Re: [PATCH 2/2] statx: move STATX_ATTR_DAX attribute handling to filesystems

[Christoph Hellwig]

On Tue, Dec 01, 2020 at 02:12:19PM -0800, Linus Torvalds wrote:
> On Tue, Dec 1, 2020 at 2:03 PM Eric Sandeen <sandeen@xxxxxxxxxxx> wrote:
> >
> > That's why I was keen to just add DAX unconditionally at this point, and if we want
> > to invent/refine meanings for the mask, we can still try to do that?
> 
> Oh Gods.  Let's *not* make this some "random filesystem choice" where
> now semantics depends on what some filesystem decided to do, and
> different filesystems have very subtly different semantics.
> 
> This all screams "please keep this in the VFS layer" so that we at
> least have _one_ place where these kinds of decisions are made.
> 
> I suspect very very few people actually end up caring about any of the
> STATX flags at all, of course. The fact that the DAX one was
> apparently entirely the wrong bit argues that this isn't all that
> important.

Agreed.  That whole support interface is just weird.  But the only
thing that remotely makes (a little bit of) sense is to just set all
bits known about by this particular kernel in the VFS.  Everything else
is going to be a complete mess.