--- /dev/null
+/*
+ * Copyright (C) 2012 David Goulet <dgoulet@efficios.com>
+ *
+ * SPDX-License-Identifier: GPL-2.0-only
+ *
+ */
+
+#define _LGPL_SOURCE
+#include <algorithm>
+#include <limits.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <unistd.h>
+#include <fcntl.h>
+#include <common/compat/time.h>
+#include <poll.h>
+
+#include <common/common.h>
+#include <common/time.h>
+#include <common/compat/errno.h>
+
+#include "inet.h"
+
+#define RECONNECT_DELAY 200 /* ms */
+
+/*
+ * INET protocol operations.
+ */
+static const struct lttcomm_proto_ops inet_ops = {
+ .bind = lttcomm_bind_inet_sock,
+ .close = lttcomm_close_inet_sock,
+ .connect = lttcomm_connect_inet_sock,
+ .accept = lttcomm_accept_inet_sock,
+ .listen = lttcomm_listen_inet_sock,
+ .recvmsg = lttcomm_recvmsg_inet_sock,
+ .sendmsg = lttcomm_sendmsg_inet_sock,
+};
+
+unsigned long lttcomm_inet_tcp_timeout;
+
+/*
+ * Creates an PF_INET socket.
+ */
+int lttcomm_create_inet_sock(struct lttcomm_sock *sock, int type, int proto)
+{
+ int val = 1, ret;
+ unsigned long timeout;
+
+ /* Create server socket */
+ if ((sock->fd = socket(PF_INET, type, proto)) < 0) {
+ PERROR("socket inet");
+ goto error;
+ }
+
+ sock->ops = &inet_ops;
+
+ /*
+ * Set socket option to reuse the address.
+ */
+ ret = setsockopt(sock->fd, SOL_SOCKET, SO_REUSEADDR, &val, sizeof(int));
+ if (ret < 0) {
+ PERROR("setsockopt inet");
+ goto error;
+ }
+ timeout = lttcomm_get_network_timeout();
+ if (timeout) {
+ ret = lttcomm_setsockopt_rcv_timeout(sock->fd, timeout);
+ if (ret) {
+ goto error;
+ }
+ ret = lttcomm_setsockopt_snd_timeout(sock->fd, timeout);
+ if (ret) {
+ goto error;
+ }
+ }
+
+ return 0;
+
+error:
+ return -1;
+}
+
+/*
+ * Bind socket and return.
+ */
+int lttcomm_bind_inet_sock(struct lttcomm_sock *sock)
+{
+ struct sockaddr_in sockaddr = sock->sockaddr.addr.sin;
+
+ return bind(sock->fd, (struct sockaddr *) &sockaddr, sizeof(sockaddr));
+}
+
+static
+int connect_no_timeout(struct lttcomm_sock *sock)
+{
+ struct sockaddr_in sockaddr = sock->sockaddr.addr.sin;
+
+ return connect(sock->fd, (struct sockaddr *) &sockaddr, sizeof(sockaddr));
+}
+
+static
+int connect_with_timeout(struct lttcomm_sock *sock)
+{
+ unsigned long timeout = lttcomm_get_network_timeout();
+ int ret, flags, connect_ret;
+ struct timespec orig_time, cur_time;
+ unsigned long diff_ms;
+ struct sockaddr_in sockaddr;
+
+ ret = fcntl(sock->fd, F_GETFL, 0);
+ if (ret == -1) {
+ PERROR("fcntl");
+ return -1;
+ }
+ flags = ret;
+
+ /* Set socket to nonblock */
+ ret = fcntl(sock->fd, F_SETFL, flags | O_NONBLOCK);
+ if (ret == -1) {
+ PERROR("fcntl");
+ return -1;
+ }
+
+ ret = lttng_clock_gettime(CLOCK_MONOTONIC, &orig_time);
+ if (ret == -1) {
+ PERROR("clock_gettime");
+ return -1;
+ }
+
+ sockaddr = sock->sockaddr.addr.sin;
+ connect_ret = connect(sock->fd, (struct sockaddr *) &sockaddr, sizeof(sockaddr));
+ if (connect_ret == -1 && errno != EAGAIN && errno != EWOULDBLOCK &&
+ errno != EINPROGRESS) {
+ goto error;
+ } else if (!connect_ret) {
+ /* Connect succeeded */
+ goto success;
+ }
+
+ DBG("Asynchronous connect for sock %d, performing polling with"
+ " timeout: %lums", sock->fd, timeout);
+ /*
+ * Perform poll loop following EINPROGRESS recommendation from
+ * connect(2) man page.
+ */
+ do {
+ struct pollfd fds;
+
+ fds.fd = sock->fd;
+ fds.events = POLLOUT;
+ fds.revents = 0;
+ ret = poll(&fds, 1, RECONNECT_DELAY);
+ if (ret < 0) {
+ goto error;
+ } else if (ret > 0) {
+ int optval;
+ socklen_t optval_len = sizeof(optval);
+
+ if (!(fds.revents & POLLOUT)) {
+ /* Either hup or error */
+ errno = EPIPE;
+ goto error;
+ }
+ /* got something */
+ ret = getsockopt(sock->fd, SOL_SOCKET,
+ SO_ERROR, &optval, &optval_len);
+ if (ret) {
+ PERROR("getsockopt");
+ goto error;
+ }
+ if (!optval) {
+ connect_ret = 0;
+ goto success;
+ } else {
+ /* Get actual connect() errno from opt_val */
+ errno = optval;
+ goto error;
+ }
+ }
+ /* ret == 0: timeout */
+ ret = lttng_clock_gettime(CLOCK_MONOTONIC, &cur_time);
+ if (ret == -1) {
+ PERROR("clock_gettime");
+ connect_ret = ret;
+ goto error;
+ }
+ if (timespec_to_ms(timespec_abs_diff(cur_time, orig_time), &diff_ms) < 0) {
+ ERR("timespec_to_ms input overflows milliseconds output");
+ connect_ret = -1;
+ goto error;
+ }
+ } while (diff_ms < timeout);
+
+ /* Timeout */
+ errno = ETIMEDOUT;
+ connect_ret = -1;
+
+success:
+ /* Restore initial flags */
+ ret = fcntl(sock->fd, F_SETFL, flags);
+ if (ret == -1) {
+ PERROR("fcntl");
+ /* Continue anyway */
+ }
+error:
+ return connect_ret;
+}
+
+/*
+ * Connect PF_INET socket.
+ */
+int lttcomm_connect_inet_sock(struct lttcomm_sock *sock)
+{
+ int ret, closeret;
+
+ if (lttcomm_get_network_timeout()) {
+ ret = connect_with_timeout(sock);
+ } else {
+ ret = connect_no_timeout(sock);
+ }
+ if (ret < 0) {
+ PERROR("connect");
+ goto error_connect;
+ }
+
+ return ret;
+
+error_connect:
+ closeret = close(sock->fd);
+ if (closeret) {
+ PERROR("close inet");
+ }
+
+ return ret;
+}
+
+/*
+ * Do an accept(2) on the sock and return the new lttcomm socket. The socket
+ * MUST be bind(2) before.
+ */
+struct lttcomm_sock *lttcomm_accept_inet_sock(struct lttcomm_sock *sock)
+{
+ int new_fd;
+ socklen_t len;
+ struct lttcomm_sock *new_sock;
+ unsigned long timeout;
+ struct sockaddr_in new_addr = {};
+
+ if (sock->proto == LTTCOMM_SOCK_UDP) {
+ /*
+ * accept(2) does not exist for UDP so simply return the passed socket.
+ */
+ new_sock = sock;
+ goto end;
+ }
+
+ new_sock = lttcomm_alloc_sock(sock->proto);
+ if (new_sock == NULL) {
+ goto error;
+ }
+
+ len = sizeof(new_addr);
+
+ /* Blocking call */
+ new_fd = accept(sock->fd, (struct sockaddr *) &new_addr, &len);
+ if (new_fd < 0) {
+ PERROR("accept inet");
+ goto error;
+ }
+ new_sock->sockaddr.addr.sin = new_addr;
+ timeout = lttcomm_get_network_timeout();
+ if (timeout) {
+ int ret;
+
+ ret = lttcomm_setsockopt_rcv_timeout(new_fd, timeout);
+ if (ret) {
+ goto error_close;
+ }
+ ret = lttcomm_setsockopt_snd_timeout(new_fd, timeout);
+ if (ret) {
+ goto error_close;
+ }
+ }
+
+ new_sock->fd = new_fd;
+ new_sock->ops = &inet_ops;
+
+end:
+ return new_sock;
+
+error_close:
+ if (close(new_fd) < 0) {
+ PERROR("accept inet close fd");
+ }
+
+error:
+ free(new_sock);
+ return NULL;
+}
+
+/*
+ * Make the socket listen using LTTNG_SESSIOND_COMM_MAX_LISTEN.
+ */
+int lttcomm_listen_inet_sock(struct lttcomm_sock *sock, int backlog)
+{
+ int ret;
+
+ if (sock->proto == LTTCOMM_SOCK_UDP) {
+ /* listen(2) does not exist for UDP so simply return success. */
+ ret = 0;
+ goto end;
+ }
+
+ /* Default listen backlog */
+ if (backlog <= 0) {
+ backlog = LTTNG_SESSIOND_COMM_MAX_LISTEN;
+ }
+
+ ret = listen(sock->fd, backlog);
+ if (ret < 0) {
+ PERROR("listen inet");
+ }
+
+end:
+ return ret;
+}
+
+/*
+ * Receive data of size len in put that data into the buf param. Using recvmsg
+ * API.
+ *
+ * Return the size of received data.
+ */
+ssize_t lttcomm_recvmsg_inet_sock(struct lttcomm_sock *sock, void *buf,
+ size_t len, int flags)
+{
+ struct msghdr msg;
+ struct iovec iov[1];
+ ssize_t ret = -1;
+ size_t len_last;
+ struct sockaddr_in addr = sock->sockaddr.addr.sin;
+
+ memset(&msg, 0, sizeof(msg));
+
+ iov[0].iov_base = buf;
+ iov[0].iov_len = len;
+ msg.msg_iov = iov;
+ msg.msg_iovlen = 1;
+
+ msg.msg_name = (struct sockaddr *) &addr;
+ msg.msg_namelen = sizeof(sock->sockaddr.addr.sin);
+
+ do {
+ len_last = iov[0].iov_len;
+ ret = recvmsg(sock->fd, &msg, flags);
+ if (ret > 0) {
+ if (flags & MSG_DONTWAIT) {
+ goto end;
+ }
+ iov[0].iov_base = ((char *) iov[0].iov_base) + ret;
+ iov[0].iov_len -= ret;
+ LTTNG_ASSERT(ret <= len_last);
+ }
+ } while ((ret > 0 && ret < len_last) || (ret < 0 && errno == EINTR));
+
+ if (ret < 0) {
+ if (errno == EAGAIN && flags & MSG_DONTWAIT) {
+ /*
+ * EAGAIN is expected in non-blocking mode and should
+ * not be reported as an error. Moreover, if no data
+ * was read, 0 must not be returned as it would be
+ * interpreted as an orderly shutdown of the socket.
+ */
+ goto end;
+ }
+ PERROR("recvmsg inet");
+ } else if (ret > 0) {
+ ret = len;
+ }
+ /* Else ret = 0 meaning an orderly shutdown. */
+end:
+ return ret;
+}
+
+/*
+ * Send buf data of size len. Using sendmsg API.
+ *
+ * Return the size of sent data.
+ */
+ssize_t lttcomm_sendmsg_inet_sock(struct lttcomm_sock *sock, const void *buf,
+ size_t len, int flags)
+{
+ struct msghdr msg;
+ struct iovec iov[1];
+ ssize_t ret = -1;
+
+ memset(&msg, 0, sizeof(msg));
+
+ iov[0].iov_base = (void *) buf;
+ iov[0].iov_len = len;
+ msg.msg_iov = iov;
+ msg.msg_iovlen = 1;
+
+ switch (sock->proto) {
+ case LTTCOMM_SOCK_UDP:
+ {
+ struct sockaddr_in addr = sock->sockaddr.addr.sin;
+
+ msg.msg_name = (struct sockaddr *) &addr;
+ msg.msg_namelen = sizeof(sock->sockaddr.addr.sin);
+ break;
+ }
+ default:
+ break;
+ }
+
+ do {
+ ret = sendmsg(sock->fd, &msg, flags);
+ } while (ret < 0 && errno == EINTR);
+ if (ret < 0) {
+ /*
+ * Only warn about EPIPE when quiet mode is deactivated.
+ * We consider EPIPE as expected.
+ */
+ if (errno != EPIPE || !lttng_opt_quiet) {
+ PERROR("sendmsg inet");
+ }
+ }
+
+ return ret;
+}
+
+/*
+ * Shutdown cleanly and close.
+ */
+int lttcomm_close_inet_sock(struct lttcomm_sock *sock)
+{
+ int ret;
+
+ /* Don't try to close an invalid marked socket */
+ if (sock->fd == -1) {
+ return 0;
+ }
+
+ ret = close(sock->fd);
+ if (ret) {
+ PERROR("close inet");
+ }
+
+ /* Mark socket */
+ sock->fd = -1;
+
+ return ret;
+}
+
+/*
+ * Return value read from /proc or else 0 if value is not found.
+ */
+static unsigned long read_proc_value(const char *path)
+{
+ int ret, fd;
+ ssize_t size_ret;
+ long r_val;
+ unsigned long val = 0;
+ char buf[64];
+
+ fd = open(path, O_RDONLY);
+ if (fd < 0) {
+ goto error;
+ }
+
+ size_ret = lttng_read(fd, buf, sizeof(buf));
+ /*
+ * Allow reading a file smaller than buf, but keep space for
+ * final \0.
+ */
+ if (size_ret < 0 || size_ret >= sizeof(buf)) {
+ PERROR("read proc failed");
+ goto error_close;
+ }
+ buf[size_ret] = '\0';
+
+ errno = 0;
+ r_val = strtol(buf, NULL, 10);
+ if (errno != 0 || r_val < -1L) {
+ val = 0;
+ goto error_close;
+ } else {
+ if (r_val > 0) {
+ val = r_val;
+ }
+ }
+
+error_close:
+ ret = close(fd);
+ if (ret) {
+ PERROR("close /proc value");
+ }
+error:
+ return val;
+}
+
+void lttcomm_inet_init(void)
+{
+ unsigned long syn_retries, fin_timeout, syn_timeout, env;
+
+ env = lttcomm_get_network_timeout();
+ if (env) {
+ lttcomm_inet_tcp_timeout = env;
+ goto end;
+ }
+
+ /* Assign default value and see if we can change it. */
+ lttcomm_inet_tcp_timeout = DEFAULT_INET_TCP_TIMEOUT;
+
+ syn_retries = read_proc_value(LTTCOMM_INET_PROC_SYN_RETRIES_PATH);
+ fin_timeout = read_proc_value(LTTCOMM_INET_PROC_FIN_TIMEOUT_PATH);
+
+ syn_timeout = syn_retries * LTTCOMM_INET_SYN_TIMEOUT_FACTOR;
+
+ /*
+ * Get the maximum between the two possible timeout value and use that to
+ * get the maximum with the default timeout.
+ */
+ lttcomm_inet_tcp_timeout = std::max(std::max(syn_timeout, fin_timeout),
+ lttcomm_inet_tcp_timeout);
+
+end:
+ DBG("TCP inet operation timeout set to %lu sec", lttcomm_inet_tcp_timeout);
+}