In message <20040114.011313.68554439.kochi@netbsd.org>, you write: 

-> Hi,
-> 
-> From: Rafal Boni <rafal@pobox.com>
-> Subject: Problems reading ACPI battery status on IBM ThinkPad T30
-> Date: Sun, 11 Jan 2004 22:55:34 -0500
-> 
-> > 	acpi0: X/RSDT: OemId <IBM   ,TP-1I   ,00002060>, AslId < LTP,00000000>
-> > 	    ACPI-1290: *** Error: Method execution failed [\_SB_.PCI0.LPC_.FDC_
-> ._INI] (Node 0xc0a24460), AE_NOT_EXIST
-> > 	    ACPI-1290: *** Error: Method execution failed [\_SB_.PCI0.LPC_.EC__
-> ._INI] (Node 0xc0a1e860), AE_NOT_EXIST
-> 
-> <snip>
-> > 	    ACPI-0178: *** Error: Method execution failed [\_SB_.NEST._STA] (No
-> de 0xc0a24320), AE_NOT_EXIST
-> 
-> These errors (many AE_NOT_EXIST) is due to missing support for
-> ACPI ECDT support.
-> 
-> I put my incomplete ECDT patch at
-> ftp://ftp.netbsd.org/pub/NetBSD/misc/kochi/acpica/acpi_ecdt_20031221.patch
-> (might not work right now)

I'll take a look at this in the next few days when I have some time and
can browse through the spec to get some clue of what this all means :-)
(getting things working better is a good carrot for getting more up
 to speed on ACPI ;-).

-> > While performing this query, the kernel outputs:
-> > 	ACPI-1290: *** Error: Method execution failed [\_SB_.PCI0.LPC_.EC__.GBS
-> T] (Node 0xc0a1e140), AE_AML_INTERNAL
-> > 	ACPI-1290: *** Error: Method execution failed [\_SB_.PCI0.LPC_.EC__.BAT
-> 0._BST] (Node 0xc0a1e000), AE_AML_INTERNAL
-> > 	acpibat0: failed to evaluate _BST: AE_AML_INTERNAL
-> 
-> These errors are source of my headache for more than half a year:(
-> (I have ThinkPad X31 and have the same problem)
-> 
-> The intel people provided a solution for this problem and I committed
-> the patch against ACPI CA 20031029, but they implemented another
-> incomplete solution in 20031203, which is incompatible with the
-> previous one, so I had to drop the patch.
-> 
-> I've been aware that FreeBSD people are arguing in the acpi-jp list
-> and I planned to import their changes if it's proven to be stable.

Ah, I see that it's had a long and ugly history :-/
Anyway, Nate Lawson's patch does seem to make my machine happier in this
regard and AFAICS there have been no resulting regressions though since
this appears to be a refcount issue I can see how you'd want to make sure
it's well shaken out lest we subtly leak memory instead.

-> Sorry for the mess.

No problem at all; in fact it is I and the rest of the ACPI users who
should be thanking you for dealing with this messy stuff and getting
it all working.

--rafal

----
Rafal Boni                                                     rafal@pobox.com
  We are all worms.  But I do believe I am a glowworm.  -- Winston Churchill