Karmic desktop livecd constant kernel ACPI exception messages: AE_LIMIT, Returned by Handler for [SystemIO] 20090521 evregion-424

[Kalman Reti]

I booted the 64-bit Karmic Livecd on my Compaq Presario machine, and am constantly getting the following messages:

[ 927.906135] ACPI Error: Illegal I/O port address/length above 64K: 0x0000000000400020/4 20090521 hwvalid-154
[ 927.906152] ACPI Exception: AE_LIMIT, Returned by Handler for [SystemIO] 20090521 evregion-424
[ 927.906164] ACPI Error (psparse-0537): Method parse/execution failed [\_GPE._L09] (Node ffff88003b5a7520), AE_LIMIT
[ 927.906217] ACPI Exception: AE_LIMIT, while evaluating GPE method [_L09] 20090521 evgpe-568

Perhaps unsurprisingly, it is also slow.

ProblemType: Bug
Architecture: amd64
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC0: ubuntu 3146 F.... pulseaudio
 /dev/snd/controlC1: ubuntu 3146 F.... pulseaudio
CRDA: Error: [Errno 2] No such file or directory
Card0.Amixer.info:
 Card hw:0 'IXP'/'ATI IXP rev 2 with ALC658D at 0xfe02a000, irq 17'
   Mixer name : 'Realtek ALC658D'
   Components : 'AC97a:414c4781'
   Controls : 38
   Simple ctrls : 24
Card1.Amixer.info:
 Card hw:1 'CX8801'/'Conexant CX8801 at 0xfa000000'
   Mixer name : 'CX88'
   Components : ''
   Controls : 3
   Simple ctrls : 2
Date: Thu Nov 5 13:44:59 2009
DistroRelease: Ubuntu 9.10
IwConfig:
 lo no wireless extensions.

 eth0 no wireless extensions.
LiveMediaBuild: Ubuntu 9.10 "Karmic Koala" - Release amd64 (20091027)
Lsusb:
 Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
 Bus 003 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
 Bus 002 Device 002: ID 058f:9360 Alcor Micro Corp. 8-in-1 Media Card Reader
 Bus 002 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
MachineType: Compaq Presario 061 EL445AA-ABA SR1750NX NA650
Package: linux-image-2.6.31-14-generic 2.6.31-14.48
ProcCmdLine: BOOT_IMAGE=/casper/vmlinuz file=/cdrom/preseed/hostname.seed boot=casper initrd=/casper/initrd.lz quiet splash --
ProcEnviron:
 LANG=en_US.UTF-8
 SHELL=/bin/bash
ProcVersionSignature: Ubuntu 2.6.31-14.48-generic
RelatedPackageVersions:
 linux-backports-modules-2.6.31-14-generic N/A
 linux-firmware 1.24
RfKill:

SourcePackage: linux
Uname: Linux 2.6.31-14-generic x86_64
XsessionErrors:
 (gnome-settings-daemon:3204): GLib-CRITICAL **: g_propagate_error: assertion `src != NULL' failed
 (gnome-settings-daemon:3204): GLib-CRITICAL **: g_propagate_error: assertion `src != NULL' failed
 (nautilus:3260): Eel-CRITICAL **: eel_preferences_get_boolean: assertion `preferences_is_initialized ()' failed
 (polkit-gnome-authentication-agent-1:3332): GLib-CRITICAL **: g_once_init_leave: assertion `initialization_value != 0' failed
 (gnome-panel:3247): Gtk-WARNING **: gtk_widget_size_allocate(): attempt to allocate widget with width -4 and height 24
dmi.bios.date: 11/15/2005
dmi.bios.vendor: Phoenix Technologies, LTD
dmi.bios.version: 3.13
dmi.board.name: Amberine M
dmi.board.vendor: ASUSTek Computer INC.
dmi.board.version: 1.03
dmi.chassis.type: 3
dmi.chassis.version: 1111
dmi.modalias: dmi:bvnPhoenixTechnologies,LTD:bvr3.13:bd11/15/2005:svnCompaqPresario061:pnEL445AA-ABASR1750NXNA650:pvr0nx1411RE101AMBEM00:rvnASUSTekComputerINC.:rnAmberineM:rvr1.03:cvn:ct3:cvr1111:
dmi.product.name: EL445AA-ABA SR1750NX NA650
dmi.product.version: 0nx1411RE101AMBEM00
dmi.sys.vendor: Compaq Presario 061

[floid (jkanowitz)]

Just noticed the same on identical hardware (SR1611NX) running 32-bit,
Linux ________ 2.6.31-15-generic #50-Ubuntu SMP Tue Nov 10 14:54:29 UTC 2009 i686 GNU/Linux.

Dumps are attached. This does not appear to be affecting performance, which is quite reasonable on an 'installed' system.

[floid (jkanowitz)]

...where offending methods are located. Not quite human-readable enough for me to debug at the moment.

[anotherstuart (stuart-astill)]

I get exactly the same error lines, several times every minute on my Compaq presario as well.

[Nels (nels-n-nelson)]

I have the same problem.

$ cat /etc/lsb-release
DISTRIB_ID=Ubuntu
DISTRIB_RELEASE=9.10
DISTRIB_CODENAME=karmic
DISTRIB_DESCRIPTION="Ubuntu 9.10"

$ uname -a
Linux dolphin 2.6.31-15-generic #50-Ubuntu SMP Tue Nov 10 14:53:52 UTC 2009 x86_64 GNU/Linux

Non stop errors.

Errors:
Dec 16 08:20:18 dolphin kernel: [977235.651601] ACPI Exception: AE_LIMIT, while evaluating GPE method [_L09] 20090521 evgpe-568
Dec 16 08:20:18 dolphin kernel: [977235.678526] ACPI Error: Illegal I/O port address/length above 64K: 0x0000000000400020/4 20090521 hwvalid-154
Dec 16 08:20:18 dolphin kernel: [977235.678538] ACPI Exception: AE_LIMIT, Returned by Handler for [SystemIO] 20090521 evregion-424
Dec 16 08:20:18 dolphin kernel: [977235.678548] ACPI Error (psparse-0537): Method parse/execution failed [\_GPE._L09] (Node ffff88003b9a7520), AE_LIMIT

$ du -hs syslog*
932K syslog
2.0M syslog.1
136K syslog.2.gz

Not good.

[Paco (adelb80)]

Hello, I got the same problem on my Compaq presario as well, not exaclty the very same messages but it looks like it...

kernel: [ 1762.911519] ACPI Exception: AE_LIMIT, while evaluating GPE method [_L09] 20090521 evgpe-568
kernel: [ 1762.928228] ACPI Error: Illegal I/O port address/length above 64K: 0x00400020/4 20090521 hwvalid-154
kernel: [ 1762.928243] ACPI Exception: AE_LIMIT, Returned by Handler for [SystemIO] 20090521 evregion-424
kernel: [ 1762.928254] ACPI Error (psparse-0537): Method parse/execution failed [\_GPE._L09] (Node f700edb0), AE_LIMIT

Any solution ?

~$ cat /etc/lsb-release
DISTRIB_ID=Ubuntu
DISTRIB_RELEASE=9.10
DISTRIB_CODENAME=karmic
DISTRIB_DESCRIPTION="Ubuntu 9.10"

~$ uname -a
Linux desktop 2.6.31-17-generic #54-Ubuntu SMP Thu Dec 10 16:20:31 UTC 2009 i686 GNU/Linux

Cheers,

Paco

[floid (jkanowitz)]

Hmm. Happened to be at the machine when I saw the last comment, so I decided to poke at it.
Attempting to recompile my SR1611NX DSDT with iasl (note that the -dc option doesn't seem to be working as advertised, but simply running `iasl DSDT.dsl` works at the expense of obliterating the contents of the directory, and `iasl -c DSDT.dsl` will try the same but create an empty .aml and a populated .src - I guess because the errors prevent it from actually generating the .aml).

...

Intel ACPI Component Architecture
ASL Optimizing Compiler version 20090521 [Jun 30 2009]
Copyright (C) 2000 - 2009 Intel Corporation
Supports ACPI Specification Revision 3.0a

DSDT.dsl 284: Method (\_WAK, 1, NotSerialized)
Warning 1080 - ^ Reserved method must return a value (_WAK)

DSDT.dsl 410: WBYT (TSAD, 0x4F, Local0)
Error 4050 - ^ Method local variable is not initialized (Local0)

DSDT.dsl 632: CSMS ()
Warning 1099 - ^ Statement is unreachable

DSDT.dsl 646: CSMS ()
Warning 1099 - ^ Statement is unreachable

DSDT.dsl 1406: Name (_PSW, 0x01)
Error 4078 - ^ Reserved name must be a control method (with arguments)

DSDT.dsl 1428: Name (_PSW, 0x01)
Error 4078 - ^ Reserved name must be a control method (with arguments)

DSDT.dsl 1450: Name (_PSW, 0x01)
Error 4078 - ^ Reserved name must be a control method (with arguments)

DSDT.dsl 2985: Field (SECF, WordAcc, NoLock, Preserve)
Error 4027 - ^ Access width is greater than region size

DSDT.dsl 2987: SCFN, 8
Error 4028 - ^ Access width of Field Unit extends beyond region limit

DSDT.dsl 3534: Method (IPHP, 2, NotSerialized)
Warning 1087 - ^ Not all control paths return a value (IPHP)

DSDT.dsl 3888: If (LEqual (HGDT (Local1), 0x01))
Warning 1092 - ^ Called method may not always return a value

DSDT.dsl 3912: If (LEqual (IPHP (0x06, Local1), Local3))
Warning 1092 - Called method may not always return a value ^

DSDT.dsl 3946: Method (HGDT, 1, NotSerialized)
Warning 1087 - ^ Not all control paths return a value (HGDT)

DSDT.dsl 3994: And (IPHP (0x07, Arg0), Arg1, Local2)
Warning 1092 - ^ Called method may not always return a value

ASL Input: DSDT.dsl - 5409 lines, 165894 bytes, 1913 keywords
Compilation complete. 6 Errors, 8 Warnings, 0 Remarks, 626 Optimizations

...

I know slightly less than enough about ACPI to be dangerous; this is what I've been able to figure out thus far:

Our troublesome GPE.L09 is a "General Purpose Event" register.. but I wish I had a clue what that "SMFN" method that relates to it is [supposed to be] doing.

As to what iasl dislikes (but MSFT Compiler Version 0x0100000E (16777230) obviously was happy to generate):

In the "STOS" method:

DSDT...

[junglist313 (junglist313)]

I get the same errors, compaq presario sr1750nx here.

[junglist313 (junglist313)]

I should mention that I am not running the live cd, this is on a regular install, ubuntu 9.10 fully upgraded.

[Banks (mike3man)]

I do not have the skill set to offer any valuable advice here, but wanted to throw in my two cents. I discovered that I was having this problem quite accidentally. I was trying the locate the reason that my computer freezes 2 minutes into a recording and just started looking at logs. I do not know that my problem is related to this issue, but my point is that I would never have known this was occurring if I had not stumbled upon the logs. Could be a bigger problem than is realized and users just do not know it is happening. I am running Mythbuntu 9.10 all updates. I did not start having the freezing problem until a couple weeks ago, January 29 to be specific. Just FYI.

[Derrick Weslowski (derrick-weslowski)]

Full 64bit 9.10 install...fully updated on a Compaq SR1710NX with the same issue.

[Jeremy Foshee (jeremyfoshee)]

Hi Kalman,

If you could also please test the latest upstream kernel available that would be great. It will allow additional upstream developers to examine the issue. Refer to https://wiki.ubuntu.com/KernelMainlineBuilds . Once you've tested the upstream kernel, please remove the 'needs-upstream-testing' tag. This can be done by clicking on the yellow pencil icon next to the tag located at the bottom of the bug description and deleting the 'needs-upstream-testing' text. Please let us know your results.

Thanks in advance.

    [This is an automated message. Apologies if it has reached you inappropriately; please just reply to this message indicating so.]

[Derrick Weslowski (derrick-weslowski)]

What I found helped out was going into my BIOS settings at boot-up time with F11. Then going into the Power tab and disabling ACPI. Then I still get about 10 errors like the one above when I startup, but at least its not the constant log filling that was happening before. I found some documentation listing the BIOS in the SR1700 series of Compaqs as broken.

http://www.columbia.edu/~ariel/acpi/acpi_howto.txt

[floid (jkanowitz)]

As I found above, it's a simple and silly type error, but the more thorny fatal incompatibility between recent iasl and the original code is beyond my familiarities. Running the simple patch through the Microsoft compiler would probably do the trick, but it would take someone with a bit more familiarity (half a clue) to 'port' around that _PSW SNAFU.

Anyone able to flag down such a person?

[johnny b (stepore)]

I'm running Lucid RC and I'm getting the same errors on an HP Pavilion:
Apr 27 23:15:02 lucid64-quinhas kernel: [176009.198578] ACPI Exception: AE_LIMIT, Returned by Handler for [SystemIO] (20090903/evregion-424)
Apr 27 23:15:02 lucid64-quinhas kernel: [176009.198590] ACPI Error (psparse-0537): Method parse/execution failed [\_GPE._L09] (Node ffff880098e2d520), AE_LIMIT
Apr 27 23:15:02 lucid64-quinhas kernel: [176009.198646] ACPI Exception: AE_LIMIT, while evaluating GPE method [_L09] (20090903/evgpe-568)
Apr 27 23:15:02 lucid64-quinhas kernel: [176009.208989] ACPI Error: Illegal I/O port address/length above 64K: 0x0000000000400020/4 (20090903/hwvalid-154)
Apr 27 23:15:02 lucid64-quinhas kernel: [176009.209000] ACPI Exception: AE_LIMIT, Returned by Handler for [SystemIO] (20090903/evregion-424)

The computer won't boot with acpi=off but I have turned off acpid and acpi-support and I still get the errors. The logs are huge.

[floid (jkanowitz)]

I did some scraping around after my last comment and determined that the closest thing to a proper venue 'upstream' would be http://acpica.org/bugzilla/ , but someone ought to do due-diligence and make sure the collateral issues in #7 here aren't magically resolved in their bleeding-edge code first.

(Again, the patch regarding the ByteAcc vs. WordAcc should fix the bug found here, but the fatal compiler errors about the _PSW method smell like they could be due to changes between the version of the ACPI standard the original authors of the HPaq code targeted vs. what iasl supports compiling. Determining the correct workaround for *that* seems to require someone who can parse the actual workings of the code, rather than just respond to 'obvious' error messages as I did. :))

I have a bunch of other stuff on my plate at the moment ["real life"/the day job], so if anyone can run with that thought I'd appreciate it.

[Frank Larimer (larimerfw)]

My "solution" was to divert the log entries to a new log:

/etc/rsyslog.d/30-ACPI.conf
# Discard ACPI messages
:msg,regex,"ACPI E" /var/log/ACPI.log
&~

This way the other logs aren't full of junk.

[Frank Larimer (larimerfw)]

Just built lucid 10.04 LTS x86_64 - it generates the same ACPI errors, every six seconds.

[Eustace Scrubb (scrubb-techgamers)]

I too see this errors too, filling my logs. It appears that my performance is suffering miserably as well. Mine is a Compaq Sr1700 series workstation and the errors that recur are below:

[187285.413470] ACPI Error (psparse-0537): Method parse/execution failed [\_GPE._L09] (Node f700fdb0), AE_LIMIT
[187285.413521] ACPI Exception: AE_LIMIT, while evaluating GPE method [_L09] (20090903/evgpe-568)

Is this bug on the official "its being worked on list?" or are we a fringe element?
Thanks,
E

[Frank Larimer (larimerfw)]

Running lucid, I eliminated this behavior by building a new kernel with this deferred patch: https://patchwork.kernel.org/patch/100574/

Of course I have to build a new kernel each time updates issue, until this patch is actually implemented in the kernel.

[Jeremy Foshee (jeremyfoshee)]

This bug report was marked as Incomplete and has not had any updated comments for quite some time. As a result this bug is being closed. Please reopen if this is still an issue in the current Ubuntu development release http://cdimage.ubuntu.com/daily-live/current/ . Also, please be sure to provide any requested information that may have been missing. To reopen the bug, click on the current status under the Status column and change the status back to "New". Thanks.

[This is an automated message. Apologies if it has reached you inappropriately; please just reply to this message indicating so.]

[floid (jkanowitz)]

To my delight, this does appear fixed with:
Linux ubuntu 2.6.35-22-generic #33-Ubuntu SMP Sun Sep 19 20:34:50 UTC 2010 i686 GNU/Linux
using the current 10.10 RC desktop CD (i386). I assume the patch referenced in https://bugs.launchpad.net/ubuntu/+source/linux/+bug/475704/comments/18 landed?

Unfortunately the machine still freezes with no display when resuming from suspend, probably unrelated.

Unfortunately the ACPI chatter is replaced by as much "serial8250: too much work for irq17" noise, but that's an unrelated bug.*

*Footnote: Apparently due to presence of this modem, which happens to be sharing IRQ17 with the Radeon and AC97 controller, go figure. Will hunt down the right bug for this one...

02:09.0 Modem: Motorola SM56 Data Fax Modem (rev 04) (prog-if 00 [Generic])
 Subsystem: Motorola Device 3020
 Flags: bus master, stepping, medium devsel, latency 64, IRQ 17
 Memory at fdcfd000 (32-bit, non-prefetchable) [size=4K]
 I/O ports at da00 [size=256]
 Capabilities: <access denied>
 Kernel driver in use: serial

[Ton van Vliet (epposan)]

Is there any chance to get this thread re-opened?

My syslog is (still) full of:

[26570.230040] ACPI Error: Illegal I/O port address/length above 64K: 0x00400020/4 (20090903/hwvalid-154)
[26570.230061] ACPI Exception: AE_LIMIT, Returned by Handler for [SystemIO] (20090903/evregion-424)
[26570.230077] ACPI Error (psparse-0537): Method parse/execution failed [\_GPE._L09] (Node f7010d98), AE_LIMIT
[26570.230142] ACPI Exception: AE_LIMIT, while evaluating GPE method [_L09] (20090903/evgpe-568)

The system is an HP Pavilion s7450

$ uname -a
Linux kato-pc 2.6.32-25-generic #45-Ubuntu SMP Sat Oct 16 19:48:22 UTC 2010 i686 GNU/Linux

$ cat /etc/lsb_release
DISTRIB_ID=Ubuntu
DISTRIB_RELEASE=10.04
DISTRIB_CODENAME=lucid
DISTRIB_DESCRIPTION="Ubuntu 10.04.1 LTS"

Thanks.

[floid (jkanowitz)]

@Ton van Vliet:

See #20 for a "by hand" patch for the actual problem - thanks to Frank Larimer for identifying it there.

This or an equivalent made it into 10.10's kernel as described in my #22, so 10.10 is the lazy option if you want to upgrade. [Do I read it right: The MS compiler generates code with malformed addresses that are supposed to be 32-bit by the spec, while MS Windows cheerfully ignores upper 16 bits? Or is this something the author of the DSDT has to directly screw up, and Windows' behavior just lets them get away with it?]

re: 10.10 on these systems, the unrelated noise from the other bug mentioned in my #22 is actually more intermittent/not nearly so much of an issue or annoyance as I first thought, if it's not already fixed anyway [haven't checked since 10.10 went final, it may have been fixed by now].

Importing or enabling the "N,000 messages: Ignoring" feature as found in certain BSDs might not be a bad idea, still...

[Ton van Vliet (epposan)]

@floyd:

Does this mean that the 'fix' might never be backported to LTS 10.04's kernel(s)?

If so, I am considering an even lazier option by applying Frank Larimers 'solution' as per #17

[floid (jkanowitz)]

@Ton van Vliet:

I'm not in charge and barely know who is. Since it hasn't been backported yet [as far as I know - I think my 10.04 machines are demanding another reboot after auto-updates], ignoring the messages is certainly one option.

Not to grouse about it [since I've gotten used to it] but the attitude to LTS releases seems to be "if something was broken, don't have any expectation that 'support' means it should be fixed, 'support' equals patches for security and major functional issues if a Y2K-type bug pops up across all systems during the 'support' period." (Actually, keeping the zoneinfo package current for the world's ever-changing DST rules is probably a good example of what 'support' really means.) So since it was doing this from day one, this bug is just an ongoing characteristic ('feature'?) of 10.04.

Of course, if we flag someone down by making noise here, the *real* patch is simple enough now that it's identified that I don't think anyone could complain about getting it in the tree.

In my brief evaluation of 10.10 I forgot to check whether the kernel with the actual patch [and any other changes] would solve the issues with suspend/resume on this hardware. That could be worth checking if suspend/resume is of interest to you, I'll have to go back and try it myself.

[Ton van Vliet (epposan)]

@floid

Thanks for sharing your view on this!

I guess I have to adapt my personal 'expectation' of the meaning of LTS.

Having no idea how many people are (unknowingly) hit by this bug, I don't feel like making noise right now.

I am too having problems with suspend/resume on the same system that has the ACPI errors.

That system also (sometimes) has start-up problems: it does not get to the log-in screen, but remains dark, and needs a 'power button reset' or a remote log-in to reboot it (via ssh, which does work). Possibly unrelated though.

After reading your post, I am now considering to give it a shot with stepping up to 10.10, even though only one of my systems suffers the ACPI
issue, but I like to keep all systems in sync.

[Ton van Vliet (epposan)]

@floid

FWIW, I ran a 'do-release-upgrade -d' on the system to 10.10 (which completed without issues)

The ACPI errors do not show up any more, however suspend/resume is still a 'no go' (it used to work fine under 8.04)

[Tim Temple (ttemple5-f)]

Compaq Presario desktop using ubuntu 10.04 LTS with no ACPI installed.

Same problem with messages.log maxed out. The continual logging bogs the system. If I edit the startup to include "acpi=off" or "pci=noacpi", the error stops logging, but NAT won't connect to the internet via eth0.
Repeating log entries:
Jun 11 10:25:51 timt-desktop kernel: [ 1308.823339] ACPI Error: Illegal I/O port address/length above 64K: 0x00400020/4 (20090903/hwvalid-154)
Jun 11 10:25:51 timt-desktop kernel: [ 1308.823357] ACPI Exception: AE_LIMIT, Returned by Handler for [SystemIO] (20090903/evregion-424)
Jun 11 10:25:51 timt-desktop kernel: [ 1308.823370] ACPI Error (psparse-0537): Method parse/execution failed [\_GPE._L09] (Node db80fd68), AE_LIMIT
Jun 11 10:25:51 timt-desktop kernel: [ 1308.823422] ACPI Exception: AE_LIMIT, while evaluating GPE method [_L09] (20090903/evgpe-568)

Same problem with Fedora and Red Hat. Red Hat says they fixed it by "Truncate I/O addresses to 16 bits for Windows compatibility."
Fedora says their source code fix is here: http://kerneltrap.org/mailarchive/git-commits-head/2009/4/5/5411984

[Brad Figg (brad-figg)]

This bug was filed against a series that is no longer supported and so is being marked as Won't Fix. If this issue still exists in a supported series, please file a new bug.

This change has been made by an automated script, maintained by the Ubuntu Kernel Team.