--CE+1k2dSO48ffgeK
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Sat, Mar 05, 2005 at 11:47:44AM +0100, Julio M. Merino Vidal wrote:
> On Sat, 2005-03-05 at 11:32 +0100, Vincent wrote:
> > Hi,
> >=20
> > the question in the subject can look silly, but, since the 2.99.13=20
> > kernel, I can't use my laptop under self-power, because it reboots=20
> > continuously without warning after 1 mn or so of use, and I haven't bee=
n=20
> > able as yet to pinpoint the bug. Conclusion : while I am using the=20
> > current kernel and userland when the laptop is connected to the mains, =
I=20
> > have to reverse to an obsolete 2.99.12 when it is not.
> >=20
> > However, I suspect sooner or later one program in the current-userland=
=20
> > will make use of a call not implemented on 2.99.12 and the system won't=
=20
> > be usable anymore. Is that right?
>=20
> When that happens, the program in question will either produce incorrect
> output or crash.  But the system will keep up and running.

As noted, /sbin/init can cause issues. Another problem may be if libc is=20
updated to use a new call that isn't there; ALL programs can crash. I=20
remember when we went to having getcwd() based on a system call. If you=20
used a new userland (that expected the system call) with an old kernel=20
(that didn't have it), most every program crashed. I don't know if init=20
crashed then, but I think every shell did. :-)

To answer the original question, it depends. As userland's expectation of=
=20
the kernel changes, different things will fail. We can't predict how fast=
=20
such changes will happen, but as they do, areas of functionality will=20
break.

Take care,

Bill

--CE+1k2dSO48ffgeK
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (NetBSD)

iD8DBQFCLIrdWz+3JHUci9cRAvnmAJ92awH4zcEup7CQGmJnVXR+FSpkxwCeKLLc
CzjHIp5lC9FMlo+7PLwNuSI=
=dyPz
-----END PGP SIGNATURE-----

--CE+1k2dSO48ffgeK--