diff --git a/lib/pls/src/internal/scheduling/task.cpp b/lib/pls/src/internal/scheduling/task.cpp
index 7c06bc7..5a4b1f4 100644
--- a/lib/pls/src/internal/scheduling/task.cpp
+++ b/lib/pls/src/internal/scheduling/task.cpp
@@ -19,17 +19,15 @@ task::task(const task &other) :
     deque_state_{other.deque_state_} {}
 
 void task::execute() {
-  {
-    PROFILE_WORK_BLOCK("execute task")
-    auto last_executing = thread_state::get()->current_task_;
-    thread_state::get()->current_task_ = this;
+  PROFILE_WORK_BLOCK("execute task")
+  auto last_executing = thread_state::get()->current_task_;
+  thread_state::get()->current_task_ = this;
 
-    execute_internal();
-
-    thread_state::get()->current_task_ = last_executing;
-  }
+  execute_internal();
+  PROFILE_END_BLOCK
 
   wait_for_all();
+  thread_state::get()->current_task_ = last_executing;
 
   if (parent_ != nullptr) {
     parent_->ref_count_--;
diff --git a/test/scheduling_tests.cpp b/test/scheduling_tests.cpp
index d99be04..710fc80 100644
--- a/test/scheduling_tests.cpp
+++ b/test/scheduling_tests.cpp
@@ -70,6 +70,9 @@ TEST_CASE("tbb task are scheduled correctly", "[internal/scheduling/fork_join_ta
       std::atomic<int> dummy_parent{1}, overall_counter{8};
       force_steal_sub_task sub_task{&dummy_parent, &overall_counter};
       scheduler::spawn_child(sub_task);
+
+      // Required, as child operates on our stack's memory!!!
+      scheduler::wait_for_all();
     });
     my_scheduler.terminate(true);
   }