Bug 5236

Summary: AE_AML_BUFFER_LIMIT: ACPI + Battery monitor not funcitioning correctly
Product: ACPI Reporter: Scott Sommerville (scottsommerville)
Component: Power-BatteryAssignee: Robert Moore (Robert.Moore)
Status: REJECTED DUPLICATE    
Severity: high CC: acpi-bugzilla
Priority: P2    
Hardware: i386   
OS: Linux   
Kernel Version: 2.6.13* Subsystem:
Regression: --- Bisected commit-id:
Attachments: ACPI DMESG WHEN UNPLUGGING AC ADAPTOR
FURTHER ACPI DMESG ERRORS
Most recent DMESG
ACPIDUMP

Description Scott Sommerville 2005-09-12 20:52:10 UTC
Most recent kernel where this bug did not occur:2.6.12
Distribution:Slackware 10.1
Hardware Environment:SONY VAIO Laptop (Model No. PCG-FX501)
Software Environment:
Problem Description:Apon kernel recompilation for ACPI, the battery status is
monitored correctly, until reboot. Then ACPI correctly notes when AC adaptor is
plugged in or not, but fails to indicate the battery status. I have tried using
battery monitors in KDE3.4, XFCE and Gnome2.10 with no success. I have a
workaround by compiling APM support instead of ACPI in the kernel, however this
is not as efficient as it only tells me the battery status in % and is unable to
tell how much time is left

Steps to reproduce: Please let me know what you would like from me next, so I
can resolve this issue. Currently I am having to revert back to kernel 2.6.10 as
this kernel version works with amy ACPI and ALPS touchpad without issue.
Comment 1 Luming Yu 2005-09-12 23:54:14 UTC
Please post dmesg with kernel with acpi debug option turned on. 
Please post acpidump output. 
Comment 2 Scott Sommerville 2005-09-13 17:49:12 UTC
Created attachment 5998 [details]
ACPI DMESG WHEN UNPLUGGING AC ADAPTOR
Comment 3 Scott Sommerville 2005-09-13 17:50:32 UTC
Created attachment 5999 [details]
FURTHER ACPI DMESG ERRORS

Please let me know if you need more info
Comment 4 Luming Yu 2005-09-14 01:50:02 UTC
Thanks for the dmesg, it shows some AML interpreter messages. 
To debug further, please post acpidump output. 
Comment 5 Robert Moore 2005-09-14 14:48:49 UTC
Yes, please post acpidump output. We also need the version of ACPICA you are 
using, it should appear in dmesg.
Comment 6 Len Brown 2005-09-14 19:10:11 UTC
acpidump is avalable in the latest pmtools here:
http://ftp.kernel.org/pub/linux/kernel/people/lenb/acpi/utils
Comment 7 Scott Sommerville 2005-09-14 20:30:08 UTC
Created attachment 6028 [details]
Most recent DMESG
Comment 8 Scott Sommerville 2005-09-14 20:30:58 UTC
Comment on attachment 6028 [details]
Most recent DMESG

Hi Robert

I assume when you say ACPICA, that is the same as ACPI subsystem revision? If
so the version according to the most recent DMESG I ran is 20050408. If this is
not it let me know. As I mentioned I ran a more recent DMESG as I recompiled
the kernel for ACPI support built in instead of ACPI modules. Therefore I have
attached a copy of my recent DMESG. I have also attached the ACPIDUMP too. If
there is any other info you need please let me know.

Reagrds
Comment 9 Scott Sommerville 2005-09-14 20:33:23 UTC
Created attachment 6029 [details]
ACPIDUMP

Hi Robert

I assume when you say ACPICA, that is the same as ACPI subsystem revision? If
so the version according to the most recent DMESG I ran is 20050408. If this is
not it let me know. As I mentioned I ran a more recent DMESG as I recompiled
the kernel for ACPI support built in instead of ACPI modules. Therefore I have
attached a copy of my recent DMESG. I have also attached the ACPIDUMP too. If
there is any other info you need please let me know.

Reagrds
Comment 10 Robert Moore 2005-09-16 10:14:35 UTC
This is starting to look like an EC driver issue:

ACPI: AC Adapter [ACAD] (on-line)
[ACPI Debug]  String: [0x25] "------------------------ _Q20 : test1"
[ACPI Debug]  Integer: 0x00000014
[ACPI Debug]  String: [0x25] "------------------------ _Q20 : test2"
     osl-0965 [48] os_wait_semaphore     : Failed to acquire semaphore
[cfeee680|1|0], AE_TIME
     osl-0965 [48] os_wait_semaphore     : Failed to acquire semaphore
[cfeee680|1|5000], AE_TIME
Comment 11 Scott Sommerville 2005-09-16 10:42:13 UTC
Would it help me by recompiling the kernel for i686 CPU support, rather than 
the current Athlon support, to troubleshoot these potential EC driver issues?
Comment 12 Luming Yu 2005-09-18 08:19:14 UTC
Please confirm if you build kernel with preempt enabled.
If yes, please try non-preempt kernel.
Also try this patch 
http://bugzilla.kernel.org/show_bug.cgi?id=4588#c23

The problem was that the thread of executing _Q20 preempted the thread of 
executing BAT1._STA after BAT1._STA acquiring a semphore. This could cause 
_Q20 timeout.
Comment 13 Scott Sommerville 2005-09-22 08:14:38 UTC
Hi Robert and Lurning

Sorry I have not got back to you sooner, but I recently started a new job and
its been a little hectic.

With regarsd to whether preemption was built in to the kernel, this was the
case. Since then I have recompiled the kernel for no preemption and what do you
know the battery monitor status is now working.

Once again thanks for the help. On another note, will the lack of preemption
built into the kernel have any negative effects for my laptop?

Regards

Scott














































Comment 14 Luming Yu 2005-09-25 05:04:46 UTC

*** This bug has been marked as a duplicate of 4588 ***