On Tue, 30 Jan 2024 at 06:39, Steven Rostedt [off-list ref] wrote:

On Tue, 30 Jan 2024 01:12:05 -0800

quoted

I suspect the solution is to make eventfs_create_dir() do the same as
the events directory case does, and actually pin the directory dentry
and save it off.

I rather not have the create do that because that happens for every event
directory.

I wasn't thinking straight yesterday - the real fix is to just do a
"d_revalidate()". It's literally why that thing exists: check whether
a dentry is still valid.

In fact, that trivially gets rid of the 'events' subdirectory issues
too, so we can stop doing that horrendous simple_recursive_removal()
too.

Let me reboot into the trivial fix. I just needed to think about this
the right way.

None of this is anything that the VFS layer has any problems with.

               Linus