tcp_connect.c 4.6 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170
  1. /*
  2. * Copyright (c) 2006 Oracle. All rights reserved.
  3. *
  4. * This software is available to you under a choice of one of two
  5. * licenses. You may choose to be licensed under the terms of the GNU
  6. * General Public License (GPL) Version 2, available from the file
  7. * COPYING in the main directory of this source tree, or the
  8. * OpenIB.org BSD license below:
  9. *
  10. * Redistribution and use in source and binary forms, with or
  11. * without modification, are permitted provided that the following
  12. * conditions are met:
  13. *
  14. * - Redistributions of source code must retain the above
  15. * copyright notice, this list of conditions and the following
  16. * disclaimer.
  17. *
  18. * - Redistributions in binary form must reproduce the above
  19. * copyright notice, this list of conditions and the following
  20. * disclaimer in the documentation and/or other materials
  21. * provided with the distribution.
  22. *
  23. * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
  24. * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
  25. * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
  26. * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
  27. * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
  28. * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  29. * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
  30. * SOFTWARE.
  31. *
  32. */
  33. #include <linux/kernel.h>
  34. #include <linux/in.h>
  35. #include <net/tcp.h>
  36. #include "rds_single_path.h"
  37. #include "rds.h"
  38. #include "tcp.h"
  39. void rds_tcp_state_change(struct sock *sk)
  40. {
  41. void (*state_change)(struct sock *sk);
  42. struct rds_connection *conn;
  43. struct rds_tcp_connection *tc;
  44. read_lock_bh(&sk->sk_callback_lock);
  45. conn = sk->sk_user_data;
  46. if (!conn) {
  47. state_change = sk->sk_state_change;
  48. goto out;
  49. }
  50. tc = conn->c_transport_data;
  51. state_change = tc->t_orig_state_change;
  52. rdsdebug("sock %p state_change to %d\n", tc->t_sock, sk->sk_state);
  53. switch (sk->sk_state) {
  54. /* ignore connecting sockets as they make progress */
  55. case TCP_SYN_SENT:
  56. case TCP_SYN_RECV:
  57. break;
  58. case TCP_ESTABLISHED:
  59. rds_connect_path_complete(&conn->c_path[0],
  60. RDS_CONN_CONNECTING);
  61. break;
  62. case TCP_CLOSE_WAIT:
  63. case TCP_CLOSE:
  64. rds_conn_drop(conn);
  65. default:
  66. break;
  67. }
  68. out:
  69. read_unlock_bh(&sk->sk_callback_lock);
  70. state_change(sk);
  71. }
  72. int rds_tcp_conn_connect(struct rds_connection *conn)
  73. {
  74. struct socket *sock = NULL;
  75. struct sockaddr_in src, dest;
  76. int ret;
  77. struct rds_tcp_connection *tc = conn->c_transport_data;
  78. mutex_lock(&tc->t_conn_lock);
  79. if (rds_conn_up(conn)) {
  80. mutex_unlock(&tc->t_conn_lock);
  81. return 0;
  82. }
  83. ret = sock_create_kern(rds_conn_net(conn), PF_INET,
  84. SOCK_STREAM, IPPROTO_TCP, &sock);
  85. if (ret < 0)
  86. goto out;
  87. rds_tcp_tune(sock);
  88. src.sin_family = AF_INET;
  89. src.sin_addr.s_addr = (__force u32)conn->c_laddr;
  90. src.sin_port = (__force u16)htons(0);
  91. ret = sock->ops->bind(sock, (struct sockaddr *)&src, sizeof(src));
  92. if (ret) {
  93. rdsdebug("bind failed with %d at address %pI4\n",
  94. ret, &conn->c_laddr);
  95. goto out;
  96. }
  97. dest.sin_family = AF_INET;
  98. dest.sin_addr.s_addr = (__force u32)conn->c_faddr;
  99. dest.sin_port = (__force u16)htons(RDS_TCP_PORT);
  100. /*
  101. * once we call connect() we can start getting callbacks and they
  102. * own the socket
  103. */
  104. rds_tcp_set_callbacks(sock, conn);
  105. ret = sock->ops->connect(sock, (struct sockaddr *)&dest, sizeof(dest),
  106. O_NONBLOCK);
  107. rdsdebug("connect to address %pI4 returned %d\n", &conn->c_faddr, ret);
  108. if (ret == -EINPROGRESS)
  109. ret = 0;
  110. if (ret == 0) {
  111. rds_tcp_keepalive(sock);
  112. sock = NULL;
  113. } else {
  114. rds_tcp_restore_callbacks(sock, conn->c_transport_data);
  115. }
  116. out:
  117. mutex_unlock(&tc->t_conn_lock);
  118. if (sock)
  119. sock_release(sock);
  120. return ret;
  121. }
  122. /*
  123. * Before killing the tcp socket this needs to serialize with callbacks. The
  124. * caller has already grabbed the sending sem so we're serialized with other
  125. * senders.
  126. *
  127. * TCP calls the callbacks with the sock lock so we hold it while we reset the
  128. * callbacks to those set by TCP. Our callbacks won't execute again once we
  129. * hold the sock lock.
  130. */
  131. void rds_tcp_conn_path_shutdown(struct rds_conn_path *cp)
  132. {
  133. struct rds_tcp_connection *tc = cp->cp_transport_data;
  134. struct socket *sock = tc->t_sock;
  135. rdsdebug("shutting down conn %p tc %p sock %p\n",
  136. cp->cp_conn, tc, sock);
  137. if (sock) {
  138. sock->ops->shutdown(sock, RCV_SHUTDOWN | SEND_SHUTDOWN);
  139. lock_sock(sock->sk);
  140. rds_tcp_restore_callbacks(sock, tc); /* tc->tc_sock = NULL */
  141. release_sock(sock->sk);
  142. sock_release(sock);
  143. }
  144. if (tc->t_tinc) {
  145. rds_inc_put(&tc->t_tinc->ti_inc);
  146. tc->t_tinc = NULL;
  147. }
  148. tc->t_tinc_hdr_rem = sizeof(struct rds_header);
  149. tc->t_tinc_data_rem = 0;
  150. }