No battery in /proc/acpi/battery (was working earlier!)

[Dean Jansen]

I'm using hoary on a Panasonic R1, and had the battery monitor working just fine under hoary (and warty before it)-
-but just after christmas, the battery disappeared?! Now when I try to find the battery, I get the following:

dean@lappy:/etc/acpi$ ls -l /proc/acpi/battery/
total 0

I have tried moving the entire acpi directory to acpi.old, and completely uninstalling acpi, acpid, and acpi-
support in synaptic, and then reinstalling them. I still have the same problem.

There is another person on the ubuntu forums who has a similar problem on a Compaq Presario 1711CL. He installed
Ubuntu at around the same time that my battery disappeared (he added the backports, and upgraded with synaptic).

http://ubuntuforums.org/showthread.php?t=12317: http://ubuntuforums.org/showthread.php?t=12317

[Matthew Garrett (mjg59)]

Could you please attach the file /var/log/dmesg and the output of the dmesg command?

[Dean Jansen (list)]

Created an attachment (id=1125)
My dmesg output

[Dean Jansen (list)]

Created an attachment (id=1126)
/var/log/dmesg

[Matthew Garrett (mjg59)]

Ok, this is due to a bug in the current ACPI code. It's fixed in upstream ACPI -
I'll try to break out the patch.

[Dean Jansen (list)]

Oh, thanks :)

[Matthew Garrett (mjg59)]

Created an attachment (id=1127)
New ACPI-CA with fix for ACPI_AML_BUFFER_LIMIT issues

This is from the linux-acpi bitkeeper tree. It should fix the problem.

[Dean Jansen (list)]

Ooop--I'm not smart enough to actually use that patch. Do you mind pointing me to a tutorial, or giving a quick
lesson?

Thanks ;)

[Matthew Garrett (mjg59)]

It'll be included in the next kernel upload.

[Fabio Massimo Di Nitto (fabbione)]

Fixed in 2.6.10-11 upload that will be available in the archive within a few hours.

[Dean Jansen (list)]

Hey, just wanted to say thanks a ton! You guys (& gals?) do a superior job--it really shows thru in every aspect of
Ubuntu. Pass on the good word :)

[Dean Jansen (list)]

Sorry to bother again, but I'm not sure if I've got this working right or not (and I'm leaving to the UK tomorrow--so I
was hoping to get my battery working tonight).

I've got my system (Hoary) upgraded as much as possible.

In my /boot/grub/menu.lst, the following image is the first: /boot/initrd.img-2.6.10-2-686 (So i'm pretty sure this is
the one I'm booting to).

In synaptic, it says that I have linux-image-2.6.10-2-686 installed, and that it is version: 2.6.10-11, and that 2.6.10-
11 is the latest version.

I tried resetting twice, but still have nothing in my /proc/acpi/battery directory.

Did I do this correctly? Or have I just not waited long enough.

Thanks and sorry for bugging you again :)

[Matthew Garrett (mjg59)]

Hrm. Could we have /var/log/dmesg and the output of dmesg again, please?

[Dean Jansen (list)]

Created an attachment (id=1146)
My second dmesg output

[Dean Jansen (list)]

Created an attachment (id=1147)
2nd var/log/dmesg

[Dean Jansen (list)]

So I tried the 2.6.9-x kernel from grub, and the battery meter works perfectly. So it seems that it is the kernel for
sure.

Thanks again--I'll be gone for a few days, so if there is any more info you need, I'll be able to get back to you next
week :)

[Matthew Garrett (mjg59)]

Ok, thanks - it seems the patch hasn't fixed the problem it was supposed to.
I'll check into this.

[Matthew Garrett (mjg59)]

*** Bug 12245 has been marked as a duplicate of this bug. ***

[Matthew Garrett (mjg59)]

Ok, this seems to be yet another upstream bug. The next kernel upload should fix
it (with luck - I can't actually reproduce it on any of the hardware I have here)

[Matt Zimmerman (mdz)]

Is this related to bug #8659? In bug #12245, it was mentioned that kacpid was
spinning

[Matthew Garrett (mjg59)]

*** Bug 8659 has been marked as a duplicate of this bug. ***

[Dean Jansen (list)]

Okay, I upgraded to the latest kernel (2.6.10-14), and still am having no luck w/ the battery monitor. Any more info that
might help you? Just let me know :)

[Fabio Massimo Di Nitto (fabbione)]

Hi Dean,
 this is why the bug is not closed. I had to upload -14 with an important fix
for amd64
to get a new hoary test CD out today. The patch has been requeued for -15 that
will be
uploaded sometime during the next week (asap basically).

Fabio

[Fabio Massimo Di Nitto (fabbione)]

Fixed in 2.6.10-15 upload that will be available on the mirrors in a few hours.
Note: the kernel includes an ABI change that will temporary break binary drivers
like NVIDIA and ati.

[Dean Jansen (list)]

I just installed 2.6.10-3 (2.6.10-15) and the battery meter is still not working :(

Anything I can do to help, just let me know.

dean@lappy:~$ uname -r
2.6.10-3-686

--Dean

[Matthew Garrett (mjg59)]

Hmm. Could we have dmesg output again, please?

[Dean Jansen (list)]

Created an attachment (id=1295)
3rd dmesg output

This is the 3rd dmesg output.

[Dean Jansen (list)]

Created an attachment (id=1296)
3rd dmesg.log

This is the 3rd dmesg log.

[Matthew Garrett (mjg59)]

Could you attach the contents of /proc/acpi/dsdt please?

[Dean Jansen (list)]

Created an attachment (id=1297)
proc.acpi.dsdt

Okay, I'm not sure if I did this right--it appears to be binary or something??
I used the following command:
dean@lappy:~$ sudo cat /proc/acpi/dsdt > /home/dean/Desktop/proc.acpi.dsdt

[Matthew Garrett (mjg59)]

Ok, it turned out that this was (another) upstream bug. We have a patch now - it
should appear in either the next kernel upload or the one after that.

[Mario Vukelic (kreuzsakra)]

I just want to add the info that I have the same problem on my laptop, HP/Compaq
nc6000 as of today, running the newest default kernel,version 2.6.10-24

mario@phonic: ~ $ uname -r
2.6.10-4-686

It has always worked before I moved from debian sid to hoary.
I'm attaching the output of the dmesg command ("dmesg.output"), /var/log/dmesg,
and the output of cat /proc/acpi/dsdt > ~/proc.acpi.dsdt

[Mario Vukelic (kreuzsakra)]

Created an attachment (id=1462)
Output of dmesg comman

[Mario Vukelic (kreuzsakra)]

Created an attachment (id=1463)
cat /proc/acpi/dstd

[Mario Vukelic (kreuzsakra)]

Created an attachment (id=1464)
/var/log/dmesg

[Andrew Bennetts (spiv)]

(In reply to comment #31)
> I just want to add the info that I have the same problem on my laptop, HP/Compaq
> nc6000 as of today, running the newest default kernel,version 2.6.10-24
>
> mario@phonic: ~ $ uname -r
> 2.6.10-4-686

That's odd. I have an nc6000 as well, and it works fine for me:

andrew@trogdor /tmp $ uname -a
Linux trogdor 2.6.10-4-686 #1 Fri Feb 25 05:28:14 UTC 2005 i686 GNU/Linux
andrew@trogdor /tmp $ cat /proc/acpi/battery/C139/info
present: yes
design capacity: 3785 mAh
last full capacity: 3785 mAh
battery technology: rechargeable
design voltage: 11100 mV
design capacity warning: 379 mAh
design capacity low: 0 mAh
capacity granularity 1: 100 mAh
capacity granularity 2: 100 mAh
model number: Primary
serial number: 00001 2004/01/05
battery type: LIon
OEM info: Hewlett-Packard

I'll attach my dmesg log as well.

[Andrew Bennetts (spiv)]

Created an attachment (id=1466)
/var/log/dmesg from Andrew's working nc6000

[Dean Jansen (list)]

Woo Hoo! The latest kernel (as of 2/27/05) seems to have cured my battery meter problem :)

dean@lappy:~$ uname -r
2.6.10-4-686

If you need any log or dmesg information, just let me know.

[Matt Zimmerman (mdz)]

Thanks for testing

[Mario Vukelic (kreuzsakra)]

It does still not work for me. I have created a diff between Andrew's
/var/log/dmesg and mine, which I will attach. For me, nothing really points at
an error, but I am no expert. However, it is clear from the diff that while
Andrew's and my laptop are both labeled nc6000, they are quite different internally.
I'll set the bug status to Reopened

mario@phonic: ~ $ ls /proc/acpi
alarm dsdt embedded_controller event fadt fan info power_resource
processor sleep thermal_zone wakeup
mario@phonic:

mario@phonic: ~ $ dpkg -l linux-image-2.6.10-4-686
Desired=Unknown/Install/Remove/Purge/Hold
| Status=Not/Installed/Config-files/Unpacked/Failed-config/Half-installed
|/ Err?=(none)/Hold/Reinst-required/X=both-problems (Status,Err: uppercase=bad)
||/ Name Version Description
+++-=========================-=========================-==================================================================
ii linux-image-2.6.10-4-686 2.6.10-27 Linux kernel image for
version 2.6.10 on PPro/Celeron/PII/PIII/PIV
mario@phonic: ~ $

mario@phonic: ~ $ uname -a
Linux phonic 2.6.10-4-686 #1 Sat Mar 12 11:12:34 GMT 2005 i686 GNU/Linux
mario@phonic: ~ $

[Mario Vukelic (kreuzsakra)]

Created an attachment (id=1665)
diff -u between andrew's /var/log/dmesg and mine

On Andrew's machine it works, on mine not

[Matthew Garrett (mjg59)]

It looks like you have different BIOS versions, which may have something to do
with it.

[Andy Pahne (ap-der-die-das)]

I have the same problem after updating from Warty to Hoary (on Dell Latitude D505).

I tried various kernels, even compiled my own, but no battery status since my
update.

My dmesg output is attached as file.

[Andy Pahne (ap-der-die-das)]

Created an attachment (id=1674)
dmesg output Dell D505

[Andy Pahne (ap-der-die-das)]

Oh no.

I just forgot to load the battery module.

So please forget about #42 und #43

Sorry for that

[Mario Vukelic (kreuzsakra)]

Aargh! I too had the battery module not loaded (thanks, Andy), so forget comment
#39 too. Sorry. Works now that I entered it in /etc/modules. It seems strange
though that I did not have to load the module explicitly when I ran sid, only
since I "sidegraded" to Hoary

[Mario Vukelic (kreuzsakra)]

Shouldn't the battery module be loaded automatically by acpid? At least it says
so in /etc/default/acpid:

# Specify modules to load on acpid's startup
# MODULES may be uncommented to load "none", contain the string "all"
# to load all acpi related modules or simply a space seperated list
# of modules to be probed.
#MODULES="battery ac processor button fan thermal"
# using all allows us to drop extra modules in drivers/acpi and have them
# be autoloaded.
MODULES="all"

[Fabio Massimo Di Nitto (fabbione)]

we got a bunch of ACPI updates in the last kernels. Can you please test them?

Fabio

[Mario Vukelic (kreuzsakra)]

Using the up-to-date kernel (linux-image-2.6.10-5-686, version 2.6.10-33),
nothing changes for me. As long as I load the battery module explicitly in
/etc/modules, battery display continues to work fine. If I rely on acpid to load
it (as set in /etc/default/acpid), it does not get loaded and subsequently my
battery display does not work

[osantana (osantana)]

I've the same problem (and other problem with
ACPI/keyboard_not_working_until_press_sysrq_key).

My notebook is an Acer Travelmate 2200 (dmesg and /proc/acpi/dsdt attached).

[osantana (osantana)]

Created an attachment (id=2023)
Acer Travelmate 2200 dmesg output

uname -a -> Linux newton 2.6.10-5-386 #1 Tue Apr 5 12:12:40 UTC 2005 i686
GNU/Linux

[osantana (osantana)]

Created an attachment (id=2024)
Acer Travelmate 2200 /proc/acpi/dsdt output

uname -a -> Linux newton 2.6.10-5-386 #1 Tue Apr 5 12:12:40 UTC 2005 i686
GNU/Linux

[Mario Vukelic (kreuzsakra)]

It seems my /etc/init.d/acpid startup script fails with a syntax error and
therefore fails to load the acpi modules. I found modprobe errors in my
/var/log/boot at that point of the boot processs where acpid should say "[ok]".
See attachment for the relevant part of the bootlog.

I have tried to remove /etc/init.d/acpid and reinstalled acpid (1.0.4-1ubuntu4)
with dpkg -i --force-confmiss acpid, but the resulting new configfile shows the
same error upon reboot

[Mario Vukelic (kreuzsakra)]

Created an attachment (id=2201)
relevant part of bootlog

bootlog as per the previous comment

[Mario Vukelic (kreuzsakra)]

Sorry for spamming. I now am not so sure anymore that what I wrote before was
true. Contrary to what I wrote, somehow reinstalling acpid did solve the
problem. Will attach one bootlog before reinstalling, and boot log after
reinstalling

[Mario Vukelic (kreuzsakra)]

Created an attachment (id=2202)
New bootlog

[Mario Vukelic (kreuzsakra)]

Created an attachment (id=2203)
Old bootlog

[Mario Vukelic (kreuzsakra)]

I forgot another thing that I did when trying to hunt down the problem. Sorry, I
was not really concentrated :(

What I did was, I checked my version of module-init-tools, and it really was the
wrong version. I downgraded it (module-init-tools 3.2-pre1-1 ->
3.1-rel-2ubuntu2) and I think this was what made the modprobe error go away.
module-init-tools must have been left over when I sidegraded Sid -> Hoary (as I
said in comment #31, thank $deity).

[Matt Zimmerman (mdz)]

So is there still an open issue here, or not?

[Mario Vukelic (kreuzsakra)]

Not for me, but I dunno about comment #49

[Matthew Garrett (mjg59)]

Travelmate is likely to be a (currently unsupported) smart battery, the original
bug seems to have been down to some sort of acpid configuration issue.