https://bugs.winehq.org/show_bug.cgi?id=54807
--- Comment #18 from Zeb Figura z.figura12@gmail.com --- I see the problem now. It *is* related to killing threads while they're holding the mutex.
The reason that adding SIGQUIT to the set isn't enough to fix it, is that the server closes the remote end of the socket, which generates EPIPE, which in turn causes wine_server_call() to abort_thread() with the mutex still held.