Closed sochsenk closed 2 years ago
I'm still analyzing, but my guess is there are boost::atomics somewhere below and Helgrind doesn't understand atomics.
I think it is a false positive due to Helgrinds inability to handle atomics. threadsanitizer couldn't find anything.
I'm using boost 1.70.0
I'm not sure if this is a known issue or a false positive. I got helgrind issues (possible data races) reported when running the following code snippet:
Note that I'm using std::thread but I got similar issues with boost::thread Besides other issues, it seems waiting for the future (r.get()) and the barrier waits, both seem to access boost::detail::get_current_thread_data(). Part of the helgrind report: