Linus Torvalds writes: (Summary) Was there some known problem during the
4.13 merge window that could have caused this?
4.13 merge window that could have caused this?
The fact that it came in with the locking merge might also just mean that it's a bug in the locking balance detection, presumably interacting badly with something else.
interacting badly with something else.
There *is* another email (from Sebastian Andrzej Siewior) that shows that same "WARNING: bad unlock balance detected", with the subject line "RCU boosting and lockdep are not playing nice", and that one blames commit cde50a67397c ("locking/rtmutex: Don't initialize lockdep when not required").
when not required").
Maybe this was already fixed elsewhere, but I'm not seeing any replies to that other email either (it was send back in mid-August).
4.13 merge window that could have caused this?
The fact that it came in with the locking merge might also just mean that it's a bug in the locking balance detection, presumably interacting badly with something else.
interacting badly with something else.
There *is* another email (from Sebastian Andrzej Siewior) that shows that same "WARNING: bad unlock balance detected", with the subject line "RCU boosting and lockdep are not playing nice", and that one blames commit cde50a67397c ("locking/rtmutex: Don't initialize lockdep when not required").
when not required").
Maybe this was already fixed elsewhere, but I'm not seeing any replies to that other email either (it was send back in mid-August).