Fix: lttng-ust-comm: wait on wrong child process

[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 7709b8dba6df5ed6cc4d5aff35c13988ff3b2114..716795e0e818028d1a6fc6fa63374abaaf0ebafa 100644 (file)
--- a/liblttng-ust/lttng-ust-comm.c
+++ b/liblttng-ust/lttng-ust-comm.c
@@ -1293,14 +1293,14 @@ open_write:
        pid = fork();
        URCU_TLS(lttng_ust_nest_count)--;
        if (pid > 0) {
-               int status;
+               int status, wait_ret;
 
                /*
                 * Parent: wait for child to return, in which case the
                 * shared memory map will have been created.
                 */
-               pid = wait(&status);
-               if (pid < 0 || !WIFEXITED(status) || WEXITSTATUS(status) != 0) {
+               wait_ret = waitpid(pid, &status, 0);
+               if (wait_ret < 0 || !WIFEXITED(status) || WEXITSTATUS(status) != 0) {
                        wait_shm_fd = -1;
                        goto end;
                }
@@ -1468,18 +1468,25 @@ void wait_for_sessiond(struct sock_info *sock_info)
 
        DBG("Waiting for %s apps sessiond", sock_info->name);
        /* Wait for futex wakeup */
-       if (uatomic_read((int32_t *) sock_info->wait_shm_mmap))
-               goto end_wait;
-
-       while (futex_async((int32_t *) sock_info->wait_shm_mmap,
-                       FUTEX_WAIT, 0, NULL, NULL, 0)) {
+       while (!uatomic_read((int32_t *) sock_info->wait_shm_mmap)) {
+               if (!futex_async((int32_t *) sock_info->wait_shm_mmap, FUTEX_WAIT, 0, NULL, NULL, 0)) {
+                       /*
+                        * Prior queued wakeups queued by unrelated code
+                        * using the same address can cause futex wait to
+                        * return 0 even through the futex value is still
+                        * 0 (spurious wakeups). Check the value again
+                        * in user-space to validate whether it really
+                        * differs from 0.
+                        */
+                       continue;
+               }
                switch (errno) {
-               case EWOULDBLOCK:
+               case EAGAIN:
                        /* Value already changed. */
                        goto end_wait;
                case EINTR:
                        /* Retry if interrupted by signal. */
-                       break;  /* Get out of switch. */
+                       break;  /* Get out of switch. Check again. */
                case EFAULT:
                        wait_poll_fallback = 1;
                        DBG(