Question on fscrypt_d_revalidate() and fstest generic/429

[Al Viro]

On Mon, May 15, 2017 at 09:51:03PM +0200, Richard Weinberger wrote:

> > I doubt it's the right solution to make fscrypt_d_revalidate() look at
> > ->i_nlink, since ->d_revalidate() is meant to validate the filename, not the
> > inode.  I think there is probably a VFS bug that is causing the dentries to not
> > be freed.
> 
> Not sure. Al? :-)

What's to tell VFS that they are garbage?  If your code doesn't do it,
VFS has no way to know.  And I certainly agree that using ->d_revalidate()
to trigger their unhashing is a bloody odd design; take it up with whoever
had come up with it...