hal not reading information about sysfs batteries correctly

I'm not sure if this is g-p-m or not, but I agree that this seems to be showing the wrong battery from before.
I suggest this because if I click on the systray battery icon and then on the name of the battery, the popup with information about charge levels and vendors and so-on contains about half as much information as it did previously.

I can confirm what Chris sees.
The only information which doesn't seem to be a default value is the vendor.
I attach a screenshot of gpm's battery information.

Hello,

I have something similar on my laptop. The g-p-m shows the battery as charging whenever I plug my wall adapter. But the battery status freezes there after then. The attachment shows two snapshots, one of them showing the mistaken battery reading (it shows 43% while actually the battery was near to be charged). The second attachment shows the correct reading. I forced the correct reading to appear by unplugging then immediately plugging the wall adapter.

Some info:
islam@islam-laptop:~$ cat /proc/version
Linux version 2.6.24-8-generic (buildd@yellow) (gcc version 4.2.3 (Ubuntu 4.2.3-1ubuntu2)) #1 SMP Thu Feb 14 20:13:27 UTC 2008
islam@islam-laptop:~$ dpkg -s gnome-power-manager | grep Version
Version: 2.21.1-0ubuntu1
islam@islam-laptop:~$ dpkg -s hal | grep Version
Version: 0.5.10-5ubuntu7

I want to add that the same thing happens at discharging. The reading never gets updated until some major event occurs such as Suspend or Shutdown. Even if the battery is near to empty, it still shows as 99% charged!!!

The correct reading, forced by triggering an event.

Also, notice how the Battery Charge Monitor Applet is affected by buggy readings as well. This suggests that the bug resides in HAL!

Some more info:
islam@islam-laptop:~$ cat /proc/acpi/battery/BAT0/info
present: yes
design capacity: 4000 mAh
last full capacity: 4000 mAh
battery technology: rechargeable
design voltage: 11100 mV
design capacity warning: 200 mAh
design capacity low: 140 mAh
capacity granularity 1: 60 mAh
capacity granularity 2: 3800 mAh
model number: GRAPE32
serial number: 38330
battery type: LION
OEM info: SANYO
islam@islam-laptop:~$ hal-find-by-capability --capability "battery"
/org/freedesktop/Hal/devices/computer_power_supply_0
islam@islam-laptop:~$ cat /proc/acpi/battery/BAT0/
alarm info state
islam@islam-laptop:~$ cat /proc/acpi/battery/BAT0/state
present: yes
capacity state: ok
charging state: charged
present rate: 0 mA
remaining capacity: 4000 mAh
present voltage: 12576 mV

(PS: the battery is really charged 100% now. The shown data is not a bug :)

I think this is a bug in hal.

lshal | grep -i bat
  [...]
  linux.sysfs_path = '/sys/devices/LNXSYSTM:00/device:00/PNP0A08:00/device:01/PNP0C09:00/PNP0C0A:00/power_supply/BAT0' (string)

hal reports different values than in linux.sysfs_path/* It seems that hal reloads values *only* on an acpi event, e.g., unplugging power. Values in the directory are getting changed while lshal reports still the same values unless I generate an acpi event using power cord (plug/unplug). Maybe other acpi events work as well.

Same happens on me. Notebook Vye S37B. If I issue a:
/etc/init.d/hal restart
I get the current reading but again, nothing gets updated.
/sys/devices/LNXST... shows the correct values.
/proc/acpi/... also shows the correct values.

gnome-power-manager is just using the information that it finds out from hal. It is actually hal that is not getting the sysfs information properly.

The related upstream Debian bug report points to a couple of patches at freedesktop.org that may resolve this bug.

I've posted test packages at http://www.trilug.org/~crimsun/hal/. Please leave feedback here.

I would like to test, but i have a 64 bit environment.
Can i compile from te hal_0.5.10.orig.tar.gz?

pbuilder is your friend.

Sorry for the spam, but I'll upload the package to my PPA.

Bruce, I tried your latest hal (ubuntu8), I built it for AMD64 architecture. It does refresh battery content almost correctly. Remaining capacity seems to be OK. But it seems that it does not report current rate at all:
lshal | grep -i bat
  battery.charge_level.current = 65080 (0xfe38) (int)
  battery.charge_level.last_full = 81980 (0x1403c) (int)
  battery.charge_level.percentage = 79 (0x4f) (int)
  battery.charge_level.rate = 0 (0x0) (int)
  battery.model = '42T4511' (string)
  battery.present = true (bool)
  battery.rechargeable.is_charging = false (bool)
  battery.rechargeable.is_discharging = true (bool)
  battery.remaining_time = 16270 (0x3f8e) (int)
  battery.reporting.current = 65080 (0xfe38) (int)
  battery.reporting.design = 84240 (0x14910) (int)
  battery.reporting.last_full = 81980 (0x1403c) (int)
  battery.reporting.rate = 0 (0x0) (int)
  battery.reporting.technology = 'Li-ion' (string)
  battery.reporting.unit = 'mWh' (string)
  battery.technology = 'lithium-ion' (string)
  battery.type = 'primary' (string)
  battery.vendor = 'SANYO' (string)
  battery.voltage.current = 12031 (0x2eff) (int)
  battery.voltage.design = 10800 (0x2a30) (int)
  battery.voltage.unit = 'mV' (string)
  info.capabilities = {'battery'} (string list)
  info.category = 'battery' (string)
  linux.sysfs_path = '/sys/devices/LNXSYSTM:00/device:00/PNP0A08:00/device:01/PNP0C09:00/PNP0C0A:00/power_supply/BAT0' (string)

Your patch has corrected the battery loading. Now i have 80Wh instead of 7.2Wh. I can also read a correct discharge time.
Thanks for this fix.

But gnome-power-manager still thinks that the laptop is working on battery. When it's not.
After i unplug/plug the power cable, the problem is solved.

On Sat, 2008-03-01 at 18:36 +0000, Lukas Hejtmanek wrote:
> Bruce, I tried your latest hal (ubuntu8), I built it for AMD64 architecture.

Sorry, but it's not mine. It's Daniel T Chen's work, I just uploaded it
to my PPA.

OK, there is a typo in sources. hald/linux/device.c line 3181, if(!hal...) should be if(hal..). After it, lshal reports correctly rate. But gnome-power-manager is still wrong about the rate, more precisely, g-p-m updates the rate only on power plug/unplug. Battery capacity (%) is updated on the fly correctly. Also, after desktop startup g-p-m reports battery is discharging even if running on power.

Here's the patch that fixes this bug:

http://gitweb.freedesktop.org/?p=hal.git;a=commit;h=85390b2b5b45253f64867895204d95bf2e1dc357

@Lucas: thanks, that's fixed in ~2.

@Khashayar: How does the addition of a battery serial number field resolve this issue?

The ~3 packages seem to work for me (amd64). g-p-m provides a (seemingly) accurate power history graph, the battery charge % is correctly reported, and it notices when I change from battery to AC and back again.

If anyone else on AMD64 wants to test these packages, they're available at http://cooperteam.net/hal

@Daniel: I'm sorry, I momentarily forgot that the summary of this bug had changed since I first filed it. When I filed it the summary was something like "incorrect profiling files created", and indeed the profiling files created in ~/.gnome2/gnome-power-manager are called 'profile-generic_id-generic..." and so on. With that patch (at least I _think_ it's that patch), the files are named 'profile-CP194245-XX....' and so on. This might be relevant when having two batteries, I don't know..

Thanks for those packages they fixed it.

I merged Daniel's branch, thank you! Will upload today.

This bug was fixed in the package hal - 0.5.10-5ubuntu8

---------------
hal (0.5.10-5ubuntu8) hardy; urgency=low

  [ Scott James Remnant ]
  * Add debian/patches/96_uinput_device_support.patch: This creates a HAL
    device for the uinput device, allowing us to set ACLs for it using
    PolicyKit/ConsoleKit.

  [ Daniel T Chen ]
  * Add debian/patches/97_fix_power_info_via_sysfs.patch: Fix battery status
    reading from /sys. Patches taken from upstream GIT head.
    (LP: #194052, #194719)

  [ Martin Pitt ]
  * Add debian/patches/02_sysfs_battery_serial.patch: Get battery serial
    number from sysfs. Patch taken fro upstream GIT head.

 -- Martin Pitt <email address hidden> Mon, 03 Mar 2008 15:02:40 +0100

Fix confirmed on my machine. Everything, battery, status-update, remaining-time, power works.
Thanks! Keep-up the good work!

Works for me too on a dell latitude d630. Although has anyone noticed that the g-p-m display doesn't change as fast as the power settings display changes in windows. In windows its almost instantaneous.

On Tue, Mar 4, 2008 at 6:28 AM, sharninder <email address hidden> wrote:

> Although has anyone noticed
> that the g-p-m display doesn't change as fast as the power settings
> display changes in windows.

Well, I've had a different experience with Windows' power management, and
that's certainly not Ubuntu's measure of quality. More to the point though,
it certainly does appear slower for me than it did on Gutsy or with the
previous Hardy packages. I just filed bug #198335 to that affect.

The windows comparison was just for reference. The gutsy comparison would have certainly been more relevant but I didn't have a gutsy notebook handy at the time. Anyway, thanks for the filing the bug.