ACPICA 20100121 regression
robert.moore at intel.com
Mon Feb 1 20:24:14 UTC 2010
The worst part of all is that if ACPICA returns TRUE for "Linux", the ASL code executes down paths that often have never been tested.
The goal of ACPICA is to be 100% compatible with the Windows ACPI implementation. As such, it returns TRUE for all Windows query strings.
Note, _OSI was never intended to be a test for "which operating system is executing". It is meant to query the "set of ACPI-related interfaces, behaviors, or features that the operating system supports" (from ACPI specification.) Thus, it is entirely appropriate for ACPICA to return TRUE for windows strings.
I guess the next question would be: why is the machine disabling things specifically for Windows 7?
>From: owner-freebsd-acpi at freebsd.org [mailto:owner-freebsd-
>acpi at freebsd.org] On Behalf Of Jung-uk Kim
>Sent: Monday, February 01, 2010 11:51 AM
>To: Rui Paulo
>Cc: freebsd-acpi at freebsd.org
>Subject: Re: ACPICA 20100121 regression
>On Monday 01 February 2010 02:36 pm, Rui Paulo wrote:
>> On 1 Feb 2010, at 19:33, Jung-uk Kim wrote:
>> > On Monday 01 February 2010 02:25 pm, Rui Paulo wrote:
>> >> On 1 Feb 2010, at 19:21, Jung-uk Kim wrote:
>> >>> On Saturday 30 January 2010 10:49 am, Rui Paulo wrote:
>> >>>> Hi,
>> >>>> Latest ACPICA can't find my ASUS010 HID. It worked fine with
>> >>>> FreeBSD 8, which has ACPICA 20090521.
>> >>>> The ASL is located at:
>> >>>> http://people.freebsd.org/~rpaulo/asus-1005ha.asl.gz
>> >>>> What I'm seeing is ACPI_ID_PROBE() returning always NULL for
>> >>>> "ASUS010" and "ATK0100" devids.
>> >>> It seems the ASL disables ASUS010 when the OS is "Windows 2009"
>> >>> (aka Windows 7). FYI, current ACPI-CA just returns okay for
>> >>> any Microsoft OSes when _OSI method is used in ASL. Thus, it
>> >>> thinks you are running Windows 7. You can comment out or
>> >>> remove line 3626-3629 and override DSDT to re-enable the
>> >>> device, I think.
>> >> You're right, but I'm left wondering why it worked with a
>> >> previous ACPICA.
>> > Because "Windows 2009" was added in 20090903. :-)
>> I understand now. Still, I think this is ACPICA's fault, but I
>> understand that other laptops may rely on this behavior from
>> ACPICA, so the fix may cause even more problems..
>I agree that it is ACPI-CA's fault but it was debated in Linux
>community for a while and they decided it is the best course of
>action for ACPI-CA, AFAIK. Basically, a lot of ACPI implementations
>out there just disable some "features" based on Windows versions.
>Even worse, many features are disabled when it matches "Linux". So,
>they decided returning the latest and greatest Windows version
>instead is the best choice. Luckily (or unluckily), not so many ACPI
>implementations match "FreeBSD". :-(
>freebsd-acpi at freebsd.org mailing list
>To unsubscribe, send any mail to "freebsd-acpi-unsubscribe at freebsd.org"
More information about the freebsd-acpi