page fault panic in device_get_softc/acpi_pcib_route_interrupt

[Nate Lawson]

John Baldwin wrote:
> 
> On Jan 8, 2005, at 1:47 PM, Nate Lawson wrote:
> 
>> Pawel Worach wrote:
>>
>>> Nate Lawson wrote:
>>> --- sys/dev/acpica/acpi_pcib.c    27 Dec 2004 05:36:47 -0000    1.53
>>> +++ sys/dev/acpica/acpi_pcib.c    8 Jan 2005 09:05:57 -0000
>>> @@ -249,11 +249,18 @@
>>>      /*
>>>       * We have to find the source device (PCI interrupt link device).
>>>       */
>>> -    if (ACPI_FAILURE(AcpiGetHandle(ACPI_ROOT_OBJECT, prt->Source, 
>>> &lnkdev))) {
>>> +    if (ACPI_FAILURE(AcpiGetHandle(acpi_get_handle(pcib), prt->Source,
>>> +    &lnkdev))) {
>>>      device_printf(pcib, "couldn't find PCI interrupt link device %s\n",
>>>          prt->Source);
>>>      goto out;
>>>      }
>>
>>
>> This change should not be committed.  I think it's correct to use \ to 
>> start the search since your _PRT contains relative references but the 
>> links are under \.
> 
> 
> So, should I change the code to force-attach the devices to also use 
> ACPI_ROOT_OBJECT for its lookup as well in prt_attach_devices()?
> 

Sure, I think that's a good idea.

-- 
Nate