HOWTO: Fix Common ACPI Problems (DSDT, ECDT, etc.)

MickKi · Veteran Joined: 08 Feb 2004 Posts: 1173

Hi All,

Like a previous poster I am getting a lot of 1061 errors. Any suggestions for fixes?

dgaffuri · Advocate Joined: 05 Jun 2005 Posts: 2078 Location: Italy

I solved 1061 errors with CPU0 adding an external declaration

MickKi · Veteran Joined: 08 Feb 2004 Posts: 1173

Thanks dgaffuri,

. . . but I don't know what is an "external declaration", or which external declaration would be appropriate in my case. Any suggestions?
_________________
Regards,
Mick

dgaffuri · Advocate Joined: 05 Jun 2005 Posts: 2078 Location: Italy

MickKi · Veteran Joined: 08 Feb 2004 Posts: 1173

Wow! It worked!

Thank you dgaffuri. I managed to fix at least some of these 1061 errors. I entered the following two lines:

TheRuler · Tux's lil' helper Joined: 06 Aug 2003 Posts: 96

Hello,

my DSDT is also broken can you help me?

ale-X · n00b Joined: 30 Apr 2006 Posts: 1

Hi to all,
I have succesfully recompiled the DSDT but my system says "battery is absent" and the dmesg says:

hakbeest · n00b Joined: 26 Aug 2004 Posts: 27

I have a buggy DSDT too, unfortunately /proc/acpi doesn't exist, so I can't cat /proc/acpi/dsdt. I get the following acpi related messages during boot:

BIOS-e820: 000000003bfd0000 - 000000003bfde000 (ACPI data)
BIOS-e820: 000000003bfde000 - 000000003c000000 (ACPI NVS)
ACPI: RSDP (v000 ACPIAM ) @ 0x00000000000f96e0
ACPI: RSDT (v001 A M I OEMRSDT 0x10000514 MSFT 0x00000097) @ 0x000000003bfd0000
ACPI: FADT (v002 A M I OEMFACP 0x10000514 MSFT 0x00000097) @ 0x000000003bfd0200
ACPI: MADT (v001 A M I OEMAPIC 0x10000514 MSFT 0x00000097) @ 0x000000003bfd0390
ACPI: MCFG (v001 A M I OEMMCFG 0x10000514 MSFT 0x00000097) @ 0x000000003bfd03f0
ACPI: OEMB (v001 A M I AMI_OEM 0x10000514 MSFT 0x00000097) @ 0x000000003bfde040
ACPI: DSDT (v001 1ABZR 1ABZRB49 0x00000b49 INTL 0x02002026) @ 0x0000000000000000
ACPI: PM-Timer IO Port: 0x808
ACPI: Local APIC address 0xfee00000
ACPI: LAPIC (acpi_id[0x01] lapic_id[0x00] enabled)
ACPI: LAPIC (acpi_id[0x02] lapic_id[0x81] disabled)
ACPI: Skipping IOAPIC probe due to 'noapic' option.
Using ACPI for processor (LAPIC) configuration information
tbxface-0109 [02] load_tables : ACPI Tables successfully acquired
ACPI Error (nssearch-0311): Bad character in ACPI Name: 43035350 [20060127]
ACPI Error (dswload-0305): [0x43035350] (NON-ASCII) Namespace lookup failure, AE_BAD_CHARACTER
ACPI Exception (psloop-0281): AE_BAD_CHARACTER, During name lookup/catalog [20060127]
ACPI Exception (tbxface-0115): AE_BAD_CHARACTER, Could not load namespace [20060127]
ACPI Exception (tbxface-0122): AE_BAD_CHARACTER, Could not load tables [20060127]
ACPI: Unable to load the System Description Tables
ACPI: Subsystem revision 20060127
ACPI: Interpreter disabled.

hakbeest · n00b Joined: 26 Aug 2004 Posts: 27

By using a patch that accepts the illegal 0x03 character (http://bugzilla.kernel.org/show_bug.cgi?id=6522) I was able to get ACPI working. It turned out that there were no other errors. My DSDT was compiled by the Intel compiler, according to MSI this one:

Intel ACPI Component Architecture ASL Compiler X2026 [Aug 31 2001]
Supports ACPI Specification Revision 2.0

which seems a bit dated as my version says:

Intel ACPI Component Architecture ASL Optimizing Compiler version 20060127 [May 8 2006]
Supports ACPI Specification Revision 3.0a

By replacing the illegal character by an 'S' I was able to decompile and recompile. I'm trying to get something definate on this illegal character from MSI, but their technical support seem to lack the requisite English and technical skills.

Augur · n00b Joined: 04 Sep 2005 Posts: 3

Hi,

I just bought an new laptop, and i am trying to fix some acpi problems but I don't understand
iasl errors

screening · n00b Joined: 16 Jun 2006 Posts: 1

Please, anyone can help me with this warnings?, i don't have idea what can i do.

Intel ACPI Component Architecture
ASL Optimizing Compiler version 20051216 [Jan 9 2006]
Copyright (C) 2000 - 2005 Intel Corporation
Supports ACPI Specification Revision 3.0

dsdt.dsl 2589: Method (RVLT, 1, NotSerialized)
Warning 2085 - ^ Not all control paths retur n a value (RVLT)

dsdt.dsl 3623: Method (OCOP, 1, NotSerialized)
Warning 2085 - ^ Not all control paths retur n a value (OCOP)

dsdt.dsl 4010: Method (PROC, 3, NotSerialized)
Warning 2085 - ^ Not all control paths retur n a value (PROC)

dsdt.dsl 4505: Store (PROC (Local0, Local1, 0x00), Local2)
Warning 2090 - ^ Called method may not al ways return a value

dsdt.dsl 6142: Release (SMBA)
Warning 2097 - Statement is unreachable ^

ASL Input: dsdt.dsl - 8425 lines, 297000 bytes, 3224 keywords
AML Output: DSDT.aml - 26244 bytes 708 named objects 2516 executable opcodes

Compilation complete. 0 Errors, 5 Warnings, 0 Remarks, 835 Optimizations

My computer is a desktop, mother ASUS P5GDC Deluxe. If anyone can help... i appreciate.

PS: sorry 4 my bad english.

Moderato · Posted: Thu Jun 22, 2006 6:02 pm Post subject: Optimizations

Hey guys,

though, the last real answer is a little longer ago, I try to ask a little question in here.

This IASL now reports "Compilation complete. 0 Errors, 0 Warnings, 0 Remarks, 36 Optimizations". I'm wondering, what it does mean by "36 Optimizations". The Code should now be "fixed" (I had just this _WAK-bug), but IASL still reports this message. Can I turn on the optimization somehow? I did not find any option, but a option to disable it. I'm not sure, this message is showing up every time and just after recompiling the table, too. However, I'm using the new, recompiled table and thought the optimizations would stuck just in there. It seem they aren't.

Any suggestions?

Gremo · Guru Joined: 27 Feb 2006 Posts: 450

different version of iasl produce different errors or NO ERRORS, what does it mean, try yourself

spindle · Apprentice Joined: 01 Dec 2003 Posts: 245

Moderato · Posted: Sun Jul 16, 2006 7:14 am Post subject:

I'd rather say yes, but keep in mind that the ACPI implementation in the Kernel is still somewhat experimental and far from being complete, though. Even if your DSDT is clean, there might remain some errors.

For me, Kernel 2.6.17 was changing the behavior when shutting down my Notebook - it's still not powering down properply, when I started a Desktop Environment like GNOME, KDE, FluxBox, BlackBox, XFCE (I'd rather say "when I did log-on in GDM"), but it's powering down, when I did not start any DE. X and GDM or the nVidia driver can't be the problems since it's powering down, when only X w/ GDM is running (but I don't have to log-in).

However, keep another thing in mind: modifing your firmware (-> your DSDT) can have unwanted "side effects", too. So, you might fixed one problem but this may _produced_ a lot more.
Also, no Kernel developer will help you with your problems, if you changed your DSDT since they may cannot locate the bug in the Kernel when it's hidden by an own DSDT.

You should try the following everytime before modifing anything:

These are things, that _can_ help with those issues.

spindle: I'm not sure about this, but if your mobo/your BIOS was quite new, it's common that older version of the IASL show errors, because there might be some ACPI calls the old IASL didn't know. Check this line:

spindle · Apprentice Joined: 01 Dec 2003 Posts: 245

Moderato: thanks for the helpful tips.

I just reverted to the default dsdt that the kernel gets from the mobo (kernel 2.6.16-gentoo-r12). I'll run like this for a while, and see if it fixes my possible overheating problem (I was seeing this issue before). I was running a modified dsdt, from an earlier attempt to fix a problem using the older 2005 IASL version. A while ago, I was trying to fix a problem with the termal reading on the CPU. It always said 40.0C and it still does anyway. I think your suspicion could be right, in trying to fix this problem I could have messed up some other important things in there.

The board is a fairly new model I think, from around summer 2005 I'm guessing (ABIT GD8). I'll check for a new BIOS too.

Moderato · Posted: Sun Jul 16, 2006 3:50 pm Post subject:

Hi spindle,

interesting issue. Are you running a Laptop?

You could also change to Kernel 2.6.17.6 for instance (as I said - for me better ACPI features and the new exploit got fixed, too).
Make sure you activited all the ACPI features in the Kernel, and, if you own a Laptop, double check whether you have implemented the thermal_zone (Symbol ACPI_THERMAL) and also check for those AMD processor "hacks" for detecting overheating, soft lockups and so on. Don't forget to enable verbose ACPI debugging in the Kernel. Even if you don't have a Laptop, enable them all.

If this still didn't help subscribe to the ACPI Mailinglist, search the ACPI bugzilla (it's also a Mailinglist) for the symptoms you observed.

In the case nothing helped and the bug is not known, send a bug report the Kernel development team or the Kernel ACPI team. (If you know, what you are doing, have really a lot of _useful_ informations towards this, also compiled your Kernel with frame pointers and unwind informations, have debug stacks, straces etc pp you could also directly send a bug report to the ACPI maintainer, but I discourage you from doing this, sending it to the Kernel guys will be fine, too, and quite more useful.)

So long, good luck.

[EDIT] By the way: I know this strange behavior of showing only "40.0C" as temperature from my Desktop, too. I'm not sure whether this is a DSDT bug or a Kernel/ACPI bug, I didn't fixed it, because I never had "real" thermal problems. I checked the DSDT of my Desktop computer and freaked out a little bit when I saw all these errors. Although, I never had problems with ACPI, so an error doesn't mean automagically that everything is b0rken in the ACPI implementation.

I also do not know how this bug behaves towards the THERMAL_ZONE which seems elementary. I'm worrying about this, now. But however, there might be some informative debug messages. [/EDIT]

spindle · Apprentice Joined: 01 Dec 2003 Posts: 245

Moderato · Posted: Mon Jul 17, 2006 2:09 pm Post subject:

Hi spindle,

ah, so I can assume we have ~the same CPU. (I got an Intel P IV 2998MHz in the Desktop and thus, the same issue.)

Yeah, playing UT2004 is NOT that CPU stressing - it's the GPU/your video card that might get really hot in this case, but the CPU is not under a real lot of load - in opposite to compiling...

Right, the BIOS is sending some messages about thermal throttling and it seems like you compiled this feature into the Kernel correctly. The bad news, however, is that this is REALLY bad. You should really search for a better cooling system. Your CPU is definitly way too hot. (Let me explain it like this... the BIOS needn't the Kernel for checking the temperature and therefore it's working fine just without it, although you cannot check the temperature via ACPI.)

I don't know whether the THERMAL_ZONE is working in Kernel 2.6.17.x correctly, but I assume the P IV just hasn't such ACPI informations.

For looking at the CPU temperature, I'd rather try lm_sensors. You need the I2C subsystem in the Kernel, if I remember correctly. Play with this program; it relates directly on the thermal sensor on your CPU.

In any case: I wish you good luck. And please, get a better cooling system.

[By the way: I also have some thermal issues on my Desktop with the P IV when playing games. In fact, the graphic card (ATI Radeon 9800 XT) is getting too hot and because my computer's case is almost small, freezes of the computer are happening, too. Even if I soft restart the computer it won't boot at all. Only after I let the computer cooling down everything works again. I think this happens because of the BIOS' thermal check.]

spindle · Apprentice Joined: 01 Dec 2003 Posts: 245

topper_harley · Posted: Wed Jul 26, 2006 1:04 pm Post subject:

I'm using a Toshiba Satellite M40-141. I have never had the fan working with any Linux distro.
The DSTD was horribly broken, I couldn't even recompile it.

I found in the DSDT repository a good DSDT for my laptop (which has a Phoenix Bios, instead of a Toshiba one). I compiled it with no errors and no warnings.
It works greatly and now I can make the fan work.

But I have a REALLY BIG problem. If I try to play a sound (any sound) the computer hags (neither the mouse cursor moves) and I have to poweroff manually.

When I reboot I cannot find any useful logs about the crash.

Now I removed alsasound from my init scripts, and I'm using a soundless computer. But I'd really like to fix this.

Any ideas?
_________________
http://topperh.ath.cx
Jabber: topper_harley@jabber.org
ICQ: 224179391
MSN: Topper_Harley80@gmail.com
Last FM

Moderato · Posted: Wed Jul 26, 2006 2:43 pm Post subject:

Are there even in dmesg no hints about Interrupts or somethings like this? Any conflicts?

I'd suggest trying the "noapic" and "nolapic" boot option first -- if it still crashes, we have to wait for other poeple and their ideas.

As an example: kernel /boot/vmlinuz-$something root=/dev/$yourdev noapic nolapic $theotheroptionsyouhave

Have a nice day.

topper_harley · Posted: Fri Jul 28, 2006 8:32 am Post subject:

Yak · Tux's lil' helper Joined: 01 Sep 2002 Posts: 107

I'm trying to fix a dsdt on an Averatec 1020-ED1. My main reason for mucking around with this is that it seems the fan isn't running at all, nothing in /proc/acpi/fan. I followed the Howto but I can't get it to recognize my new DSDT.aml, dmesg keeps saying

Yak · Tux's lil' helper Joined: 01 Sep 2002 Posts: 107

Seems I was on the right track about the initramfs, so guess I'll answer my own question. Thanks go to Rob for his Tech Tip: How to use initramfs

Howto put DSDT.aml into an initrd image:

Make a directory, this will be the root of the image. Move the DSDT.aml file into this directory and change to it. Use find to list the files and pipe through cpio and gzip and output to the image file. For example: