Kernel Bug Tracker – Bug 3974
BATTERY state incorrect - okay in 2.4 - on Compaq Presario
Last modified: 2007-07-12 17:19:28 UTC
Since 2.6.9 kernel I have experienced problems with the battery state
on my Compaq Presario 2104EA laptop. I have noticed the problem when
running the ubuntu 220.127.116.11 kernel, which I believe is patched with a
recent acpi release, and the gentoo-dev-sources 2.6.9. I just
installed a 2.4.26 kernel, and have found that it works perfectly.
The problems are quite difficult to describe. Essentially when i boot with the
power adapter plugged in, acpid will give me the correct reading of
the battery. When there is a CHANGE of AC adapter state (i.e. i unplug
it), acpid does not give me the correct battery status, which remains
at the level which it was at. The same problem happens if I start the
computer without AC power, and plug it in: in this case the battery
status begins to go down correctly, but when AC is inserted, the level does not
go up as it should.
AFAICS the AC state is read correctly, and everything else (thermal etc) seems fine.
The results when I tested it with the 2.4 kernel have convinced me
that this is not a hardware problem.
I would be happy to provide any more information, and I hope i've been clear.
Many thanks in advance for your help.
Hardware Environment: Compaq Presario 2104EA
Steps to reproduce:
please attach the output from dmesg -s64000 and the output
from acpidmp, found in /usr/sbin, or in pmtools here:
Created attachment 4333 [details]
Output from acpidmp
Attached as request the output from acpidmp.
Created attachment 4334 [details]
Output of dmesg
As requested, output from dmesg -s 64000. However there is a problem: I'm
pretty sure this is not long enough for your needs, i.e. does not go all the
way back. I don't know how to make it longer. If you let me know, i'll do my
Created attachment 4357 [details]
Output from dmesg
Have recompiled my kernel with a bigger buffer size, so now here is the whole
of dmesg. Hope this helps, and that this problem can be tracked down.
I experience the same problems, first I suspected the BIOS is at fault, but a
well known commercial operative system does not show the same symptoms.
I am using a 18.104.22.168 kernel on a Pentium-M 745/ICH-4 Laptop (Littlebit Razor
I will try to attach the different logs...
Created attachment 5071 [details]
dmesg output, baier
Created attachment 5072 [details]
output of acpidmp, baier
Created attachment 5073 [details]
dmidecode output, baier
Created attachment 5074 [details]
Created attachment 5075 [details]
interrupts as in /proc/interrupts, baier
This possibly is the EC issue. Please try the latest -mm tree, which includes
the burst mode EC read/write patch.
I don't know about Mathew East, but at least in my case acpi info is much vivid
with kernel rc4-mm2. Hope the relevant patch gets rapidly integrated into
stable, such that swsusp2 patch applies cleanly again...
I still have this problem: no battery status. Recently I installed Windows to
ensure this wasn't a hardware problem: works fine. I've been without battery
status since 2.6.9 :/
I still have this problem with the current Ubuntu unstable kernel (2.6.12 plus
Ubuntu patches). I was thinking that perhaps it would help to try an unstable
kernel: could you point me to a kernel that I can try to see if this problem is
solved in your latest development tree?
I'm really keen to sort this problem out because being without battery status is
a huge blocker for me!
Could you try linux-2.6.13 with patch filed at bug 3851 (id: 5574)
Please try both burst_mode=0 and burst_mode=1
ok I've downloaded that kernel and have applied that patch, it's building now so
I will let you know. I'm not sure how to set burst_mode=0/1 however: I can
compile a kernel but I am not very knowledgeable about how it works!
Ok I booted that kernel but I did not have any battery status changes: I booted
up with AC power enabled, then took the plug out: no change in battery status.
boot kernel with option: ec_burst=1
Yes, that fixes it! It also fixes my inability to change the screen brightness,
which i noted in Ubuntu bugzilla here:
Thanks for your help! Will this patch be available in the stable kernel soon?
*** Bug 3516 has been marked as a duplicate of this bug. ***
*** Bug 5048 has been marked as a duplicate of this bug. ***
Does Linux-2.6.20 work without any cmdline parameters?
Please reopen this bug if it's still present with kernel 2.6.22.