Re: [PATCH] ACPI: don't walk tables if ACPI was disabled

[Len Brown <lenb@kernel.org>]

On Tue, 24 Jun 2008, Ingo Molnar wrote:

> * Vegard Nossum <vegard.nossum@gmail.com> wrote:
> 
> > On Fri, Jun 20, 2008 at 11:27 PM, Vegard Nossum <vegard.nossum@gmail.com> wrote:
> > > So I guess this function, pnpbios_init() needs the check as well. In
> > > fact, it has this:
> > >
> > > #ifdef CONFIG_PNPACPI
> > >        if (!acpi_disabled && !pnpacpi_disabled) {
> > >                pnpbios_disabled = 1;
> > >                printk(KERN_INFO "PnPBIOS: Disabled by ACPI PNP\n");
> > >                return -ENODEV;
> > >        }
> > > #endif                          /* CONFIG_ACPI */
> > >
> > > ...I guess that should be changed to say if (acpi_disabled ||
> > > pnpacpi_disabled)? Or... I don't understand the purpose of the
> > > original test. But it seems to be there since the beginning of time
> > > (or, well, v2.6.12-rc2).
> > 
> > Nope. I found the introduction of the change in the historical git repository:
> > 
> > commit 4723ebe898a32262ed49fe677897ccea47e72ff4
> > Author: Adam Belay <ambx1@neo.rr.com>
> > Date:   Sun Oct 24 15:07:32 2004 -0400
> > 
> >     [PNPBIOS] disable if ACPI is active
> > 
> >     As further ACPI pnp functionaility is implemented it is no longer
> >     safe to run ACPI and PNPBIOS concurrently.
> > 
> >     We therefore take the following approach:
> >     - attempt to enable ACPI support
> >     - if ACPI fails (blacklist etc.) enable pnpbios support
> >     - if ACPI support is not compiled in the kernel enable pnpbios support
> > 
> >     Signed-off-by: Adam Belay <ambx1@neo.rr.com>
> > 
> > and now I understand the purpose of the check; pnpbios does not depend 
> > on ACPI; ACPI/pnpacpi is incompatible with pnpbios.
> 
> wow, rather old bug - i guess lockdep made it more visible.

No, that commit was not a bug, it was correct, and still is,
for pnpACPI and pnpBIOS must be mutually exclusive.

The thing that changed was the RTC specific code.

-Len