Re: [GIT pull] timer updates for 4.19

From: Thomas Gleixner
Date: Sun Aug 26 2018 - 16:18:15 EST

On Sun, 26 Aug 2018, Linus Torvalds wrote:

> On Sun, Aug 26, 2018 at 2:39 AM Thomas Gleixner <tglx@xxxxxxxxxxxxx> wrote:
> >
> > New defines for the compat time* types so they can be shared between 32bit
> > and 64bit builds. Not used yet, but merging them now allows the actual
> > conversions to be merged through different maintainer trees without
> > dependencies
> So I pulled this, then looked at it a bit more, and went "that's stupid".
> Why introduce a completely useless new name for something we already have?
> Why can't the 32-bit code just use the "compat" names instead? Even
> for 32-bit systems, it's about compatibility with the old world order.
> So what's the advantage of calling it "old" over just calling it "compat"?
> I do not for a *second* believe that "compat" is somehow confusing to
> 32-bit architectures. They all have the new 64-bit time code already,
> it's not like there is any confusion what-so-ever about what is going
> on here.
> The fact is, those 32-bit interfaces are FOR COMPATIBILITY with old
> binaries. What the hell is confusing about that? No, the real
> confusion would be to have yet another name for the same thing for no
> good reason.

We still have compat interfaces for 32bit on 64bit even with the new 2038
safe timespec/val variants because pointer size is different. And for the
old style timespec/val interfaces we need yet another 'compat' interface
for both 32bit native and 32bit on 64bit. That's why we need a different