device atpic to be deprecated?

[Conrad J. Sabatier]

On Mon, 3 Apr 2006 11:52:25 -0400
John Baldwin <jhb at freebsd.org> wrote:
> On Sunday 02 April 2006 07:07, Conrad J. Sabatier wrote:
> > On Wed, 29 Mar 2006 13:15:54 -0500
> > John Baldwin <jhb at freebsd.org> wrote:
> > > 
> > > I have to make sure it really works for everyone first though
> > > before removing it would really be viable. :-/
> > 
> > So, would it be necessary to upgrade to HEAD in order to make sure
> > that this problem won't still occur on my box?  Or has this stuff
> > already been merged to STABLE?
> 
> This particular bunch of code is identical in HEAD and 6.x right
> now, so we can probably debug it on STABLE just fine.

Well, some testing with my latest STABLE source update (Fri, Apr 7)
exhibits the same problems.

I can boot the kernel and mount root OK with or without ACPI enabled
(via the loader tunable hint.acpi.0.disabled), but not without atpic.

I can even boot OK without "device acpi" compiled into the kernel
explicitly (and with no acpi module, either, of course, as this
doesn't get built under amd64), but without atpic and the associated
ACPI bits it pulls in, it's a definite show-stopper.

Another interesting side effect I've noticed is that without acpi, the
built-in nVidia sound chip (with the snd_ich driver compiled in) is not
recognized either.

-- 
Conrad J. Sabatier <conrads at cox.net> -- "In Unix veritas"