On Sat, 25 Nov 2023 at 21:08, Al Viro <viro@xxxxxxxxxxxxxxxxxx> wrote:
>
> On Sat, Nov 25, 2023 at 08:59:54PM -0800, Linus Torvalds wrote:
> >
> > because I for some reason (probably looking
> > at Mateusz' original patch too much) re-implemented file_free() as
> > fput_immediate()..
>
> file_free() was with RCU delay at that time, IIRC.
Ahh, indeed. So it was the SLAB_TYPESAFE_BY_RCU changes that basically
made my fput_immediate() and file_free() be the same, and thus it all
simplifies to your nice version.
> I don't think that
> cost of one test and (rarely) branch on each final fput() is going to
> be measurable.
Nope. Me likey.
Linus
