Re: one more ACPI Error (utglobal-0125): Unknown exception code: 0xFFFFFFEA [Re: 2.6.18-rc4-mm3]

["Bjorn Helgaas" <bjorn.helgaas@hp.com>]

> On Thu, 2006-08-31 at 17:06 -0600, Bjorn Helgaas wrote:
>> Problem 1: acpi_reserve_io_ranges() needs to return an acpi_status
>> like AE_OK or AE_CTRL_TERMINATE, not a -EINVAL.
>
> Sure great sounds.  I understand AE_OK is a 0 return so I can change it
> to AE_CTRL_TERMINATE.  I don't want  acpi_reserve_io_ranges to return a
> happy state when if finds a resource type is doesn't know.

Except that when the motherboard driver claims a device, it really
should claim all the resources used by the device.  It currently only
claims I/O port resources, but I think it should also claim MMIO
resources.  Otherwise, the system resource accounting is screwed up,
and resources consumed by the motherboard device could be mistakenly
allocated to another device.

> Kame (who helped me greatly in tracking down the source my troubles)
> thinks that the root cause is that the device (my memory_device) has
> both a _HID and _CID. The driver for _HID is different for _CID and the
> driver for _CID is found before _HID and I pass the wrong device up the
> chain.

Ok, this is starting to make sense.  It sounds like your memory
device has _HID of PNP0C80 and _CID of PNP0C01 (or PNP0C02).

The current ACPI driver binding algorithm in acpi_bus_find_driver()
looks at each driver, checking whether it can match either the _HID
or the _CID of a device.  Since we try the motherboard driver first,
it matches the memory device _CID.

I couldn't find a specific reference in the spec, but this seems
intuitively sub-optimal.  It seems like it'd be better to look
first for a driver that can claim the _HID (which is more specific),
and only fall back to checking the _CIDs if no _HID-specific driver
is found.

This looks fairly easy to do in ACPI.  Not so easy in PNPACPI,
since I don't think PNP has the concept of _HID vs _CID.  Maybe
Len will chime in with an opinion.

Bjorn