debug.acpi.avoid, how do i use it to tell acpi to not control the keyboard?

[Nate Lawson]

john at utzweb.net wrote:
>> John L.Utz III wrote:
>>> At Mon, 12 Jun 2006 10:17:11 -0700,
>>> Nate Lawson wrote:
>>> wouldnt this be the one to avoid, assuming that 'KBC' is
>>> KeyboardController?
>>>
>>>                 Device (KBC)
>>>                 {
>>>                     Name (_HID, EisaId ("PNP0303"))
>>>                     Name (CRS, ResourceTemplate ()
>>>                     {
>>>                         IO (Decode16, 0x0060, 0x0060, 0x10, 0x01)
>>>                         IO (Decode16, 0x0064, 0x0064, 0x04, 0x01)
>>>                         IRQNoFlags () {1}
>>>                     })
>>>                     Method (_CRS, 0, NotSerialized)
>>>                     {
>>>                         Return (CRS)
>>>                     }
>>>                 }
>>>
>>>
>>>
>>> can you clarify your comment a trifle?
>> It would take a long time to explain it fully.  ACPI is not controlling
>> your keyboard.  The BIOS just decides whether or not to send ACPI hotkey
>> events (different from keystrokes) or consume them itself via SMI.
> 
> so, the act of loading acpi.ko implements a callback for the bios that
> informs it that an acpi cognizant os is available?
> 
> it seems to me that i would like to tell it to not use acpi and stick with
> smi because if i dont load acpi the buttons work almost perfectly. the
> only bit of grief i have is that without acpi, the powerbutton shuts off
> hard instead of performing a shutdown

Yep, that's your choice.

>> For some reason, the notify isn't getting delivered to ACPI along
>> the way.
> 
> some do, some dont, suspend get's responded to, power button get's
> responded to.
> 
> what commands do i use to break on buttons in acpidb?

Just stick a printf in acpi_button.c?

-- 
Nate