Fix: liblttng-ust-fork deadlock

[lttng-ust.git] / liblttng-ust / lttng-ust-comm.c

diff --git a/liblttng-ust/lttng-ust-comm.c b/liblttng-ust/lttng-ust-comm.c
index 0116b5f744160cab7d4077146d4a532f606359d8..8c864f8ce595621434cfac4a08dec1aa705d5671 100644 (file)
--- a/liblttng-ust/lttng-ust-comm.c
+++ b/liblttng-ust/lttng-ust-comm.c
@@ -78,6 +78,12 @@ static sem_t constructor_wait;
  */
 static int sem_count = { 2 };
 
+/*
+ * Counting nesting within lttng-ust. Used to ensure that calling fork()
+ * from liblttng-ust does not execute the pre/post fork handlers.
+ */
+static int __thread lttng_ust_nest_count;
+
 /*
  * Info about socket and associated listener thread.
  */
@@ -469,7 +475,9 @@ int get_wait_shm(struct sock_info *sock_info, size_t mmap_size)
         * If the open failed because the file did not exist, try
         * creating it ourself.
         */
+       lttng_ust_nest_count++;
        pid = fork();
+       lttng_ust_nest_count--;
        if (pid > 0) {
                int status;
 
@@ -837,6 +845,7 @@ int get_timeout(struct timespec *constructor_timeout)
 void __attribute__((constructor)) lttng_ust_init(void)
 {
        struct timespec constructor_timeout;
+       sigset_t sig_all_blocked, orig_parent_mask;
        int timeout_mode;
        int ret;
 
@@ -873,16 +882,39 @@ void __attribute__((constructor)) lttng_ust_init(void)
        if (ret) {
                ERR("Error setting up to local apps");
        }
-       ret = pthread_create(&local_apps.ust_listener, NULL,
-                       ust_listener_thread, &local_apps);
 
+       /* A new thread created by pthread_create inherits the signal mask
+        * from the parent. To avoid any signal being received by the
+        * listener thread, we block all signals temporarily in the parent,
+        * while we create the listener thread.
+        */
+       sigfillset(&sig_all_blocked);
+       ret = pthread_sigmask(SIG_SETMASK, &sig_all_blocked, &orig_parent_mask);
+       if (ret) {
+               ERR("pthread_sigmask: %s", strerror(ret));
+       }
+
+       ret = pthread_create(&global_apps.ust_listener, NULL,
+                       ust_listener_thread, &global_apps);
+       if (ret) {
+               ERR("pthread_create global: %s", strerror(ret));
+       }
        if (local_apps.allowed) {
-               ret = pthread_create(&global_apps.ust_listener, NULL,
-                               ust_listener_thread, &global_apps);
+               ret = pthread_create(&local_apps.ust_listener, NULL,
+                               ust_listener_thread, &local_apps);
+               if (ret) {
+                       ERR("pthread_create local: %s", strerror(ret));
+               }
        } else {
                handle_register_done(&local_apps);
        }
 
+       /* Restore original signal mask in parent */
+       ret = pthread_sigmask(SIG_SETMASK, &orig_parent_mask, NULL);
+       if (ret) {
+               ERR("pthread_sigmask: %s", strerror(ret));
+       }
+
        switch (timeout_mode) {
        case 1: /* timeout wait */
                do {
@@ -956,12 +988,14 @@ void __attribute__((destructor)) lttng_ust_exit(void)
        /* cancel threads */
        ret = pthread_cancel(global_apps.ust_listener);
        if (ret) {
-               ERR("Error cancelling global ust listener thread");
+               ERR("Error cancelling global ust listener thread: %s",
+                       strerror(ret));
        }
        if (local_apps.allowed) {
                ret = pthread_cancel(local_apps.ust_listener);
                if (ret) {
-                       ERR("Error cancelling local ust listener thread");
+                       ERR("Error cancelling local ust listener thread: %s",
+                               strerror(ret));
                }
        }
        /*
@@ -993,6 +1027,8 @@ void ust_before_fork(sigset_t *save_sigset)
        sigset_t all_sigs;
        int ret;
 
+       if (lttng_ust_nest_count)
+               return;
        /* Disable signals */
        sigfillset(&all_sigs);
        ret = sigprocmask(SIG_BLOCK, &all_sigs, save_sigset);
@@ -1018,6 +1054,8 @@ static void ust_after_fork_common(sigset_t *restore_sigset)
 
 void ust_after_fork_parent(sigset_t *restore_sigset)
 {
+       if (lttng_ust_nest_count)
+               return;
        DBG("process %d", getpid());
        rcu_bp_after_fork_parent();
        /* Release mutexes and reenable signals */
@@ -1035,6 +1073,8 @@ void ust_after_fork_parent(sigset_t *restore_sigset)
  */
 void ust_after_fork_child(sigset_t *restore_sigset)
 {
+       if (lttng_ust_nest_count)
+               return;
        DBG("process %d", getpid());
        /* Release urcu mutexes */
        rcu_bp_after_fork_child();