exit gotcha + remove DelayedInit state + rocm-smi + cleanup (#110)

* exit gotcha + remove DelayedInit state + cleanup

- exit gotcha which wraps exit, quick_exit, abort
- minor refactor of mpi gotchas
- removed some redundant code in omnitrace_finalized_hidden
- exclude instrumenting functions starting with dlopen and dlsym
- exclude instrumenting exit, quick_exit, and abort functions
- update timemory submodule with support for new gotcha_invoker with (gotcha_data, <function pointer>, args...)

* Improved rocm_smi error handling

source/bin/omnitrace/details.cpp

@@ -46,7 +46,7 @@ get_whole_function_names()
         "sigandset", "sigdelset", "sigemptyset", "sigfillset", "sighold", "sigisemptyset",
         "sigismember", "sigorset", "sigrelse", "sigvec", "strtok", "strstr", "sbrk",
         "strxfrm", "atexit", "ompt_start_tool", "nanosleep", "cfree", "tolower",
-        "toupper", "fileno", "fileno_unlocked",
+        "toupper", "fileno", "fileno_unlocked", "exit", "quick_exit", "abort",
         // below are functions which never terminate
         "rocr::core::Signal::WaitAny", "rocr::core::Runtime::AsyncEventsLoop",
         "rocr::core::BusyWaitSignal::WaitAcquire",

source/bin/omnitrace/module_function.cpp

@@ -435,8 +435,8 @@ module_function::is_routine_constrained() const
         "std::(basic_|)(ifstream|ios|istream|ostream|stream))",
         regex_opts);
     static std::regex leading(
-        "^(_|\\.|frame_dummy|transaction clone|virtual "
-        "thunk|non-virtual thunk|\\(|targ|kmp_threadprivate_|Kokkos::Profiling::)",
+        "^(_|\\.|frame_dummy|transaction clone|virtual thunk|non-virtual thunk|"
+        "\\(|targ|kmp_threadprivate_|Kokkos::Profiling::|dlopen|dlsym)",
         regex_opts);
     static std::regex trailing(
         "(_|\\.part\\.[0-9]+|\\.constprop\\.[0-9]+|\\.|\\.[0-9]+)$", regex_opts);