Re: kernel structures in 2.0.29->2.0.30

Raul Miller (rdm@test.legislate.com)
Fri, 25 Apr 1997 02:39:51 -0400


> Raul Miller <rdm@test.legislate.com> writes:
> > by the people who are interested in afs. But it's really silly
> > for people to try to enforce a binary interface where a source
> > interface was designed.

On Apr 25, Stefan Monnier wrote
> I doubt that they'd have bother with versioning info if it really
> was supposed to be only source-level compatible. And it is true
> that the fundamental idea behind a "stable release" is that upgrades
> should only fix things and not bring any changes which might break
> other programs/services.

You're right, I phrased that poorly.

A better way of describing it is that the kernel module system
is a view into the guts of Linux. As Linux evolves, the view
is subject to change. This should not surprise anyone.

And, in my opinion, it's not right to stick a lead weight into
Linux's guts to try and prevent Linux from being improved. Nor
is it necessary, as you can write filters that obscure the
view enough to make it appear as if nothing had been fixed.

But the point is that, from the view of a subsystem such as AFS,
kernel module interface can and will drift. A simple shim between
the binary module and the raw kernel ought to fill in the gaps.

Once again: there's no reason that source code can't be written
to fit between the AFS module and the various linux kernel
versions. This would be an easier task if AFS had been engineered
for this approach, but it's doable even now.

-- 
Raul