Re: [PATCH 2/2] vfs: support statx(..., NULL, AT_EMPTY_PATH, ...)

[Linus Torvalds <torvalds@linux-foundation.org>]

On Wed, 3 Jul 2024 at 09:54, Xi Ruoyao <xry111@xry111.site> wrote:
>
> > Honestly, 'statx' is disgusting. I don't understand why anybody pushes
> > that thing that nobody actually uses or cares about.
>
> Hmm why it was added in the first place then?  Why not just NAK it?

There are valid uses of statx - they are just VERY very few and far between.

For example, if you want the "change cookie" or any number of other
special things that aren't standard, you have to use statx.

But _normal_ programs will never do that. It's unportable, and it
really is a specialty interface.

Pushing 'statx' as a replacement for 'stat' is just crazy. It's a
different thing. It's not a "better" thing. It's an extension, yes,
but "extension" doesn't mean "better".

That's true when it was mis-designed in certain ways that we now have
to fix up, but PARTICULARLY true when it's nonstandard and no other OS
has it.

> And should we add stat_time64, fstat_time64, and fstatat_time64 to stop
> using statx on 32-bit platforms too as it's disgusting?

We already have 'stat64' for 32-bit platforms. We have had it for over
25 years - it predates not only the kernel git tree, it predates the
BK tree too.

I think stat64 was introduced in 2.3.34. That is literally last century.

Anybody who tries to make this about 2037 is being actively dishonest.

Why are people even discussing this pointless thing?

            Linus