 kfree(con);
 409         }
 410 }
 411
 412 /*
 413  * initialize a new connection.
 414  */
 415 void ceph_con_init(struct ceph_messenger *msgr, struct ceph_connection *con)
 416 {
 417         dout("con_init %p\n", con);
 418         memset(con, 0, sizeof(*con));
 419         atomic_set(&con->nref, 1);
 420         con->msgr = msgr;
 421         mutex_init(&con->mutex);
 422         INIT_LIST_HEAD(&con->out_queue);
 423         INIT_LIST_HEAD(&con->out_sent);
 424         INIT_DELAYED_WORK(&con->work, con_work);
 425 }
 426 EXPORT_SYMBOL(ceph_con_init);
 427
 428
 429 /*
 430  * We maintain a global counter to order connection attempts.  Get
 431  * a unique seq greater than @gt.
 432  */
 433 static u32 get_global_seq(struct ceph_messenger *msgr, u32 gt)
 434 {
 435         u32 ret;
 436
 437         spin_lock(&msgr->global_seq_lock);
 438         if (msgr->global_seq < gt)
 439                 msgr->global_seq = gt;
 440         ret = ++msgr->global_seq;
 441         spin_unlock(&msgr->global_seq_lock);
 442         return ret;
 443 }
 444
 445
 446 /*
 447  * Prepare footer for currently outgoing message, and finish things
 448  * off.  Assumes out_kvec* are already valid.. we just add on to the end.
 449  */
 450 static void prepare_write_message_footer(struct ceph_connection *con, int v)
 451 {
 452         struct ceph_msg *m = con->out_msg;
 453
 454         dout("prepare_write_message_footer %p\n", con);
 455         con->out_kvec_is_msg = true;
 456         con->out_kvec[v].iov_base = &m->footer;
 457         con->out_kvec[v].iov_len = sizeof(m->footer);
 458         con->out_kvec_bytes += sizeof(m->footer);
 459         con->out_kvec_left++;
 460         con->out_more = m->more_to_follow;
 461         con->out_msg_done = true;
 462 }
 463
 464 /*
 465  * Prepare headers for the next outgoing message.
 466  */
 467 static void prepare_write_message(struct ceph_connection *con)
 468 {
 469         struct ceph_msg *m;
 470         int v = 0;
 471
 472         con->out_kvec_bytes = 0;
 473         con->out_kvec_is_msg = true;
 474         con->out_msg_done = false;
 475
 476         /* Sneak an ack in there first?  If we can get it into the same
 477          * TCP packet that's a good thing. */
 478         if (con->in_seq > con->in_seq_acked) {
 479                 con->in_seq_acked = con->in_seq;
 480                 con->out_kvec[v].iov_base = &tag_ack;
 481                 con->out_kvec[v++].iov_len = 1;
 482                 con->out_temp_ack = cpu_to_le64(con->in_seq_acked);
 483                 con->out_kvec[v].iov_base = &con->out_temp_ack;
 484                 con->out_kvec[v++].iov_len = sizeof(con->out_temp_ack);
 485                 con->out_kvec_bytes = 1 + sizeof(con->out_temp_ack);
 486         }
 487
 488         m = list_first_entry(&con->out_queue,
 489                        struct ceph_msg, list_head);
 490         con->out_msg = m;
 491
 492         /* put message on sent list */
 493         ceph_msg_get(m);
 494         list_move_tail(&m->list_head, &con->out_sent);
 495
 496         /*
 497          * only assign outgoing seq # if we haven't sent this message
 498          * yet.  if it is requeued, resend with it's original seq.
 499          */
 500         if (m->needs_out_seq) {
 501                 m->hdr.seq = cpu_to_le64(++con->out_seq);
 502                 m->needs_out_seq = false;
 503         }
 504
 505         dout("prepare_write_message %p seq %lld type %d len %d+%d+%d %d pgs\n",
 506              m, con->out_seq, le16_to_cpu(m->hdr.type),
 507              le32_to_cpu(m->hdr.front_len), le32_to_cpu(m->hdr.middle_len),
 508              le32_to_cpu(m->hdr.data_len),
 509              m->nr_pages);
 510         BUG_ON(le32_to_cpu(m->hdr.front_len) != m->front.iov_len);
 511
 512         /* tag + hdr + front + middle */
 513         con->out_kvec[v].iov_base = &tag_msg;
 514         con->out_kvec[v++].iov_len = 1;
 515         con->out_kvec[v].iov_base = &m->hdr;
 516         con->out_kvec[v++].iov_len = sizeof(m->hdr);
 517         con->out_kvec[v++] = m->front;
 518         if (m->middle)
 519                 con->out_kvec[v++] = m->middle->vec;
 520         con->out_kvec_left = v;
 521         con->out_kvec_bytes += 1 + sizeof(m->hdr) + m->front.iov_len +
 522                 (m->middle ? m->middle->vec.iov_len : 0);
 523         con->out_kvec_cur = con->out_kvec;
 524
 525         /* fill in crc (except data pages), footer */
 526         con->out_msg->hdr.crc =
 527                 cpu_to_le32(crc32c(0, (void *)&m->hdr,
 528                                       sizeof(m->hdr) - sizeof(m->hdr.crc)));
 529         con->out_msg->footer.flags = CEPH_MSG_FOOTER_COMPLETE;
 530         con->out_msg->footer.front_crc =
 531                 cpu_to_le32(crc32c(0, m->front.iov_base, m->front.iov_len));
 532         if (m->middle)
 533                 con->out_msg->footer.middle_crc =
 534                         cpu_to_le32(crc32c(0, m->middle->vec.iov_base,
 535                                            m->middle->vec.iov_len));
 536         else
 537                 con->out_msg->footer.middle_crc = 0;
 538         con->out_msg->footer.data_crc = 0;
 539         dout("prepare_write_message front_crc %u data_crc %u\n",
 540              le32_to_cpu(con->out_msg->footer.front_crc),
 541              le32_to_cpu(con->out_msg->footer.middle_crc));
 542
 543         /* is there a data payload? */
 544         if (le32_to_cpu(m->hdr.data_len) > 0) {
 545                 /* initialize page iterator */
 546                 con->out_msg_pos.page = 0;
 547                 if (m->pages)
 548                         con->out_msg_pos.page_pos = m->page_alignment;
 549                 else
 550                         con->out_msg_pos.page_pos = 0;
 551                 con->out_msg_pos.data_pos = 0;
 552                 con->out_msg_pos.did_page_crc = 0;
 553                 con->out_more = 1;  /* data + footer will follow */
 554         } else {
 555                 /* no, queue up footer too and be done */
 556                 prepare_write_message_footer(con, v);
 557         }
 558
 559         set_bit(WRITE_PENDING, &con->state);
 560 }
 561
 562 /*
 563  * Prepare an ack.
 564  */
 565 static void prepare_write_ack(struct ceph_connection *con)
 566 {
 567         dout("prepare_write_ack %p %llu -> %llu\n", con,
 568              con->in_seq_acked, con->in_seq);
 569         con->in_seq_acked = con->in_seq;
 570
 571         con->out_kvec[0].iov_base = &tag_ack;
 572         con->out_kvec[0].iov_len = 1;
 573         con->out_temp_ack = cpu_to_le64(con->in_seq_acked);
 574         con->out_kvec[1].iov_base = &con->out_temp_ack;
 575         con->out_kvec[1].iov_len = sizeof(con->out_temp_ack);
 576         con->out_kvec_left = 2;
 577         con->out_kvec_bytes = 1 + sizeof(con->out_temp_ack);
 578         con->out_kvec_cur = con->out_kvec;
 579         con->out_more = 1;  /* more will follow.. eventually.. */
 580         set_bit(WRITE_PENDING, &con->state);
 581 }
 582
 583 /*
 584  * Prepare to write keepalive byte.
 585  */
 586 static void prepare_write_keepalive(struct ceph_connection *con)
 587 {
 588         dout("prepare_write_keepalive %p\n", con);
 589         con->out_kvec[0].iov_base = &tag_keepalive;
 590         con->out_kvec[0].iov_len = 1;
 591         con->out_kvec_left = 1;
 592         con->out_kvec_bytes = 1;
 593         con->out_kvec_cur = con->out_kvec;
 594         set_bit(WRITE_PENDING, &con->state);
 595 }
 596
 597 /*
 598  * Connection negotiation.
 599  */
 600
 601 static int prepare_connect_authorizer(struct ceph_connection *con)
 602 {
 603         void *auth_buf;
 604         int auth_len = 0;
 605         int auth_protocol = 0;
 606
 607         mutex_unlock(&con->mutex);
 608         if (con->ops->get_authorizer)
 609                 con->ops->get_authorizer(con, &auth_buf, &auth_len,
 610                                          &auth_protocol, &con->auth_reply_buf,
 611                                          &con->auth_reply_buf_len,
 612                                          con->auth_retry);
 613         mutex_lock(&con->mutex);
 614
 615         if (test_bit(CLOSED, &con->state) ||
 616             test_bit(OPENING, &con->state))
 617                 return -EAGAIN;
 618
 619         con->out_connect.authorizer_protocol = cpu_to_le32(auth_protocol);
 620         con->out_connect.authorizer_len = cpu_to_le32(auth_len);
 621
 622         if (auth_len) {
 623                 con->out_kvec[con->out_kvec_left].iov_base = auth_buf;
 624                 con->out_kvec[con->out_kvec_left].iov_len = auth_len;
 625                 con->out_kvec_left++;
 626                 con->out_kvec_bytes += auth_len;
 627         }
 628         return 0;
 629 }
 630
 631 /*
 632  * We connected to a peer and are saying hello.
 633  */
 634 static void prepare_write_banner(struct ceph_messenger *msgr,
 635                                  struct ceph_connection *con)
 636 {
 637         int len = strlen(CEPH_BANNER);
 638
 639         con->out_kvec[0].iov_base = CEPH_BANNER;
 640         con->out_kvec[0].iov_len = len;
 641         con->out_kvec[1].iov_base = &msgr->my_enc_addr;
 642         con->out_kvec[1].iov_len = sizeof(msgr->my_enc_addr);
 643         con->out_kvec_left = 2;
 644         con->out_kvec_bytes = len + sizeof(msgr->my_enc_addr);
 645         con->out_kvec_cur = con->out_kvec;
 646         con->out_more = 0;
 647         set_bit(WRITE_PENDING, &con->state);
 648 }
 649
 650 static int prepare_write_connect(struct ceph_messenger *msgr,
 651                                  struct ceph_connection *con,
 652                                  int after_banner)
 653 {
 654         unsigned global_seq = get_global_seq(con->msgr, 0);
 655         int proto;
 656
 657         switch (con->peer_name.type) {
 658         case CEPH_ENTITY_TYPE_MON:
 659                 proto = CEPH_MONC_PROTOCOL;
 660                 break;
 661         case CEPH_ENTITY_TYPE_OSD:
 662                 proto = CEPH_OSDC_PROTOCOL;
 663                 break;
 664         case CEPH_ENTITY_TYPE_MDS:
 665                 proto = CEPH_MDSC_PROTOCOL;
 666                 break;
 667         default:
 668                 BUG();
 669         }
 670
 671         dout("prepare_write_connect %p cseq=%d gseq=%d proto=%d\n", con,
 672              con->connect_seq, global_seq, proto);
 673
 674         con->out_connect.features = cpu_to_le64(msgr->supported_features);
 675         con->out_connect.host_type = cpu_to_le32(CEPH_ENTITY_TYPE_CLIENT);
 676         con->out_connect.connect_seq = cpu_to_le32(con->connect_seq);
 677         con->out_connect.global_seq = cpu_to_le32(global_seq);
 678         con->out_connect.protocol_version = cpu_to_le32(proto);
 679         con->out_connect.flags = 0;
 680
 681         if (!after_banner) {
 682                 con->out_kvec_left = 0;
 683                 con->out_kvec_bytes = 0;
 684         }
 685         con->out_kvec[con->out_kvec_left].iov_base = &con->out_connect;
 686         con->out_kvec[con->out_kvec_left].iov_len = sizeof(con->out_connect);
 687         con->out_kvec_left++;
 688         con->out_kvec_bytes += sizeof(con->out_connect);
 689         con->out_kvec_cur = con->out_kvec;
 690         con->out_more = 0;
 691         set_bit(WRITE_PENDING, &con->state);
 692
 693         return prepare_connect_authorizer(con);
 694 }
 695
 696
 697 /*
 698  * write as much of pending kvecs to the socket as we can.
 699  *  1 -> done
 700  *  0 -> socket full, but more to do
 701  * <0 -> error
 702  */
 703 static int write_partial_kvec(struct ceph_connection *con)
 704 {
 705         int ret;
 706
 707         dout("write_partial_kvec %p %d left\n", con, con->out_kvec_bytes);
 708         while (con->out_kvec_bytes > 0) {
 709                 ret = ceph_tcp_sendmsg(con->sock, con->out_kvec_cur,
 710                                        con->out_kvec_left, con->out_kvec_bytes,
 711                                        con->out_more);
 712                 if (ret <= 0)
 713                         goto out;
 714                 con->out_kvec_bytes -= ret;
 715                 if (con->out_kvec_bytes == 0)
 716                         break;            /* done */
 717                 while (ret > 0) {
 718                         if (ret >= con->out_kvec_cur->iov_len) {
 719                                 ret -= con->out_kvec_cur->iov_len;
 720                                 con->out_kvec_cur++;
 721                                 con->out_kvec_left--;
 722                         } else {
 723                                 con->out_kvec_cur->iov_len -= ret;
 724                                 con->out_kvec_cur->iov_base += ret;
 725                                 ret = 0;
 726                                 break;
 727                         }
 728                 }
 729         }
 730         con->out_kvec_left = 0;
 731         con->out_kvec_is_msg = false;
 732         ret = 1;
 733 out:
 734         dout("write_partial_kvec %p %d left in %d kvecs ret = %d\n", con,
 735              con->out_kvec_bytes, con->out_kvec_left, ret);
 736         return ret;  /* done! */
 737 }
 738
 739 #ifdef CONFIG_BLOCK
 740 static void init_bio_iter(struct bio *bio, struct bio **iter, int *seg)
 741 {
 742         if (!bio) {
 743                 *iter = NULL;
 744                 *seg = 0;
 745                 return;
 746         }
 747         *iter = bio;
 748         *seg = bio->bi_idx;
 749 }
 750
 751 static void iter_bio_next(struct bio **bio_iter, int *seg)
 752 {
 753         if (*bio_iter == NULL)
 754                 return;
 755
 756         BUG_ON(*seg >= (*bio_iter)->bi_vcnt);
 757
 758         (*seg)++;
 759         if (*seg == (*bio_iter)->bi_vcnt)
 760                 init_bio_iter((*bio_iter)->bi_next, bio_iter, seg);
 761 }
 762 #endif
 763
 764 /*
 765  * Write as much message data payload as we can.  If we finish, queue
 766  * up the footer.
 767  *  1 -> done, footer is now queued in out_kvec[].
 768  *  0 -> socket full, but more to do
 769  * <0 -> error
 770  */
 771 static int write_partial_msg_pages(struct ceph_connection *con)
 772 {
 773         struct ceph_msg *msg = con->out_msg;
 774         unsigned data_len = le32_to_cpu(msg->hdr.data_len);
 775         size_t len;
 776         int crc = con->msgr->nocrc;
 777         int ret;
 778         int total_max_write;
 779         int in_trail = 0;
 780         size_t trail_len = (msg->trail ? msg->trail->length : 0);
 781
 782         dout("write_partial_msg_pages %p msg %p page %d/%d offset %d\n",
 783              con, con->out_msg, con->out_msg_pos.page, con->out_msg->nr_pages,
 784              con->out_msg_pos.page_pos);
 785
 786 #ifdef CONFIG_BLOCK
 787         if (msg->bio && !msg->bio_iter)
 788                 init_bio_iter(msg->bio, &msg->bio_iter, &msg->bio_seg);
 789 #endif
 790
 791         while (data_len > con->out_msg_pos.data_pos) {
 792                 struct page *page = NULL;
 793                 void *kaddr = NULL;
 794                 int max_write = PAGE_SIZE;
 795                 int page_shift = 0;
 796
 797                 total_max_write = data_len - trail_len -
 798                         con->out_msg_pos.data_pos;
 799
 800                 /*
 801                  * if we are calculating the data crc (the default), we need
 802                  * to map the page.  if our pages[] has been revoked, use the
 803                  * zero page.
 804                  */
 805
 806                 /* have we reached the trail part of the data? */
 807                 if (con->out_msg_pos.data_pos >= data_len - trail_len) {
 808                         in_trail = 1;
 809
 810                         total_max_write = data_len - con->out_msg_pos.data_pos;
 811
 812                         page = list_first_entry(&msg->trail->head,
 813                                                 struct page, lru);
 814                         if (crc)
 815                                 kaddr = kmap(page);
 816                         max_write = PAGE_SIZE;
 817                 } else if (msg->pages) {
 818                         page = msg->pages[con->out_msg_pos.page];
 819                         if (crc)
 820                                 kaddr = kmap(page);
 821                 } else if (msg->pagelist) {
 822                         page = list_first_entry(&msg->pagelist->head,
 823                                                 struct page, lru);
 824                         if (crc)
 825                                 kaddr = kmap(page);
 826 #ifdef CONFIG_BLOCK
 827                 } else if (msg->bio) {
 828                         struct bio_vec *bv;
 829
 830                         bv = bio_iovec_idx(msg->bio_iter, msg->bio_seg);
 831                         page = bv->bv_page;
 832                         page_shift = bv->bv_offset;
 833                         if (crc)
 834                                 kaddr = kmap(page) + page_shift;
 835                         max_write = bv->bv_len;
 836 #endif
 837                 } else {
 838                         page = con->msgr->zero_page;
 839                         if (crc)
 840                                 kaddr = page_address(con->msgr->zero_page);
 841                 }
 842                 len = min_t(int, max_write - con->out_msg_pos.page_pos,
 843                             total_max_write);
 844
 845                 if (crc && !con->out_msg_pos.did_page_crc) {
 846                         void *base = kaddr + con->out_msg_pos.page_pos;
 847                         u32 tmpcrc = le32_to_cpu(con->out_msg->footer.data_crc);
 848
 849                         BUG_ON(kaddr == NULL);
 850                         con->out_msg->footer.data_crc =
 851                                 cpu_to_le32(crc32c(tmpcrc, base, len));
 852                         con->out_msg_pos.did_page_crc = 1;
 853                 }
 854                 ret = kernel_sendpage(con->sock, page,
 855                                       con->out_msg_pos.page_pos + page_shift,
 856                                       len,
 857                                       MSG_DONTWAIT | MSG_NOSIGNAL |
 858                                       MSG_MORE);
 859
 860                 if (crc &&
 861                     (msg->pages || msg->pagelist || msg->bio || in_trail))
 862                         kunmap(page);
 863
 864                 if (ret == -EAGAIN)
 865                         ret = 0;
 866                 if (ret <= 0)
 867                         goto out;
 868
 869                 con->out_msg_pos.data_pos += ret;
 870                 con->out_msg_pos.page_pos += ret;
 871                 if (ret == len) {
 872                         con->out_msg_pos.page_pos = 0;
 873                         con->out_msg_pos.page++;
 874                         con->out_msg_pos.did_page_crc = 0;
 875                         if (in_trail)
 876                                 list_move_tail(&page->lru,
 877                                                &msg->trail->head);
 878                         else if (msg->pagelist)
 879                                 list_move_tail(&page->lru,
 880                                                &msg->pagelist->head);
 881 #ifdef CONFIG_BLOCK
 882                         else if (msg->bio)
 883                                 iter_bio_next(&msg->bio_iter, &msg->bio_seg);
 884 #endif
 885                 }
 886         }
 887
 888         dout("write_partial_msg_pages %p msg %p done\n", con, msg);
 889
 890         /* prepare and queue up footer, too */
 891         if (!crc)
 892                 con->out_msg->footer.flags |= CEPH_MSG_FOOTER_NOCRC;
 893         con->out_kvec_bytes = 0;
 894         con->out_kvec_left = 0;
 895         con->out_kvec_cur = con->out_kvec;
 896         prepare_write_message_footer(con, 0);
 897         ret = 1;
 898 out:
 899         return ret;
 900 }
 901
 902 /*
 903  * write some zeros
 904  */
 905 static int write_partial_skip(struct ceph_connection *con)
 906 {
 907         int ret;
 908
 909         while (con->out_skip > 0) {
 910                 struct kvec iov = {
 911                         .iov_base = page_address(con->msgr->zero_page),
 912                         .iov_len = min(con->out_skip, (int)PAGE_CACHE_SIZE)
 913                 };
 914
 915                 ret = ceph_tcp_sendmsg(con->sock, &iov, 1, iov.iov_len, 1);
 916                 if (ret <= 0)
 917                         goto out;
 918                 con->out_skip -= ret;
 919         }
 920         ret = 1;
 921 out:
 922         return ret;
 923 }
 924
 925 /*
 926  * Prepare to read connection handshake, or an ack.
 927  */
 928 static void prepare_read_banner(struct ceph_connection *con)
 929 {
 930         dout("prepare_read_banner %p\n", con);
 931         con->in_base_pos = 0;
 932 }
 933
 934 static void prepare_read_connect(struct ceph_connection *con)
 935 {
 936         dout("prepare_read_connect %p\n", con);
 937         con->in_base_pos = 0;
 938 }
 939
 940 static void prepare_read_ack(struct ceph_connection *con)
 941 {
 942         dout("prepare_read_ack %p\n", con);
 943         con->in_base_pos = 0;
 944 }
 945
 946 static void prepare_read_tag(struct ceph_connection *con)
 947 {
 948         dout("prepare_read_tag %p\n", con);
 949         con->in_base_pos = 0;
 950         con->in_tag = CEPH_MSGR_TAG_READY;
 951 }
 952
 953 /*
 954  * Prepare to read a message.
 955  */
 956 static int prepare_read_message(struct ceph_connection *con)
 957 {
 958         dout("prepare_read_message %p\n", con);
 959         BUG_ON(con->in_msg != NULL);
 960         con->in_base_pos = 0;
 961         con->in_front_crc = con->in_middle_crc = con->in_data_crc = 0;
 962         return 0;
 963 }
 964
 965
 966 static int read_partial(struct ceph_connection *con,
 967                         int *to, int size, void *object)
 968 {
 969         *to += size;
 970         while (con->in_base_pos < *to) {
 971                 int left = *to - con->in_base_pos;
 972                 int have = size - left;
 973                 int ret = ceph_tcp_recvmsg(con->sock, object + have, left);
 974                 if (ret <= 0)
 975                         return ret;
 976                 con->in_base_pos += ret;
 977         }
 978         return 1;
 979 }
 980
 981
 982 /*
 983  * Read all or part of the connect-side handshake on a new connection
 984  */
 985 static int read_partial_banner(struct ceph_connection *con)
 986 {
 987         int ret, to = 0;
 988
 989         dout("read_partial_banner %p at %d\n", con, con->in_base_pos);
 990
 991         /* peer's banner */
 992         ret = read_partial(con, &to, strlen(CEPH_BANNER), con->in_banner);
 993         if (ret <= 0)
 994                 goto out;
 995         ret = read_partial(con, &to, sizeof(con->actual_peer_addr),
 996                            &con->actual_peer_addr);
 997         if (ret <= 0)
 998                 goto out;
 999         ret = read_partial(con, &to, sizeof(con->peer_addr_for_me),
1000                            &con->peer_addr_for_me);
1001         if (ret <= 0)
1002                 goto out;
1003 out:
1004         return ret;
1005 }
1006
1007 static int read_partial_connect(struct ceph_connection *con)
1008 {
1009         int ret, to = 0;
1010
1011         dout("read_partial_connect %p at %d\n", con, con->in_base_pos);
1012
1013         ret = read_partial(con, &to, sizeof(con->in_reply), &con->in_reply);
1014         if (ret <= 0)
1015                 goto out;
1016         ret = read_partial(con, &to, le32_to_cpu(con->in_reply.authorizer_len),
1017                            con->auth_reply_buf);
1018         if (ret <= 0)
1019                 goto out;
1020
1021         dout("read_partial_connect %p tag %d, con_seq = %u, g_seq = %u\n",
1022              con, (int)con->in_reply.tag,
1023              le32_to_cpu(con->in_reply.connect_seq),
1024              le32_to_cpu(con->in_reply.global_seq));
1025 out:
1026         return ret;
1027
1028 }
1029
1030 /*
1031  * Verify the hello banner looks okay.
1032  */
1033 static int verify_hello(struct ceph_connection *con)
1034 {
1035         if (memcmp(con->in_banner, CEPH_BANNER, strlen(CEPH_BANNER))) {
1036                 pr_err("connect to %s got bad banner\n",
1037                        ceph_pr_addr(&con->peer_addr.in_addr));
1038                 con->error_msg = "protocol error, bad banner";
1039                 return -1;
1040         }
1041         return 0;
1042 }
1043
1044 static bool addr_is_blank(struct sockaddr_storage *ss)
1045 {
1046         switch (ss->ss_family) {
1047         case AF_INET:
1048                 return ((struct sockaddr_in *)ss)->sin_addr.s_addr == 0;
1049         case AF_INET6:
1050                 return
1051                      ((struct sockaddr_in6 *)ss)->sin6_addr.s6_addr32[0] == 0 &&
1052                      ((struct sockaddr_in6 *)ss)->sin6_addr.s6_addr32[1] == 0 &&
1053                      ((struct sockaddr_in6 *)ss)->sin6_addr.s6_addr32[2] == 0 &&
1054                      ((struct sockaddr_in6 *)ss)->sin6_addr.s6_addr32[3] == 0;
1055         }
1056         return false;
1057 }
1058
1059 static int addr_port(struct sockaddr_storage *ss)
1060 {
1061         switch (ss->ss_family) {
1062         case AF_INET:
1063                 return ntohs(((struct sockaddr_in *)ss)->sin_port);
1064         case AF_INET6:
1065                 return ntohs(((struct sockaddr_in6 *)ss)->sin6_port);
1066         }
1067         return 0;
1068 }
1069
1070 static void addr_set_port(struct sockaddr_storage *ss, int p)
1071 {
1072         switch (ss->ss_family) {
1073         case AF_INET:
1074                 ((struct sockaddr_in *)ss)->sin_port = htons(p);
1075                 break;
1076         case AF_INET6:
1077                 ((struct sockaddr_in6 *)ss)->sin6_port = htons(p);
1078                 break;
1079         }
1080 }
1081
1082 /*
1083  * Unlike other *_pton function semantics, zero indicates success.
1084  */
1085 static int ceph_pton(const char *str, size_t len, struct sockaddr_storage *ss,
1086                 char delim, const char **ipend)
1087 {
1088         struct sockaddr_in *in4 = (void *)ss;
1089         struct sockaddr_in6 *in6 = (void *)ss;
1090
1091         memset(ss, 0, sizeof(*ss));
1092
1093         if (in4_pton(str, len, (u8 *)&in4->sin_addr.s_addr, delim, ipend)) {
1094                 ss->ss_family = AF_INET;
1095                 return 0;
1096         }
1097
1098         if (in6_pton(str, len, (u8 *)&in6->sin6_addr.s6_addr, delim, ipend)) {
1099                 ss->ss_family = AF_INET6;
1100                 return 0;
1101         }
1102
1103         return -EINVAL;
1104 }
1105
1106 /*
1107  * Extract hostname string and resolve using kernel DNS facility.
1108  */
1109 #ifdef CONFIG_CEPH_LIB_USE_DNS_RESOLVER
1110 static int ceph_dns_resolve_name(const char *name, size_t namelen,
1111                 struct sockaddr_storage *ss, char delim, const char **ipend)
1112 {
1113         const char *end, *delim_p;
1114         char *colon_p, *ip_addr = NULL;
1115         int ip_len, ret;
1116
1117         /*
1118          * The end of the hostname occurs immediately preceding the delimiter or
1119          * the port marker (':') where the delimiter takes precedence.
1120          */
1121         delim_p = memchr(name, delim, namelen);
1122         colon_p = memchr(name, ':', namelen);
1123
1124         if (delim_p && colon_p)
1125                 end = delim_p < colon_p ? delim_p : colon_p;
1126         else if (!delim_p && colon_p)
1127                 end = colon_p;
1128         else {
1129                 end = delim_p;
1130                 if (!end) /* case: hostname:/ */
1131                         end = name + namelen;
1132         }
1133
1134         if (end <= name)
1135                 return -EINVAL;
1136
1137         /* do dns_resolve upcall */
1138         ip_len = dns_query(NULL, name, end - name, NULL, &ip_addr, NULL);
1139         if (ip_len > 0)
1140                 ret = ceph_pton(ip_addr, ip_len, ss, -1, NULL);
1141         else
1142                 ret = -ESRCH;
1143
1144         kfree(ip_addr);
1145
1146         *ipend = end;
1147
1148         pr_info("resolve '%.*s' (ret=%d): %s\n", (int)(end - name), name,
1149                         ret, ret ? "failed" : ceph_pr_addr(ss));
1150
1151         return ret;
1152 }
1153 #else
1154 static inline int ceph_dns_resolve_name(const char *name, size_t namelen,
1155                 struct sockaddr_storage *ss, char delim, const char **ipend)
1156 {
1157         return -EINVAL;
1158 }
1159 #endif
1160
1161 /*
1162  * Parse a server name (IP or hostname). If a valid IP address is not found
1163  * then try to extract a hostname to resolve using userspace DNS upcall.
1164  */
1165 static int ceph_parse_server_name(const char *name, size_t namelen,
1166                         struct sockaddr_storage *ss, char delim, const char **ipend)
1167 {
1168         int ret;
1169
1170         ret = ceph_pton(name, namelen, ss, delim, ipend);
1171         if (ret)
1172                 ret = ceph_dns_resolve_name(name, namelen, ss, delim, ipend);
1173
1174         return ret;
1175 }
1176
1177 /*
1178  * Parse an ip[:port] list into an addr array.  Use the default
1179  * monitor port if a port isn't specified.
1180  */
1181 int ceph_parse_ips(const char *c, const char *end,
1182                    struct ceph_entity_addr *addr,
1183                    int max_count, int *count)
1184 {
1185         int i, ret = -EINVAL;
1186         const char *p = c;
1187
1188         dout("parse_ips on '%.*s'\n", (int)(end-c), c);
1189         for (i = 0; i < max_count; i++) {
1190                 const char *ipend;
1191                 struct sockaddr_storage *ss = &addr[i].in_addr;
1192                 int port;
1193                 char delim = ',';
1194
1195                 if (*p == '[') {
1196                         delim = ']';
1197                         p++;
1198                 }
1199
1200                 ret = ceph_parse_server_name(p, end - p, ss, delim, &ipend);
1201                 if (ret)
1202                         goto bad;
1203                 ret = -EINVAL;
1204
1205                 p = ipend;
1206
1207                 if (delim == ']') {
1208                         if (*p != ']') {
1209                                 dout("missing matching ']'\n");
1210                                 goto bad;
1211                         }
1212                         p++;
1213                 }
1214
1215                 /* port? */
1216                 if (p < end && *p == ':') {
1217                         port = 0;
1218                         p++;
1219                         while (p < end && *p >= '0' && *p <= '9') {
1220                                 port = (port * 10) + (*p - '0');
1221                                 p++;
1222                         }
1223                         if (port > 65535 || port == 0)
1224                                 goto bad;
1225                 } else {
1226                         port = CEPH_MON_PORT;
1227                 }
1228
1229                 addr_set_port(ss, port);
1230
1231                 dout("parse_ips got %s\n", ceph_pr_addr(ss));
1232
1233                 if (p == end)
1234                         break;
1235                 if (*p != ',')
1236                         goto bad;
1237                 p++;
1238         }
1239
1240         if (p != end)
1241                 goto bad;
1242
1243         if (count)
1244                 *count = i + 1;
1245         return 0;
1246
1247 bad:
1248         pr_err("parse_ips bad ip '%.*s'\n", (int)(end - c), c);
1249         return ret;
1250 }
1251 EXPORT_SYMBOL(ceph_parse_ips);
1252
1253 static int process_banner(struct ceph_connection *con)
1254 {
1255         dout("process_banner on %p\n", con);
1256
1257         if (verify_hello(con) < 0)
1258                 return -1;
1259
1260         ceph_decode_addr(&con->actual_peer_addr);
1261         ceph_decode_addr(&con->peer_addr_for_me);
1262
1263         /*
1264          * Make sure the other end is who we wanted.  note that the other
1265          * end may not yet know their ip address, so if it's 0.0.0.0, give
1266          * them the benefit of the doubt.
1267          */
1268         if (memcmp(&con->peer_addr, &con->actual_peer_addr,
1269                    sizeof(con->peer_addr)) != 0 &&
1270             !(addr_is_blank(&con->actual_peer_addr.in_addr) &&
1271               con->actual_peer_addr.nonce == con->peer_addr.nonce)) {
1272                 pr_warning("wrong peer, want %s/%d, got %s/%d\n",
1273                            ceph_pr_addr(&con->peer_addr.in_addr),
1274                            (int)le32_to_cpu(con->peer_addr.nonce),
1275                            ceph_pr_addr(&con->actual_peer_addr.in_addr),
1276                            (int)le32_to_cpu(con->actual_peer_addr.nonce));
1277                 con->error_msg = "wrong peer at address";
1278                 return -1;
1279         }
1280
1281         /*
1282          * did we learn our address?
1283          */
1284         if (addr_is_blank(&con->msgr->inst.addr.in_addr)) {
1285                 int port = addr_port(&con->msgr->inst.addr.in_addr);
1286
1287                 memcpy(&con->msgr->inst.addr.in_addr,
1288                        &con->peer_addr_for_me.in_addr,
1289                        sizeof(con->peer_addr_for_me.in_addr));
1290                 addr_set_port(&con->msgr->inst.addr.in_addr, port);
1291                 encode_my_addr(con->msgr);
1292                 dout("process_banner learned my addr is %s\n",
1293                      ceph_pr_addr(&con->msgr->inst.addr.in_addr));
1294         }
1295
1296         set_bit(NEGOTIATING, &con->state);
1297         prepare_read_connect(con);
1298         return 0;
1299 }
1300
1301 static void fail_protocol(struct ceph_connection *con)
1302 {
1303         reset_connection(con);
1304         set_bit(CLOSED, &con->state);  /* in case there's queued work */
1305
1306         mutex_unlock(&con->mutex);
1307         if (con->ops->bad_proto)
1308                 con->ops->bad_proto(con);
1309         mutex_lock(&con->mutex);
1310 }
1311
1312 static int process_connect(struct ceph_connection *con)
1313 {
1314         u64 sup_feat = con->msgr->supported_features;
1315         u64 req_feat = con->msgr->required_features;
1316         u64 server_feat = le64_to_cpu(con->in_reply.features);
1317         int ret;
1318
1319         dout("process_connect on %p tag %d\n", con, (int)con->in_tag);
1320
1321         switch (con->in_reply.tag) {
1322         case CEPH_MSGR_TAG_FEATURES:
1323                 pr_err("%s%lld %s feature set mismatch,"
1324                        " my %llx < server's %llx, missing %llx\n",
1325                        ENTITY_NAME(con->peer_name),
1326                        ceph_pr_addr(&con->peer_addr.in_addr),
1327                        sup_feat, server_feat, server_feat & ~sup_feat);
1328                 con->error_msg = "missing required protocol features";
1329                 fail_protocol(con);
1330                 return -1;
1331
1332         case CEPH_MSGR_TAG_BADPROTOVER:
1333                 pr_err("%s%lld %s protocol version mismatch,"
1334                        " my %d != server's %d\n",
1335                        ENTITY_NAME(con->peer_name),
1336                        ceph_pr_addr(&con->peer_addr.in_addr),
1337                        le32_to_cpu(con->out_connect.protocol_version),
1338                        le32_to_cpu(con->in_reply.protocol_version));
1339                 con->error_msg = "protocol version mismatch";
1340                 fail_protocol(con);
1341                 return -1;
1342
1343         case CEPH_MSGR_TAG_BADAUTHORIZER:
1344                 con->auth_retry++;
1345                 dout("process_connect %p got BADAUTHORIZER attempt %d\n", con,
1346                      con->auth_retry);
1347                 if (con->auth_retry == 2) {
1348                         con->error_msg = "connect authorization failure";
1349                         return -1;
1350                 }
1351                 con->auth_retry = 1;
1352                 ret = prepare_write_connect(con->msgr, con, 0);
1353                 if (ret < 0)
1354                         return ret;
1355                 prepare_read_connect(con);
1356                 break;
1357
1358         case CEPH_MSGR_TAG_RESETSESSION:
1359                 /*
1360                  * If we connected with a large connect_seq but the peer
1361                  * has no record of a session with us (no connection, or
1362                  * connect_seq == 0), they will send RESETSESION to indicate
1363                  * that they must have reset their session, and may have
1364                  * dropped messages.
1365                  */
1366                 dout("process_connect got RESET peer seq %u\n",
1367                      le32_to_cpu(con->in_connect.connect_seq));
1368                 pr_err("%s%lld %s connection reset\n",
1369                        ENTITY_NAME(con->peer_name),
1370                        ceph_pr_addr(&con->peer_addr.in_addr));
1371                 reset_connection(con);
1372                 prepare_write_connect(con->msgr, con, 0);
1373                 prepare_read_connect(con);
1374
1375                 /* Tell ceph about it. */
1376                 mutex_unlock(&con->mutex);
1377                 pr_info("reset on %s%lld\n", ENTITY_NAME(con->peer_name));
1378                 if (con->ops->peer_reset)
1379                         con->ops->peer_reset(con);
1380                 mutex_lock(&con->mutex);
1381                 if (test_bit(CLOSED, &con->state) ||
1382                     test_bit(OPENING, &con->state))
1383                         return -EAGAIN;
1384                 break;
1385
1386         case CEPH_MSGR_TAG_RETRY_SESSION:
1387                 /*
1388                  * If we sent a smaller connect_seq than the peer has, try
1389                  * again with a larger value.
1390                  */
1391                 dout("process_connect got RETRY my seq = %u, peer_seq = %u\n",
1392                      le32_to_cpu(con->out_connect.connect_seq),
1393                      le32_to_cpu(con->in_connect.connect_seq));
1394                 con->connect_seq = le32_to_cpu(con->in_connect.connect_seq);
1395                 prepare_write_connect(con->msgr, con, 0);
1396                 prepare_read_connect(con);
1397                 break;
1398
1399         case CEPH_MSGR_TAG_RETRY_GLOBAL:
1400                 /*
1401                  * If we sent a smaller global_seq than the peer has, try
1402                  * again with a larger value.
1403                  */
1404                 dout("process_connect got RETRY_GLOBAL my %u peer_gseq %u\n",
1405                      con->peer_global_seq,
1406                      le32_to_cpu(con->in_connect.global_seq));
1407                 get_global_seq(con->msgr,
1408                                le32_to_cpu(con->in_connect.global_seq));
1409                 prepare_write_connect(con->msgr, con, 0);
1410                 prepare_read_connect(con);
1411                 break;
1412
1413         case CEPH_MSGR_TAG_READY:
1414                 if (req_feat & ~server_feat) {
1415                         pr_err("%s%lld %s protocol feature mismatch,"
1416                                " my required %llx > server's %llx, need %llx\n",
1417                                ENTITY_NAME(con->peer_name),
1418                                ceph_pr_addr(&con->peer_addr.in_addr),
1419                                req_feat, server_feat, req_feat & ~server_feat);
1420                         con->error_msg = "missing required protocol features";
1421                         fail_protocol(con);
1422                         return -1;
1423                 }
1424                 clear_bit(CONNECTING, &con->state);
1425                 con->peer_global_seq = le32_to_cpu(con->in_reply.global_seq);
1426                 con->connect_seq++;
1427                 con->peer_features = server_feat;
1428                 dout("process_connect got READY gseq %d cseq %d (%d)\n",
1429                      con->peer_global_seq,
1430                      le32_to_cpu(con->in_reply.connect_seq),
1431                      con->connect_seq);
1432                 WARN_ON(con->connect_seq !=
1433                         le32_to_cpu(con->in_reply.connect_seq));
1434
1435                 if (con->in_reply.flags & CEPH_MSG_CONNECT_LOSSY)
1436                         set_bit(LOSSYTX, &con->state);
1437
1438                 prepare_read_tag(con);
1439                 break;
1440
1441         case CEPH_MSGR_TAG_WAIT:
1442                 /*
1443                  * If there is a connection race (we are opening
1444                  * connections to each other), one of us may just have
1445                  * to WAIT.  This shouldn't happen if we are the
1446                  * client.
1447                  */
1448                 pr_err("process_connect got WAIT as client\n");
1449                 con->error_msg = "protocol error, got WAIT as client";
1450                 return -1;
1451
1452         default:
1453                 pr_err("connect protocol error, will retry\n");
1454                 con->error_msg = "protocol error, garbage tag during connect";
1455                 return -1;
1456         }
1457         return 0;
1458 }
1459
1460
1461 /*
1462  * read (part of) an ack
1463  */
1464 static int read_partial_ack(struct ceph_connection *con)
1465 {
1466         int to = 0;
1467
1468         return read_partial(con, &to, sizeof(con->in_temp_ack),
1469                             &con->in_temp_ack);
1470 }
1471
1472
1473 /*
1474  * We can finally discard anything that's been acked.
1475  */
1476 static void process_ack(struct ceph_connection *con)
1477 {
1478         struct ceph_msg *m;
1479         u64 ack = le64_to_cpu(con->in_temp_ack);
1480         u64 seq;
1481
1482         while (!list_empty(&con->out_sent)) {
1483                 m = list_first_entry(&con->out_sent, struct ceph_msg,
1484                                      list_head);
1485                 seq = le64_to_cpu(m->hdr.seq);
1486                 if (seq > ack)
1487                         break;
1488                 dout("got ack for seq %llu type %d at %p\n", seq,
1489                      le16_to_cpu(m->hdr.type), m);
1490                 m->ack_stamp = jiffies;
1491                 ceph_msg_remove(m);
1492         }
1493         prepare_read_tag(con);
1494 }
1495
1496
1497
1498
1499 static int read_partial_message_section(struct ceph_connection *con,
1500                                         struct kvec *section,
1501                                         unsigned int sec_len, u32 *crc)
1502 {
1503         int ret, left;
1504
1505         BUG_ON(!section);
1506
1507         while (section->iov_len < sec_len) {
1508                 BUG_ON(section->iov_base == NULL);
1509                 left = sec_len - section->iov_len;
1510                 ret = ceph_tcp_recvmsg(con->sock, (char *)section->iov_base +
1511                                        section->iov_len, left);
1512                 if (ret <= 0)
1513                         return ret;
1514                 section->iov_len += ret;
1515                 if (section->iov_len == sec_len)
1516                         *crc = crc32c(0, section->iov_base,
1517                                       section->iov_len);
1518         }
1519
1520         return 1;
1521 }
1522
1523 static struct ceph_msg *ceph_alloc_msg(struct ceph_connection *con,
1524                                 struct ceph_msg_header *hdr,
1525                                 int *skip);
1526
1527
1528 static int read_partial_message_pages(struct ceph_connection *con,
1529                                       struct page **pages,
1530                                       unsigned data_len, int datacrc)
1531 {
1532         void *p;
1533         int ret;
1534         int left;
1535
1536         left = min((int)(data_len - con->in_msg_pos.data_pos),
1537                    (int)(PAGE_SIZE - con->in_msg_pos.page_pos));
1538         /* (page) data */
1539         BUG_ON(pages == NULL);
1540         p = kmap(pages[con->in_msg_pos.page]);
1541         ret = ceph_tcp_recvmsg(con->sock, p + con->in_msg_pos.page_pos,
1542                                left);
1543         if (ret > 0 && datacrc)
1544                 con->in_data_crc =
1545                         crc32c(con->in_data_crc,
1546                                   p + con->in_msg_pos.page_pos, ret);
1547         kunmap(pages[con->in_msg_pos.page]);
1548         if (ret <= 0)
1549                 return ret;
1550         con->in_msg_pos.data_pos += ret;
1551         con->in_msg_pos.page_pos += ret;
1552         if (con->in_msg_pos.page_pos == PAGE_SIZE) {
1553                 con->in_msg_pos.page_pos = 0;
1554                 con->in_msg_pos.page++;
1555         }
1556
1557         return ret;
1558 }
1559
1560 #ifdef CONFIG_BLOCK
1561 static int read_partial_message_bio(struct ceph_connection *con,
1562                                     struct bio **bio_iter, int *bio_seg,
1563                                     unsigned data_len, int datacrc)
1564 {
1565         struct bio_vec *bv = bio_iovec_idx(*bio_iter, *bio_seg);
1566         void *p;
1567         int ret, left;
1568
1569         if (IS_ERR(bv))
1570                 return PTR_ERR(bv);
1571
1572         left = min((int)(data_len - con->in_msg_pos.data_pos),
1573                    (int)(bv->bv_len - con->in_msg_pos.page_pos));
1574
1575         p = kmap(bv->bv_page) + bv->bv_offset;
1576
1577         ret = ceph_tcp_recvmsg(con->sock, p + con->in_msg_pos.page_pos,
1578                                left);
1579         if (ret > 0 && datacrc)
1580                 con->in_data_crc =
1581                         crc32c(con->in_data_crc,
1582                                   p + con->in_msg_pos.page_pos, ret);
1583         kunmap(bv->bv_page);
1584         if (ret <= 0)
1585                 return ret;
1586         con->in_msg_pos.data_pos += ret;
1587         con->in_msg_pos.page_pos += ret;
1588         if (con->in_msg_pos.page_pos == bv->bv_len) {
1589                 con->in_msg_pos.page_pos = 0;
1590                 iter_bio_next(bio_iter, bio_seg);
1591         }
1592
1593         return ret;
1594 }
1595 #endif
1596
1597 /*
1598  * read (part of) a message.
1599  */
1600 static int read_partial_message(struct ceph_connection *con)
1601 {
1602         struct ceph_msg *m = con->in_msg;
1603         int ret;
1604         int to, left;
1605         unsigned front_len, middle_len, data_len;
1606         int datacrc = con->msgr->nocrc;
1607         int skip;
1608         u64 seq;
1609
1610         dout("read_partial_message con %p msg %p\n", con, m);
1611
1612         /* header */
1613         while (con->in_base_pos < sizeof(con->in_hdr)) {
1614                 left = sizeof(con->in_hdr) - con->in_base_pos;
1615                 ret = ceph_tcp_recvmsg(con->sock,
1616                                        (char *)&con->in_hdr + con->in_base_pos,
1617                                        left);
1618                 if (ret <= 0)
1619                         return ret;
1620                 con->in_base_pos += ret;
1621                 if (con->in_base_pos == sizeof(con->in_hdr)) {
1622                         u32 crc = crc32c(0, (void *)&con->in_hdr,
1623                                  sizeof(con->in_hdr) - sizeof(con->in_hdr.crc));
1624                         if (crc != le32_to_cpu(con->in_hdr.crc)) {
1625                                 pr_err("read_partial_message bad hdr "
1626                                        " crc %u != expected %u\n",
1627                                        crc, con->in_hdr.crc);
1628                                 return -EBADMSG;
1629                         }
1630                 }
1631         }
1632         front_len = le32_to_cpu(con->in_hdr.front_len);
1633         if (front_len > CEPH_MSG_MAX_FRONT_LEN)
1634                 return -EIO;
1635         middle_len = le32_to_cpu(con->in_hdr.middle_len);
1636         if (middle_len > CEPH_MSG_MAX_DATA_LEN)
1637                 return -EIO;
1638         data_len = le32_to_cpu(con->in_hdr.data_len);
1639         if (data_len > CEPH_MSG_MAX_DATA_LEN)
1640                 return -EIO;
1641
1642         /* verify seq# */
1643         seq = le64_to_cpu(con->in_hdr.seq);
1644         if ((s64)seq - (s64)con->in_seq < 1) {
1645                 pr_info("skipping %s%lld %s seq %lld expected %lld\n",
1646                         ENTITY_NAME(con->peer_name),
1647                         ceph_pr_addr(&con->peer_addr.in_addr),
1648                         seq, con->in_seq + 1);
1649                 con->in_base_pos = -front_len - middle_len - data_len -
1650                         sizeof(m->footer);
1651                 con->in_tag = CEPH_MSGR_TAG_READY;
1652                 return 0;
1653         } else if ((s64)seq - (s64)con->in_seq > 1) {
1654                 pr_err("read_partial_message bad seq %lld expected %lld\n",
1655                        seq, con->in_seq + 1);
1656                 con->error_msg = "bad message sequence # for incoming message";
1657                 return -EBADMSG;
1658         }
1659
1660         /* allocate message? */
1661         if (!con->in_msg) {
1662                 dout("got hdr type %d front %d data %d\n", con->in_hdr.type,
1663                      con->in_hdr.front_len, con->in_hdr.data_len);
1664                 skip = 0;
1665                 con->in_msg = ceph_alloc_msg(con, &con->in_hdr, &skip);
1666                 if (skip) {
1667                         /* skip this message */
1668                         dout("alloc_msg said skip message\n");
1669                         BUG_ON(con->in_msg);
1670                         con->in_base_pos = -front_len - middle_len - data_len -
1671                                 sizeof(m->footer);
1672                         con->in_tag = CEPH_MSGR_TAG_READY;
1673                         con->in_seq++;
1674                         return 0;
1675                 }
1676                 if (!con->in_msg) {
1677                         con->error_msg =
1678                                 "error allocating memory for incoming message";
1679                         return -ENOMEM;
1680                 }
1681                 m = con->in_msg;
1682                 m->front.iov_len = 0;    /* haven't read it yet */
1683                 if (m->middle)
1684                         m->middle->vec.iov_len = 0;
1685
1686                 con->in_msg_pos.page = 0;
1687                 if (m->pages)
1688                         con->in_msg_pos.page_pos = m->page_alignment;
1689                 else
1690                         con->in_msg_pos.page_pos = 0;
1691                 con->in_msg_pos.data_pos = 0;
1692         }
1693
1694         /* front */
1695         ret = read_partial_message_section(con, &m->front, front_len,
1696                                            &con->in_front_crc);
1697         if (ret <= 0)
1698                 return ret;
1699
1700         /* middle */
1701         if (m->middle) {
1702                 ret = read_partial_message_section(con, &m->middle->vec,
1703                                                    middle_len,
1704                                                    &con->in_middle_crc);
1705                 if (ret <= 0)
1706                         return ret;
1707         }
1708 #ifdef CONFIG_BLOCK
1709         if (m->bio && !m->bio_iter)
1710                 init_bio_iter(m->bio, &m->bio_iter, &m->bio_seg);
1711 #endif
1712
1713         /* (page) data */
1714         while (con->in_msg_pos.data_pos < data_len) {
1715                 if (m->pages) {
1716                         ret = read_partial_message_pages(con, m->pages,
1717                                                  data_len, datacrc);
1718                         if (ret <= 0)
1719                                 return ret;
1720 #ifdef CONFIG_BLOCK
1721                 } else if (m->bio) {
1722
1723                         ret = read_partial_message_bio(con,
1724                                                  &m->bio_iter, &m->bio_seg,
1725                                                  data_len, datacrc);
1726                         if (ret <= 0)
1727                                 return ret;
1728 #endif
1729                 } else {
1730                         BUG_ON(1);
1731                 }
1732         }
1733
1734         /* footer */
1735         to = sizeof(m->hdr) + sizeof(m->footer);
1736         while (con->in_base_pos < to) {
1737                 left = to - con->in_base_pos;
1738                 ret = ceph_tcp_recvmsg(con->sock, (char *)&m->footer +
1739                                        (con->in_base_pos - sizeof(m->hdr)),
1740                                        left);
1741                 if (ret <= 0)
1742                         return ret;
1743                 con->in_base_pos += ret;
1744         }
1745         dout("read_partial_message got msg %p %d (%u) + %d (%u) + %d (%u)\n",
1746              m, front_len, m->footer.front_crc, middle_len,
1747              m->footer.middle_crc, data_len, m->footer.data_crc);
1748
1749         /* crc ok? */
1750         if (con->in_front_crc != le32_to_cpu(m->footer.front_crc)) {
1751                 pr_err("read_partial_message %p front crc %u != exp. %u\n",
1752                        m, con->in_front_crc, m->footer.front_crc);
1753                 return -EBADMSG;
1754         }
1755         if (con->in_middle_crc != le32_to_cpu(m->footer.middle_crc)) {
1756                 pr_err("read_partial_message %p middle crc %u != exp %u\n",
1757                        m, con->in_middle_crc, m->footer.middle_crc);
1758                 return -EBADMSG;
1759         }
1760         if (datacrc &&
1761             (m->footer.flags & CEPH_MSG_FOOTER_NOCRC) == 0 &&
1762             con->in_data_crc != le32_to_cpu(m->footer.data_crc)) {
1763                 pr_err("read_partial_message %p data crc %u != exp. %u\n", m,
1764                        con->in_data_crc, le32_to_cpu(m->footer.data_crc));
1765                 return -EBADMSG;
1766         }
1767
1768         return 1; /* done! */
1769 }
1770
1771 /*
1772  * Process message.  This happens in the worker thread.  The callback should
1773  * be careful not to do anything that waits on other incoming messages or it
1774  * may deadlock.
1775  */
1776 static void process_message(struct ceph_connection *con)
1777 {
1778         struct ceph_msg *msg;
1779
1780         msg = con->in_msg;
1781         con->in_msg = NULL;
1782
1783         /* if first message, set peer_name */
1784         if (con->peer_name.type == 0)
1785                 con->peer_name = msg->hdr.src;
1786
1787         con->in_seq++;
1788         mutex_unlock(&con->mutex);
1789
1790         dout("===== %p %llu from %s%lld %d=%s len %d+%d (%u %u %u) =====\n",
1791              msg, le64_to_cpu(msg->hdr.seq),
1792              ENTITY_NAME(msg->hdr.src),
1793              le16_to_cpu(msg->hdr.type),
1794              ceph_msg_type_name(le16_to_cpu(msg->hdr.type)),
1795              le32_to_cpu(msg->hdr.front_len),
1796              le32_to_cpu(msg->hdr.data_len),
1797              con->in_front_crc, con->in_middle_crc, con->in_data_crc);
1798         con->ops->dispatch(con, msg);
1799
1800         mutex_lock(&con->mutex);
1801         prepare_read_tag(con);
1802 }
1803
1804
1805 /*
1806  * Write something to the socket.  Called in a worker thread when the
1807  * socket appears to be writeable and we have something ready to send.
1808  */
1809 static int try_write(struct ceph_connection *con)
1810 {
1811         struct ceph_messenger *msgr = con->msgr;
1812         int ret = 1;
1813
1814         dout("try_write start %p state %lu nref %d\n", con, con->state,
1815              atomic_read(&con->nref));
1816
1817 more:
1818         dout("try_write out_kvec_bytes %d\n", con->out_kvec_bytes);
1819
1820         /* open the socket first? */
1821         if (con->sock == NULL) {
1822                 prepare_write_banner(msgr, con);
1823                 prepare_write_connect(msgr, con, 1);
1824                 prepare_read_banner(con);
1825                 set_bit(CONNECTING, &con->state);
1826                 clear_bit(NEGOTIATING, &con->state);
1827
1828                 BUG_ON(con->in_msg);
1829                 con->in_tag = CEPH_MSGR_TAG_READY;
1830                 dout("try_write initiating connect on %p new state %lu\n",
1831                      con, con->state);
1832                 con->sock = ceph_tcp_connect(con);
1833                 if (IS_ERR(con->sock)) {
1834                         con->sock = NULL;
1835                         con->error_msg = "connect error";
1836                         ret = -1;
1837                         goto out;
1838                 }
1839         }
1840
1841 more_kvec:
1842         /* kvec data queued? */
1843         if (con->out_skip) {
1844                 ret = write_partial_skip(con);
1845                 if (ret <= 0)
1846                         goto out;
1847         }
1848         if (con->out_kvec_left) {
1849                 ret = write_partial_kvec(con);
1850                 if (ret <= 0)
1851                         goto out;
1852         }
1853
1854         /* msg pages? */
1855         if (con->out_msg) {
1856                 if (con->out_msg_done) {
1857                         ceph_msg_put(con->out_msg);
1858                         con->out_msg = NULL;   /* we're done with this one */
1859                         goto do_next;
1860                 }
1861
1862                 ret = write_partial_msg_pages(con);
1863                 if (ret == 1)
1864                         goto more_kvec;  /* we need to send the footer, too! */
1865                 if (ret == 0)
1866                         goto out;
1867                 if (ret < 0) {
1868                         dout("try_write write_partial_msg_pages err %d\n",
1869                              ret);
1870                         goto out;
1871                 }
1872         }
1873
1874 do_next:
1875         if (!test_bit(CONNECTING, &con->state)) {
1876                 /* is anything else pending? */
1877                 if (!list_empty(&con->out_queue)) {
1878                         prepare_write_message(con);
1879                         goto more;
1880                 }
1881                 if (con->in_seq > con->in_seq_acked) {
1882                         prepare_write_ack(con);
1883                         goto more;
1884                 }
1885                 if (test_and_clear_bit(KEEPALIVE_PENDING, &con->state)) {
1886                         prepare_write_keepalive(con);
1887                         goto more;
1888                 }
1889         }
1890
1891         /* Nothing to do! */
1892         clear_bit(WRITE_PENDING, &con->state);
1893         dout("try_write nothing else to write.\n");
1894         ret = 0;
1895 out:
1896         dout("try_write done on %p ret %d\n", con, ret);
1897         return ret;
1898 }
1899
1900
1901
1902 /*
1903  * Read what we can from the socket.
1904  */
1905 static int try_read(struct ceph_connection *con)
1906 {
1907         int ret = -1;
1908
1909         if (!con->sock)
1910                 return 0;
1911
1912         if (test_bit(STANDBY, &con->state))
1913                 return 0;
1914
1915         dout("try_read start on %p\n", con);
1916
1917 more:
1918         dout("try_read tag %d in_base_pos %d\n", (int)con->in_tag,
1919              con->in_base_pos);
1920
1921         /*
1922          * process_connect and process_message drop and re-take
1923          * con->mutex.  make sure we handle a racing close or reopen.
1924          */
1925         if (test_bit(CLOSED, &con->state) ||
1926             test_bit(OPENING, &con->state)) {
1927                 ret = -EAGAIN;
1928                 goto out;
1929         }
1930
1931         if (test_bit(CONNECTING, &con->state)) {
1932                 if (!test_bit(NEGOTIATING, &con->state)) {
1933                         dout("try_read connecting\n");
1934                         ret = read_partial_banner(con);
1935                         if (ret <= 0)
1936                                 goto out;
1937                         ret = process_banner(con);
1938                         if (ret < 0)
1939                                 goto out;
1940                 }
1941                 ret = read_partial_connect(con);
1942                 if (ret <= 0)
1943                         goto out;
1944                 ret = process_connect(con);
1945                 if (ret < 0)
1946                         goto out;
1947                 goto more;
1948         }
1949
1950         if (con->in_base_pos < 0) {
1951                 /*
1952                  * skipping + discarding content.
1953                  *
1954                  * FIXME: there must be a better way to do this!
1955                  */
1956                 static char buf[1024];
1957                 int skip = min(1024, -con->in_base_pos);
1958                 dout("skipping %d / %d bytes\n", skip, -con->in_base_pos);
1959                 ret = ceph_tcp_recvmsg(con->sock, buf, skip);
1960                 if (ret <= 0)
1961                         goto out;
1962                 con->in_base_pos += ret;
1963                 if (con->in_base_pos)
1964                         goto more;
1965         }
1966         if (con->in_tag == CEPH_MSGR_TAG_READY) {
1967                 /*
1968                  * what's next?
1969                  */
1970                 ret = ceph_tcp_recvmsg(con->sock, &con->in_tag, 1);
1971                 if (ret <= 0)
1972                         goto out;
1973                 dout("try_read got tag %d\n", (int)con->in_tag);
1974                 switch (con->in_tag) {
1975                 case CEPH_MSGR_TAG_MSG:
1976                         prepare_read_message(con);
1977                         break;
1978                 case CEPH_MSGR_TAG_ACK:
1979                         prepare_read_ack(con);
1980                         break;
1981                 case CEPH_MSGR_TAG_CLOSE:
1982                         set_bit(CLOSED, &con->state);   /* fixme */
1983                         goto out;
1984                 default:
1985                         goto bad_tag;
1986                 }
1987         }
1988         if (con->in_tag == CEPH_MSGR_TAG_MSG) {
1989                 ret = read_partial_message(con);
1990                 if (ret <= 0) {
1991                         switch (ret) {
1992                         case -EBADMSG:
1993                                 con->error_msg = "bad crc";
1994                                 ret = -EIO;
1995                                 break;
1996                         case -EIO:
1997                                 con->error_msg = "io error";
1998                                 break;
1999                         }
2000                         goto out;
2001                 }
2002                 if (con->in_tag == CEPH_MSGR_TAG_READY)
2003                         goto more;
2004                 process_message(con);
2005                 goto more;
2006         }
2007         if (con->in_tag == CEPH_MSGR_TAG_ACK) {
2008                 ret = read_partial_ack(con);
2009                 if (ret <= 0)
2010                         goto out;
2011                 process_ack(con);
2012                 goto more;
2013         }
2014
2015 out:
2016         dout("try_read done on %p ret %d\n", con, ret);
2017         return ret;
2018
2019 bad_tag:
2020         pr_err("try_read bad con->in_tag = %d\n", (int)con->in_tag);
2021         con->error_msg = "protocol error, garbage tag";
2022         ret = -1;
2023         goto out;
2024 }
2025
2026
2027 /*
2028  * Atomically queue work on a connection.  Bump @con reference to
2029  * avoid races with connection teardown.
2030  */
2031 static void queue_con(struct ceph_connection *con)
2032 {
2033         if (test_bit(DEAD, &con->state)) {
2034                 dout("queue_con %p ignoring: DEAD\n",
2035                      con);
2036                 return;
2037         }
2038
2039         if (!con->ops->get(con)) {
2040                 dout("queue_con %p ref count 0\n", con);
2041                 return;
2042         }
2043
2044         if (!queue_delayed_work(ceph_msgr_wq, &con->work, 0)) {
2045                 dout("queue_con %p - already queued\n", con);
2046                 con->ops->put(con);
2047         } else {
2048                 dout("queue_con %p\n", con);
2049         }
2050 }
2051
2052 /*
2053  * Do some work on a connection.  Drop a connection ref when we're done.
2054  */
2055 static void con_work(struct work_struct *work)
2056 {
2057         struct ceph_connection *con = container_of(work, struct ceph_connection,
2058                                                    work.work);
2059         int ret;
2060
2061         mutex_lock(&con->mutex);
2062 restart:
2063         if (test_and_clear_bit(BACKOFF, &con->state)) {
2064                 dout("con_work %p backing off\n", con);
2065                 if (queue_delayed_work(ceph_msgr_wq, &con->work,
2066                                        round_jiffies_relative(con->delay))) {
2067                         dout("con_work %p backoff %lu\n", con, con->delay);
2068                         mutex_unlock(&con->mutex);
2069                         return;
2070                 } else {
2071                         con->ops->put(con);
2072                         dout("con_work %p FAILED to back off %lu\n", con,
2073                              con->delay);
2074                 }
2075         }
2076
2077         if (test_bit(STANDBY, &con->state)) {
2078                 dout("con_work %p STANDBY\n", con);
2079                 goto done;
2080         }
2081         if (test_bit(CLOSED, &con->state)) { /* e.g. if we are replaced */
2082                 dout("con_work CLOSED\n");
2083                 con_close_socket(con);
2084                 goto done;
2085         }
2086         if (test_and_clear_bit(OPENING, &con->state)) {
2087                 /* reopen w/ new peer */
2088                 dout("con_work OPENING\n");
2089                 con_close_socket(con);
2090         }
2091
2092         if (test_and_clear_bit(SOCK_CLOSED, &con->state))
2093                 goto fault;
2094
2095         ret = try_read(con);
2096         if (ret == -EAGAIN)
2097                 goto restart;
2098         if (ret < 0)
2099                 goto fault;
2100
2101         ret = try_write(con);
2102         if (ret == -EAGAIN)
2103                 goto restart;
2104         if (ret < 0)
2105                 goto fault;
2106
2107 done:
2108         mutex_unlock(&con->mutex);
2109 done_unlocked:
2110         con->ops->put(con);
2111         return;
2112
2113 fault:
2114         mutex_unlock(&con->mutex);
2115         ceph_fault(con);     /* error/fault path */
2116         goto done_unlocked;
2117 }
2118
2119
2120 /*
2121  * Generic error/fault handler.  A retry mechanism is used with
2122  * exponential backoff
2123  */
2124 static void ceph_fault(struct ceph_connection *con)
2125 {
2126         pr_err("%s%lld %s %s\n", ENTITY_NAME(con->peer_name),
2127                ceph_pr_addr(&con->peer_addr.in_addr), con->error_msg);
2128         dout("fault %p state %lu to peer %s\n",
2129              con, con->state, ceph_pr_addr(&con->peer_addr.in_addr));
2130
2131         if (test_bit(LOSSYTX, &con->state)) {
2132                 dout("fault on LOSSYTX channel\n");
2133                 goto out;
2134         }
2135
2136         mutex_lock(&con->mutex);
2137         if (test_bit(CLOSED, &con->state))
2138                 goto out_unlock;
2139
2140         con_close_socket(con);
2141
2142         if (con->in_msg) {
2143                 ceph_msg_put(con->in_msg);
2144                 con->in_msg = NULL;
2145         }
2146
2147         /* Requeue anything that hasn't been acked */
2148         list_splice_init(&con->out_sent, &con->out_queue);
2149
2150         /* If there are no messages queued or keepalive pending, place
2151          * the connection in a STANDBY state */
2152         if (list_empty(&con->out_queue) &&
2153             !test_bit(KEEPALIVE_PENDING, &con->state)) {
2154                 dout("fault %p setting STANDBY clearing WRITE_PENDING\n", con);
2155                 clear_bit(WRITE_PENDING, &con->state);
2156                 set_bit(STANDBY, &con->state);
2157         } else {
2158                 /* retry after a delay. */
2159                 if (con->delay == 0)
2160                         con->delay = BASE_DELAY_INTERVAL;
2161                 else if (con->delay < MAX_DELAY_INTERVAL)
2162                         con->delay *= 2;
2163                 con->ops->get(con);
2164                 if (queue_delayed_work(ceph_msgr_wq, &con->work,
2165                                        round_jiffies_relative(con->delay))) {
2166                         dout("fault queued %p delay %lu\n", con, con->delay);
2167                 } else {
2168                         con->ops->put(con);
2169                         dout("fault failed to queue %p delay %lu, backoff\n",
2170                              con, con->delay);
2171                         /*
2172                          * In many cases we see a socket state change
2173                          * while con_work is running and end up
2174                          * queuing (non-delayed) work, such that we
2175                          * can't backoff with a delay.  Set a flag so
2176                          * that when con_work restarts we schedule the
2177                          * delay then.
2178                          */
2179                         set_bit(BACKOFF, &con->state);
2180                 }
2181         }
2182
2183 out_unlock:
2184         mutex_unlock(&con->mutex);
2185 out:
2186         /*
2187          * in case we faulted due to authentication, invalidate our
2188          * current tickets so that we can get new ones.
2189          */
2190         if (con->auth_retry && con->ops->invalidate_authorizer) {
2191                 dout("calling invalidate_authorizer()\n");
2192                 con->ops->invalidate_authorizer(con);
2193         }
2194
2195         if (con->ops->fault)
2196                 con->ops->fault(con);
2197 }
2198
2199
2200
2201 /*
2202  * create a new messenger instance
2203  */
2204 struct ceph_messenger *ceph_messenger_create(struct ceph_entity_addr *myaddr,
2205                                              u32 supported_features,
2206                                              u32 required_features)
2207 {
2208         struct ceph_messenger *msgr;
2209
2210         msgr = kzalloc(sizeof(*msgr), GFP_KERNEL);
2211         if (msgr == NULL)
2212                 return ERR_PTR(-ENOMEM);
2213
2214         msgr->supported_features = supported_features;
2215         msgr->required_features = required_features;
2216
2217         spin_lock_init(&msgr->global_seq_lock);
2218
2219         /* the zero page is needed if a request is "canceled" while the message
2220          * is being written over the socket */
2221         msgr->zero_page = __page_cache_alloc(GFP_KERNEL | __GFP_ZERO);
2222         if (!msgr->zero_page) {
2223                 kfree(msgr);
2224                 return ERR_PTR(-ENOMEM);
2225         }
2226         kmap(msgr->zero_page);
2227
2228         if (myaddr)
2229                 msgr->inst.addr = *myaddr;
2230
2231         /* select a random nonce */
2232         msgr->inst.addr.type = 0;
2233         get_random_bytes(&msgr->inst.addr.nonce, sizeof(msgr->inst.addr.nonce));
2234         encode_my_addr(msgr);
2235
2236         dout("messenger_create %p\n", msgr);
2237         return msgr;
2238 }
2239 EXPORT_SYMBOL(ceph_messenger_create);
2240
2241 void ceph_messenger_destroy(struct ceph_messenger *msgr)
2242 {
2243         dout("destroy %p\n", msgr);
2244         kunmap(msgr->zero_page);
2245         __free_page(msgr->zero_page);
2246         kfree(msgr);
2247         dout("destroyed messenger %p\n", msgr);
2248 }
2249 EXPORT_SYMBOL(ceph_messenger_destroy);
2250
2251 static void clear_standby(struct ceph_connection *con)
2252 {
2253         /* come back from STANDBY? */
2254         if (test_and_clear_bit(STANDBY, &con->state)) {
2255                 mutex_lock(&con->mutex);
2256                 dout("clear_standby %p and ++connect_seq\n", con);
2257                 con->connect_seq++;
2258                 WARN_ON(test_bit(WRITE_PENDING, &con->state));
2259                 WARN_ON(test_bit(KEEPALIVE_PENDING, &con->state));
2260                 mutex_unlock(&con->mutex);
2261         }
2262 }
2263
2264 /*
2265  * Queue up an outgoing message on the given connection.
2266  */
2267 void ceph_con_send(struct ceph_connection *con, struct ceph_msg *msg)
2268 {
2269         if (test_bit(CLOSED, &con->state)) {
2270                 dout("con_send %p closed, dropping %p\n", con, msg);
2271                 ceph_msg_put(msg);
2272                 return;
2273         }
2274
2275         /* set src+dst */
2276         msg->hdr.src = con->msgr->inst.name;
2277
2278         BUG_ON(msg->front.iov_len != le32_to_cpu(msg->hdr.front_len));
2279
2280         msg->needs_out_seq = true;
2281
2282         /* queue */
2283         mutex_lock(&con->mutex);
2284         BUG_ON(!list_empty(&msg->list_head));
2285         list_add_tail(&msg->list_head, &con->out_queue);
2286         dout("----- %p to %s%lld %d=%s len %d+%d+%d -----\n", msg,
2287              ENTITY_NAME(con->peer_name), le16_to_cpu(msg->hdr.type),
2288              ceph_msg_type_name(le16_to_cpu(msg->hdr.type)),
2289              le32_to_cpu(msg->hdr.front_len),
2290              le32_to_cpu(msg->hdr.middle_len),
2291              le32_to_cpu(msg->hdr.data_len));
2292         mutex_unlock(&con->mutex);
2293
2294         /* if there wasn't anything waiting to send before, queue
2295          * new work */
2296         clear_standby(con);
2297         if (test_and_set_bit(WRITE_PENDING, &con->state) == 0)
2298                 queue_con(con);
2299 }
2300 EXPORT_SYMBOL(ceph_con_send);
2301
2302 /*
2303  * Revoke a message that was previously queued for send
2304  */
2305 void ceph_con_revoke(struct ceph_connection *con, struct ceph_msg *msg)
2306 {
2307         mutex_lock(&con->mutex);
2308         if (!list_empty(&msg->list_head)) {
2309                 dout("con_revoke %p msg %p - was on queue\n", con, msg);
2310                 list_del_init(&msg->list_head);
2311                 ceph_msg_put(msg);
2312                 msg->hdr.seq = 0;
2313         }
2314         if (con->out_msg == msg) {
2315                 dout("con_revoke %p msg %p - was sending\n", con, msg);
2316                 con->out_msg = NULL;
2317                 if (con->out_kvec_is_msg) {
2318                         con->out_skip = con->out_kvec_bytes;
2319                         con->out_kvec_is_msg = false;
2320                 }
2321                 ceph_msg_put(msg);
2322                 msg->hdr.seq = 0;
2323         }
2324         mutex_unlock(&con->mutex);
2325 }
2326
2327 /*
2328  * Revoke a message that we may be reading data into
2329  */
2330 void ceph_con_revoke_message(struct ceph_connection *con, struct ceph_msg *msg)
2331 {
2332         mutex_lock(&con->mutex);
2333         if (con->in_msg && con->in_msg == msg) {
2334                 unsigned front_len = le32_to_cpu(con->in_hdr.front_len);
2335                 unsigned middle_len = le32_to_cpu(con->in_hdr.middle_len);
2336                 unsigned data_len = le32_to_cpu(con->in_hdr.data_len);
2337
2338                 /* skip rest of message */
2339                 dout("con_revoke_pages %p msg %p revoked\n", con, msg);
2340                         con->in_base_pos = con->in_base_pos -
2341                                 sizeof(struct ceph_msg_header) -
2342                                 front_len -
2343                                 middle_len -
2344                                 data_len -
2345                                 sizeof(struct ceph_msg_footer);
2346                 ceph_msg_put(con->in_msg);
2347                 con->in_msg = NULL;
2348                 con->in_tag = CEPH_MSGR_TAG_READY;
2349                 con->in_seq++;
2350         } else {
2351                 dout("con_revoke_pages %p msg %p pages %p no-op\n",
2352                      con, con->in_msg, msg);
2353         }
2354         mutex_unlock(&con->mutex);
2355 }
2356
2357 /*
2358  * Queue a keepalive byte to ensure the tcp connection is alive.
2359  */
2360 void ceph_con_keepalive(struct ceph_connection *con)
2361 {
2362         dout("con_keepalive %p\n", con);
2363         clear_standby(con);
2364         if (test_and_set_bit(KEEPALIVE_PENDING, &con->state) == 0 &&
2365             test_and_set_bit(WRITE_PENDING, &con->state) == 0)
2366                 queue_con(con);
2367 }
2368 EXPORT_SYMBOL(ceph_con_keepalive);
2369
2370
2371 /*
2372  * construct a new message with given type, size
2373  * the new msg has a ref count of 1.
2374  */
2375 struct ceph_msg *ceph_msg_new(int type, int front_len, gfp_t flags,
2376                               bool can_fail)
2377 {
2378         struct ceph_msg *m;
2379
2380         m = kmalloc(sizeof(*m), flags);
2381         if (m == NULL)
2382                 goto out;
2383         kref_init(&m->kref);
2384         INIT_LIST_HEAD(&m->list_head);
2385
2386         m->hdr.tid = 0;
2387         m->hdr.type = cpu_to_le16(type);
2388         m->hdr.priority = cpu_to_le16(CEPH_MSG_PRIO_DEFAULT);
2389         m->hdr.version = 0;
2390         m->hdr.front_len = cpu_to_le32(front_len);
2391         m->hdr.middle_len = 0;
2392         m->hdr.data_len = 0;
2393         m->hdr.data_off = 0;
2394         m->hdr.reserved = 0;
2395         m->footer.front_crc = 0;
2396         m->footer.middle_crc = 0;
2397         m->footer.data_crc = 0;
2398         m->footer.flags = 0;
2399         m->front_max = front_len;
2400         m->front_is_vmalloc = false;
2401         m->more_to_follow = false;
2402         m->ack_stamp = 0;
2403         m->pool = NULL;
2404
2405         /* middle */
2406         m->middle = NULL;
2407
2408         /* data */
2409         m->nr_pages = 0;
2410         m->page_alignment = 0;
2411         m->pages = NULL;
2412         m->pagelist = NULL;
2413         m->bio = NULL;
2414         m->bio_iter = NULL;
2415         m->bio_seg = 0;
2416         m->trail = NULL;
2417
2418         /* front */
2419         if (front_len) {
2420                 if (front_len > PAGE_CACHE_SIZE) {
2421                         m->front.iov_base = __vmalloc(front_len, flags,
2422                                                       PAGE_KERNEL);
2423                         m->front_is_vmalloc = true;
2424                 } else {
2425                         m->front.iov_base = kmalloc(front_len, flags);
2426                 }
2427                 if (m->front.iov_base == NULL) {
2428                         dout("ceph_msg_new can't allocate %d bytes\n",
2429                              front_len);
2430                         goto out2;
2431                 }
2432         } else {
2433                 m->front.iov_base = NULL;
2434         }
2435         m->front.iov_len = front_len;
2436
2437         dout("ceph_msg_new %p front %d\n", m, front_len);
2438         return m;
2439
2440 out2:
2441         ceph_msg_put(m);
2442 out:
2443         if (!can_fail) {
2444                 pr_err("msg_new can't create type %d front %d\n", type,
2445                        front_len);
2446                 WARN_ON(1);
2447         } else {
2448                 dout("msg_new can't create type %d front %d\n", type,
2449                      front_len);
2450         }
2451         return NULL;
2452 }
2453 EXPORT_SYMBOL(ceph_msg_new);
2454
2455 /*
2456  * Allocate "middle" portion of a message, if it is needed and wasn't
2457  * allocated by alloc_msg.  This allows us to read a small fixed-size
2458  * per-type header in the front and then gracefully fail (i.e.,
2459  * propagate the error to the caller based on info in the front) when
2460  * the middle is too large.
2461  */
2462 static int ceph_alloc_middle(struct ceph_connection *con, struct ceph_msg *msg)
2463 {
2464         int type = le16_to_cpu(msg->hdr.type);
2465         int middle_len = le32_to_cpu(msg->hdr.middle_len);
2466
2467         dout("alloc_middle %p type %d %s middle_len %d\n", msg, type,
2468              ceph_msg_type_name(type), middle_len);
2469         BUG_ON(!middle_len);
2470         BUG_ON(msg->middle);
2471
2472         msg->middle = ceph_buffer_new(middle_len, GFP_NOFS);
2473         if (!msg->middle)
2474                 return -ENOMEM;
2475         return 0;
2476 }
2477
2478 /*
2479  * Generic message allocator, for incoming messages.
2480  */
2481 static struct ceph_msg *ceph_alloc_msg(struct ceph_connection *con,
2482                                 struct ceph_msg_header *hdr,
2483                                 int *skip)
2484 {
2485         int type = le16_to_cpu(hdr->type);
2486         int front_len = le32_to_cpu(hdr->front_len);
2487         int middle_len = le32_to_cpu(hdr->middle_len);
2488         struct ceph_msg *msg = NULL;
2489         int ret;
2490
2491         if (con->ops->alloc_msg) {
2492                 mutex_unlock(&con->mutex);
2493                 msg = con->ops->alloc_msg(con, hdr, skip);
2494                 mutex_lock(&con->mutex);
2495                 if (!msg || *skip)
2496                         return NULL;
2497         }
2498         if (!msg) {
2499                 *skip = 0;
2500                 msg = ceph_msg_new(type, front_len, GFP_NOFS, false);
2501                 if (!msg) {
2502                         pr_err("unable to allocate msg type %d len %d\n",
2503                                type, front_len);
2504                         return NULL;
2505                 }
2506                 msg->page_alignment = le16_to_cpu(hdr->data_off);
2507         }
2508         memcpy(&msg->hdr, &con->in_hdr, sizeof(con->in_hdr));
2509
2510         if (middle_len && !msg->middle) {
2511                 ret = ceph_alloc_middle(con, msg);
2512                 if (ret < 0) {
2513                         ceph_msg_put(msg);
2514                         return NULL;
2515                 }
2516         }
2517
2518         return msg;
2519 }
2520
2521
2522 /*
2523  * Free a generically kmalloc'd message.
2524  */
2525 void ceph_msg_kfree(struct ceph_msg *m)
2526 {
2527         dout("msg_kfree %p\n", m);
2528         if (m->front_is_vmalloc)
2529                 vfree(m->front.iov_base);
2530         else
2531                 kfree(m->front.iov_base);
2532         kfree(m);
2533 }
2534
2535 /*
2536  * Drop a msg ref.  Destroy as needed.
2537  */
2538 void ceph_msg_last_put(struct kref *kref)
2539 {
2540         struct ceph_msg *m = container_of(kref, struct ceph_msg, kref);
2541
2542         dout("ceph_msg_put last one on %p\n", m);
2543         WARN_ON(!list_empty(&m->list_head));
2544
2545         /* drop middle, data, if any */
2546         if (m->middle) {
2547                 ceph_buffer_put(m->middle);
2548                 m->middle = NULL;
2549         }
2550         m->nr_pages = 0;
2551         m->pages = NULL;
2552
2553         if (m->pagelist) {
2554                 ceph_pagelist_release(m->pagelist);
2555                 kfree(m->pagelist);
2556                 m->pagelist = NULL;
2557         }
2558
2559         m->trail = NULL;
2560
2561         if (m->pool)
2562                 ceph_msgpool_put(m->pool, m);
2563         else
2564                 ceph_msg_kfree(m);
2565 }
2566 EXPORT_SYMBOL(ceph_msg_last_put);
2567
2568 void ceph_msg_dump(struct ceph_msg *msg)
2569 {
2570         pr_debug("msg_dump %p (front_max %d nr_pages %d)\n", msg,
2571                  msg->front_max, msg->nr_pages);
2572         print_hex_dump(KERN_DEBUG, "header: ",
2573                        DUMP_PREFIX_OFFSET, 16, 1,
2574                        &msg->hdr, sizeof(msg->hdr), true);
2575         print_hex_dump(KERN_DEBUG, " front: ",
2576                        DUMP_PREFIX_OFFSET, 16, 1,
2577                        msg->front.iov_base, msg->front.iov_len, true);
2578         if (msg->middle)
2579                 print_hex_dump(KERN_DEBUG, "middle: ",
2580                                DUMP_PREFIX_OFFSET, 16, 1,
2581                                msg->middle->vec.iov_base,
2582                                msg->middle->vec.iov_len, true);
2583         print_hex_dump(KERN_DEBUG, "footer: ",
2584                        DUMP_PREFIX_OFFSET, 16, 1,
2585                        &msg->footer, sizeof(msg->footer), true);
2586 }
2587 EXPORT_SYMBOL(ceph_msg_dump);