Hi Pavel,
æ åï2010-09-25 æ 14:44 +0200ïPavel Troller æåï
> Hi!
> I'm a fresh owner of GX723 notebook by MSI. It looks relatively
> linux-friendly, but I've found some ACPI-related issues.
> 1) Booting 2.6.35.5 kernel natively, the Fn keys (brightness control,
> video switching) are non-functional and even a single keypress causes,
> that ACPI events (like AC adapter status changes or lid switch operation) are
> not generated anymore. After a while (about 10 minutes), the following
> lines pop to the system log:
> Sep 25 13:53:18 beam kernel: ACPI Error (psparse-0537): Method parse/execution failed [\_SB_.PCI0.SBRG.EC__._QB7] (Node ffff88013fc1ff00), AE_AML_INFINITE_LOOP
> Sep 25 13:42:23 beam kernel: ACPI Error (psparse-0537): Method parse/execution failed [\_SB_.PCI0.SBRG.EC__._QB8] (Node ffff88013fc1ff20), AE_AML_INFINITE_LOOP
In _QB7 and _QB8 have the following infinite while loop:
If (_OSI ("Windows 2006"))
{
If (LEqual (^^^P0P2.VGA.VVID, 0x10DE))
{
Store (One, ^^^P0P2.VGA.MSIB)
Store (Zero, ^^^P0P2.VGA.MSIM)
Store (One, ^^^P0P2.VGA.MSIE) /* defalut
is one */
Sleep (0x0A)
Notify (^^^P0P2.VGA, 0xCB)
While (^^^P0P2.VGA.MSIE) /* infinite */
{
Sleep (0x0A)
}
}
}
The MSIE is just a named object declared in VGA device but it's not any
hardware register. So, kernel's acpi parser is right, it's a infinite
loop.
> I've dumped and disassembled the DSDT and looked at the above methods. What
> felt into my eyes immediately, were _OSI("Windows 2006") calls and then
> conditional code execution in these methods. So my next boot was with
> acpi_osi="!Windows 2006" and it fully repaired this problem. Now, the
> brightness keys really do regulate brightness (causing no acpi events),
> the VGA switcher generates the VGA ACPI event and of course other ACPI
> events are being generated without any problems.
> So, it would be good to either look at the problem and fix it in Linux,
> or put vista to the blacklist for this notebook type (in the case it is the
> DSDT bug).
Don't know why MSI do a named object value check in a infinite loop, I
thought it's a workaround in DSDT with Windows Vista's nVIDIA driver for
brightness control.
Could you please kindly provide the dmidecode dump to me? Then we can
send out patch to put this MSI notebook to blacklist.
> 2) Even with vista string disabled, the following warning is still printed
> to the log on every boot:
> Sep 25 14:09:03 beam kernel: ACPI Warning for \_SB_.PCI0.P0P2.VGA_.MXMI: Excess arguments - needs 1, found 2 (20100428/nspredef-319)
> Sep 25 14:09:03 beam kernel: ACPI Warning for \_SB_.PCI0.P0P2.VGA_.MXMS: Excess arguments - needs 1, found 2 (20100428/nspredef-319)
> It looks like a DSDT bug. I hope it is harmless.
> I've flashed the BIOS to the latest available (1.11).
> The DSDT table is attached.
>
I thought it should be ok.
Thank's a lot!
Joey Lee
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
