Misc updates (#48)

- reworked `add_critical_trace`
- `get_use_thread_sampling` / `"OMNITRACE_USE_THREAD_SAMPLING"` option
- `get_cpu_cid_stack_lock`
- reworked finalization messaging
- significant updates to pthread_gotcha
  - shutdown stability
  - `"start_thread"` entries
- `rocm_smi` stability 
- roctracer_callbacks add critical trace entries on the callback thread
- reworked CPU CID initialization
- thread_sampler stability

[ROCm/rocprofiler-systems commit: 9b25d4b3b5]

projects/rocprofiler-systems/source/lib/omnitrace/include/library/runtime.hpp

@@ -86,4 +86,7 @@ create_cpu_cid_entry(int64_t _tid = threading::get_id());
 cpu_cid_pair_t
 get_cpu_cid_entry(uint64_t _cid, int64_t _tid = threading::get_id());
 
+tim::mutex_t&
+get_cpu_cid_stack_lock(int64_t _tid = threading::get_id());
+
 }  // namespace omnitrace