On Tue, Apr 08, 2025 at 09:41:33PM -0400, Maxim Levitsky wrote: > Allow to specify the lockdep subclass in mutex_trylock > instead of hardcoding it to 0. We disable a whole bunch of checks for trylock, simply because they do not wait, therefore they cannot deadlock. But I can't remember if they disable all the cases required to make subclasses completely redundant -- memory suggests they do, but I've not verified. Please expand this Changelog to include definite proof that subclasses make sense with trylock.
