Re: [Qemu-devel] [PATCH for-2.0] dsdt: tweak ACPI ID for hotplug resource device

["Michael S. Tsirkin" <mst@redhat.com>]

On Mon, Apr 07, 2014 at 09:46:34AM +0200, Igor Mammedov wrote:
> On Sun, 6 Apr 2014 12:47:37 +0300
> "Michael S. Tsirkin" <mst@redhat.com> wrote:
> 
> > ACPI0004 seems too new:
> > Windows XP complains about an unrecognized device.
> > This is a regression since 1.7.
> > Use PNP0A06 instead - Generic Container Device.
> > 
> > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> > ---
> > 
> > Note: Igor has RFC patches to use PNP0C02 under PCI0,
> > but that's not ready for 2.0.
> > Igor, could you comment on whether PNP0C02 is preferable for
> > some OS-es?
> > 
> > For PNP0A06 ACPI spec says:
> >     PNP0A06
> >     Generic Container Device. This device behaves exactly the same as the
> >     PNP0A05 device.
> >     This was originally known as Extended I/O Bus. This ID should only be
> >     used for containers
> >     that do not produce resources for consumption by child devices. Any
> >     system resources
> >     claimed by a PNP0A06 device’s _CRS object must be consumed by the
> >     container itself.
> > 
> > PNP0C02 is not in ACPI spec. It does appear in MicroSoft legacy PNP ID
> > document:
> >     PNP0C02         General ID for reserving resources required by Plug and
> >     Play motherboard registers. (Not specific to a particular device.)
> > 
> > Thoughts?
> From testing Windows doesn't check for conflicts _CRS provided by PNP0C02,
> so while using it will let XP not to complain about unknown device, it will
> also make later Windows OSes silently ignore conflicts if any.
> 
> When XP's complaining for the first time about unknown device, it could be
> told to ignore it, so it would stop to complain in future. i.e. user have to
> silence only once.
> 
> Wouldn't it be better to use ACPI0004 that gives minor annoyance to XP user
> but provides error checking with later OS versions?

OK I quickly tested with Win7 and it does check for conflicts
when EisaId("PNP0A06") is used.

Taking this into account, ack this patch?

> > 
> >  hw/i386/acpi-dsdt-cpu-hotplug.dsl | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/hw/i386/acpi-dsdt-cpu-hotplug.dsl b/hw/i386/acpi-dsdt-cpu-hotplug.dsl
> > index dee4843..34aab5a 100644
> > --- a/hw/i386/acpi-dsdt-cpu-hotplug.dsl
> > +++ b/hw/i386/acpi-dsdt-cpu-hotplug.dsl
> > @@ -93,7 +93,7 @@ Scope(\_SB) {
> >      }
> >  
> >      Device(CPU_HOTPLUG_RESOURCE_DEVICE) {
> > -        Name(_HID, "ACPI0004")
> > +        Name(_HID, EisaId("PNP0A06"))
> >  
> >          Name(_CRS, ResourceTemplate() {
> >              IO(Decode16, CPU_STATUS_BASE, CPU_STATUS_BASE, 0, CPU_STATUS_LEN)