Fix: honor send timeout on unix socket connect

Needed if we want to hold the ust_lock() while we connect to the session
daemon without blocking the application forever if the session daemon is
hung on SIGSTOP.

This only triggers if we launchs _many_ applications with a session
daemon SIGSTOP'd (e.g. 1000 in parallel), so we fill the socket queue,
and applications hang there until the session daemon is SIGCONT'd.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

diff --git a/include/ust-comm.h b/include/ust-comm.h
index fb0c0bbc773b854bfc97af19b93a18e5e7c1b914..6a08b306b1b9ade6b44cf722d8b3d26a73281600 100644 (file)
--- a/include/ust-comm.h
+++ b/include/ust-comm.h
@@ -190,7 +190,8 @@ struct ustcomm_notify_channel_reply {
  */
 
 extern int ustcomm_create_unix_sock(const char *pathname);
-extern int ustcomm_connect_unix_sock(const char *pathname);
+extern int ustcomm_connect_unix_sock(const char *pathname,
+               long timeout);
 extern int ustcomm_accept_unix_sock(int sock);
 extern int ustcomm_listen_unix_sock(int sock);
 extern int ustcomm_close_unix_sock(int sock);

diff --git a/liblttng-ust-comm/lttng-ust-comm.c b/liblttng-ust-comm/lttng-ust-comm.c
index ecc23821fb6fa29624dcbc25564c5011eed4aafb..7ba23c527ce0e9b3c2c621e4fe66acf3af33c37d 100644 (file)
--- a/liblttng-ust-comm/lttng-ust-comm.c
+++ b/liblttng-ust-comm/lttng-ust-comm.c
@@ -95,7 +95,7 @@ const char *lttng_ust_strerror(int code)
  *
  * Connect to unix socket using the path name.
  */
-int ustcomm_connect_unix_sock(const char *pathname)
+int ustcomm_connect_unix_sock(const char *pathname, long timeout)
 {
        struct sockaddr_un sun;
        int fd, ret;
@@ -110,6 +110,15 @@ int ustcomm_connect_unix_sock(const char *pathname)
                ret = -errno;
                goto error;
        }
+       if (timeout >= 0) {
+               /* Give at least 10ms. */
+               if (timeout < 10)
+                       timeout = 10;
+               ret = ustcomm_setsockopt_snd_timeout(fd, timeout);
+               if (ret < 0) {
+                       WARN("Error setting connect socket send timeout");
+               }
+       }
        ret = fcntl(fd, F_SETFD, FD_CLOEXEC);
        if (ret < 0) {
                PERROR("fcntl");

diff --git a/liblttng-ust/lttng-ust-comm.c b/liblttng-ust/lttng-ust-comm.c
index 07c1be09eb46aa2f53903ebb91f42369c290aa87..34b930d284f891c96be11a87f1fd2ff4c924c307 100644 (file)
--- a/liblttng-ust/lttng-ust-comm.c
+++ b/liblttng-ust/lttng-ust-comm.c
@@ -451,7 +451,7 @@ int setup_local_apps(void)
 }
 
 /*
- * Get notify_sock timeout, in ms.
+ * Get socket timeout, in ms.
  * -1: wait forever. 0: don't wait. >0: timeout, in ms.
  */
 static
@@ -471,12 +471,20 @@ long get_timeout(void)
        return constructor_delay_ms;
 }
 
+/* Timeout for notify socket send and recv. */
 static
 long get_notify_sock_timeout(void)
 {
        return get_timeout();
 }
 
+/* Timeout for connecting to cmd and notify sockets. */
+static
+long get_connect_sock_timeout(void)
+{
+       return get_timeout();
+}
+
 /*
  * Return values: -1: wait forever. 0: don't wait. 1: timeout wait.
  */
@@ -1351,7 +1359,8 @@ restart:
         * first connect registration message.
         */
        /* Connect cmd socket */
-       ret = ustcomm_connect_unix_sock(sock_info->sock_path);
+       ret = ustcomm_connect_unix_sock(sock_info->sock_path,
+               get_connect_sock_timeout());
        if (ret < 0) {
                DBG("Info: sessiond not accepting connections to %s apps socket", sock_info->name);
                prev_connect_failed = 1;
@@ -1407,7 +1416,8 @@ restart:
        ust_unlock();
 
        /* Connect notify socket */
-       ret = ustcomm_connect_unix_sock(sock_info->sock_path);
+       ret = ustcomm_connect_unix_sock(sock_info->sock_path,
+               get_connect_sock_timeout());
        if (ret < 0) {
                DBG("Info: sessiond not accepting connections to %s apps socket", sock_info->name);
                prev_connect_failed = 1;