GOTCHA + Kokkos + tasking + more (#47)

* GOTCHA + Kokkos + tasking + more

- update gotcha with fix for dlsym(RTLD_NEXT, ...)
- support for standalone KOKKOS_PROFILE_LIBRARY
- remove extra flags for omnitrace-user
- roctracer and critical_trace namespaces in tasking
- generic tasking functions, e.g. join(), shutdown(), etc.
- omnitrace_init_tooling_hidden in api.hpp
- ompt.cpp uses OMNITRACE_USE_OMPT
- kokkosp uses user_region instead of omnitrace component
- re-enable recycling thread ids
- more generic _{push,pop}_perfetto functors
- fix for thread_data::instance(construct_on_init, ...)
- fix for omnitrace-headers interface target
- omnitrace_watch_for_change

source/lib/omnitrace/include/library/api.hpp

@@ -70,6 +70,7 @@ extern "C"
 
     // these are the real implementations for internal calling convention
     void omnitrace_init_library_hidden(void) OMNITRACE_HIDDEN_API;
+    bool omnitrace_init_tooling_hidden(void) OMNITRACE_HIDDEN_API;
     void omnitrace_init_hidden(const char*, bool, const char*) OMNITRACE_HIDDEN_API;
     void omnitrace_finalize_hidden(void) OMNITRACE_HIDDEN_API;
     void omnitrace_set_env_hidden(const char* env_name,

source/lib/omnitrace/include/library/components/fwd.hpp

@@ -52,6 +52,11 @@ TIMEMORY_DECLARE_TYPE_TRAIT(name, typename Tp)
     {                                                                                    \
         static constexpr auto value = NAME;                                              \
     };                                                                                   \
+    template <>                                                                          \
+    struct name<type_list<__VA_ARGS__>>                                                  \
+    {                                                                                    \
+        static constexpr auto value = NAME;                                              \
+    };                                                                                   \
     }                                                                                    \
     }
 

source/lib/omnitrace/include/library/ptl.hpp

@@ -32,22 +32,55 @@ namespace omnitrace
 {
 namespace tasking
 {
+void
+setup();
+
+void
+join();
+
+void
+shutdown();
+
+//--------------------------------------------------------------------------------------//
+//
+//      roctracer
+//
+//--------------------------------------------------------------------------------------//
+
+namespace roctracer
+{
 std::mutex&
-get_roctracer_mutex();
+get_mutex();
 
 PTL::ThreadPool&
-get_roctracer_thread_pool();
+get_thread_pool();
 
 PTL::TaskGroup<void>&
-get_roctracer_task_group();
+get_task_group();
 
+bool
+get_thread_pool_is_active();
+}  // namespace roctracer
+
+//--------------------------------------------------------------------------------------//
+//
+//      critical trace
+//
+//--------------------------------------------------------------------------------------//
+
+namespace critical_trace
+{
 std::mutex&
-get_critical_trace_mutex();
+get_mutex();
 
 PTL::ThreadPool&
-get_critical_trace_thread_pool();
+get_thread_pool();
 
 PTL::TaskGroup<void>&
-get_critical_trace_task_group();
+get_task_group();
+
+bool
+get_thread_pool_is_active();
+}  // namespace critical_trace
 }  // namespace tasking
 }  // namespace omnitrace

source/lib/omnitrace/include/library/thread_data.hpp

@@ -134,8 +134,8 @@ template <typename... Args>
 typename thread_data<Tp, Tag, MaxThreads>::instance_array_t&
 thread_data<Tp, Tag, MaxThreads>::instances(construct_on_init, Args&&... _args)
 {
-    static auto _v = [&]() {
-        auto _internal = instance_array_t{};
+    static auto& _v = [&]() -> instance_array_t& {
+        auto& _internal = instances();
         for(size_t i = 0; i < MaxThreads; ++i)
             _internal.at(i) = unique_ptr_t<Tp>{ new Tp(std::forward<Args>(_args)...) };
         return _internal;