On Tue, Jul 13, 2021 at 11:58 PM Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx> wrote: > > On Tue, Jul 13, 2021 at 7:53 AM Christian Brauner > <christian.brauner@xxxxxxxxxx> wrote: > > > > Instead of naming all these $something_helper I would follow the > > underscore naming pattern we usually do, i.e. instead of e.g. > > rmdir_helper do __rmdir() or __do_rmdir(). > > That's certainly a pattern we have, but I don't necessarily love it. > > It would be even better if we'd have names that actually explain > what/why the abstraction exists. In this case, it's the "possibly > retry due to ESTALE", but I have no idea how to sanely name that. > Making it "try_rmdir()" or something like that is the best I can come > up with right now. > > On a similar note, the existing "do_rmdir()" and friends aren't > wonderful names either, but we expose that name out so changing it is > probably not worth it. But right now we have "vfs_rmdir()" and > "do_rmdir()", and they are just different levels of the "rmdir stack", > without the name really describing where in the stack they are. > > Naming is hard, and I don't think the double underscores have been > wonderful either. Naming *is* hard, I do not have any good ideas here, so I just went with try_rmdir(). Christian, Linus, let me know if that is not good enough. -- Dmitry Kadashev
