https://bugs.winehq.org/show_bug.cgi?id=54807
--- Comment #20 from Zeb Figura z.figura12@gmail.com --- (In reply to Roman from comment #19)
Great that you find a reason, but it fixable?
I sent https://gitlab.winehq.org/wine/wine/-/merge_requests/5118. The problem with that merge request is that, while it's probably reliably good enough for this application, it isn't really enough to solve the broader problem of kernel-side code getting into an inconsistent state when threads are terminated. It needs more thought.