Bluetooth keyboard battery interpreted as laptop battery

[Øyvind Stegard]

My Logitech DiNovo Edge keyboard's internal battery status is available in:
/sys/class/power_supply/hid-00:07:61:76:C9:F1-battery

for the first time ever in Ubuntu 12.10. This would seem like a good thing, except that upower indicator reports it as a regular laptop/system battery, and even proposing to power off my desktop system if the keyboard battery happens to drain completely. That is unacceptable behaviour. I rectified it by manually editing settings-daemon power settings with dconf-editor, so the critical action was set to "nothing". (The regular power system settings GUI would NOT let me select this value, which is pretty strange.)

Here's what upower says:
$ upower -d
Device: /org/freedesktop/UPower/devices/battery_hid_00o07o61o76oC9oF1_battery
  native-path: /sys/devices/pci0000:00/0000:00:1d.2/usb8/8-1/8-1.1/8-1.1:1.0/bluetooth/hci0/hci0:11/0005:046D:B309.0002/power_supply/hid-00:07:61:76:C9:F1-battery
  model: Logitech diNovo Edge
  power supply: no
  updated: Wed Dec 5 12:15:38 2012 (27 seconds ago)
  has history: yes
  has statistics: yes
  battery
    present: yes
    rechargeable: yes
    state: discharging
    energy: 0 Wh
    energy-empty: 0 Wh
    energy-full: 0 Wh
    energy-full-design: 0 Wh
    energy-rate: 0 W
    percentage: 35%
    capacity: 100%

Daemon:
  daemon-version: 0.9.17
  can-suspend: yes
  can-hibernate no
  on-battery: no
  on-low-battery: no
  lid-is-closed: no
  lid-is-present: no
  is-docked: no

ProblemType: Bug
DistroRelease: Ubuntu 12.10
Package: upower 0.9.17-1build1
ProcVersionSignature: Ubuntu 3.5.0-19.30-generic 3.5.7
Uname: Linux 3.5.0-19-generic x86_64
NonfreeKernelModules: nvidia
ApportVersion: 2.6.1-0ubuntu8
Architecture: amd64
Date: Wed Dec 5 11:55:30 2012
InstallationDate: Installed on 2012-06-02 (185 days ago)
InstallationMedia: Ubuntu 12.04 LTS "Precise Pangolin" - Release amd64 (20120425)
MarkForUpload: True
SourcePackage: upower
UpgradeStatus: Upgraded to quantal on 2012-12-04 (0 days ago)

[Øyvind Stegard (oyvindstegard)]

While having apport collection info about upower, it made the kernel crash with nullpointer deref in kernel proc khidpd. Reboot failed, kernel was semi-hung afterwords. I suspect the crash might be triggered by reading/probing stuff about bluetooth or whatever. I have no time at the moment to report this as a separate bug, sorry. Traces appended below.

[ 1053.449584] ------------[ cut here ]------------
[ 1053.449593] WARNING: at /build/buildd/linux-3.5.0/fs/sysfs/dir.c:529 sysfs_add_one+0xca/0x100()
[ 1053.449595] Hardware name: P5E WS Pro
[ 1053.449596] sysfs: cannot create duplicate filename '/devices/pci0000:00/0000:00:1d.2/usb8/8-1/8-1.1/8-1.1:1.0/bluetooth/hci0/hci0:11'
[ 1053.449598] Modules linked in: joydev hidp parport_pc ppdev rfcomm bnep dm_crypt snd_hda_codec_hdmi kvm_intel kvm gpio_ich btusb microcode bluetooth nvidia(PO) snd_hda_codec_realtek snd_seq_midi serio_raw snd_rawmidi mceusb rc_core lpc_ich snd_seq_midi_event snd_seq asus_atk0110 snd_hda_intel snd_hda_codec snd_hwdep snd_pcm snd_seq_device snd_timer mac_hid shpchp snd coretemp soundcore snd_page_alloc x38_edac edac_core lp parport hid_generic firewire_ohci firewire_core usbhid crc_itu_t hid pata_marvell sky2 usb_storage
[ 1053.449640] Pid: 62, comm: kworker/u:5 Tainted: P O 3.5.0-19-generic #30-Ubuntu
[ 1053.449642] Call Trace:
[ 1053.449648] [<ffffffff81051c1f>] warn_slowpath_common+0x7f/0xc0
[ 1053.449651] [<ffffffff81051d16>] warn_slowpath_fmt+0x46/0x50
[ 1053.449655] [<ffffffff811f2baa>] sysfs_add_one+0xca/0x100
[ 1053.449658] [<ffffffff811f2dd8>] create_dir+0x78/0xd0
[ 1053.449661] [<ffffffff811f3166>] sysfs_create_dir+0x86/0xe0
[ 1053.449665] [<ffffffff8132b14c>] kobject_add_internal+0x9c/0x220
[ 1053.449668] [<ffffffff813342c1>] ? vsnprintf+0x461/0x600
[ 1053.449672] [<ffffffff8108e116>] ? update_group_power+0x26/0x1c0
[ 1053.449675] [<ffffffff8132b607>] kobject_add+0x67/0xc0
[ 1053.449680] [<ffffffff8141cc66>] device_add+0x106/0x6c0
[ 1053.449683] [<ffffffff8141b4b1>] ? dev_set_name+0x41/0x50
[ 1053.449694] [<ffffffffa0bb4465>] hci_conn_add_sysfs+0x45/0x80 [bluetooth]
[ 1053.449702] [<ffffffffa0ba7550>] hci_conn_complete_evt.isra.47+0xd0/0x390 [bluetooth]
[ 1053.449709] [<ffffffffa0baa1e0>] hci_event_packet+0x1720/0x2580 [bluetooth]
[ 1053.449713] [<ffffffff8103fa39>] ? default_spin_lock_flags+0x9/0x10
[ 1053.449719] [<ffffffffa0b9e84b>] hci_rx_work+0x1bb/0x310 [bluetooth]
[ 1053.449724] [<ffffffff816830ae>] ? _raw_spin_lock+0xe/0x20
[ 1053.449727] [<ffffffff8107086a>] process_one_work+0x12a/0x420
[ 1053.449733] [<ffffffffa0b9e690>] ? hci_send_frame+0x90/0x90 [bluetooth]
[ 1053.449736] [<ffffffff8107140e>] worker_thread+0x12e/0x2f0
[ 1053.449739] [<ffffffff810712e0>] ? manage_workers.isra.26+0x200/0x200
[ 1053.449742] [<ffffffff81075f03>] kthread+0x93/0xa0
[ 1053.449746] [<ffffffff8168c624>] kernel_thread_helper+0x4/0x10
[ 1053.449749] [<ffffffff81075e70>] ? kthread_freezable_should_stop+0x70/0x70
[ 1053.449752] [<ffffffff8168c620>] ? gs_change+0x13/0x13
[ 1053.449754] ---[ end trace 9e712b7e826b5280 ]---
[ 1053.449756] ------------[ cut here ]------------
[ 1053.449760] WARNING: at /build/buildd/linux-3.5.0/lib/kobject.c:19...

[Øyvind Stegard (oyvindstegard)]

Looks like the kernel crash I reported above is actually bug 888828.

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Arnaud Thevenet (aeromousse)]

Hello there,

I have this issue too.
Notice that with the Bluetooth USB dongle shipped with the keyboard, the "wrong" battery disappear and the keyboard is working fine.
So I am keeping the USB dongle plugged in, as Ubuntu do not work well with the laptop bluetooth device connected to the DiNovo Edge. With the dongle, I can let the bluetooth-wifi hard switch on, anyway. At least it does not interfere.

[Marcos Sánchez (arkanus)]

Please note problem also exists with other devices; i have a cheap bt mouse from DX and a Genius mouse/presenter both show the same behaviour descibed here.

[David Morris (dmorris68)]

Same problem with my Razer Pro|Click BT mouse on my Sony VAIO's built-in BT radio. This looks like a bug in the Unity battery applet, not in upowerd, but I could be wrong. upower-d displays the mouse battery with the proper device name and "no" to power supply (the laptop battery reports "yes" to power supply). But when I open the battery info app from Unity, it shows the mouse as a laptop battery. Not only that, but the "default" battery meter displayed in the Unity menu bar is for the mouse and NOT the laptop battery. So I have to click it to see the laptop's battery level.

[Ralph Meijer (ralphmeijer)]

I played around with several udev rules and inspected the upower source code. It appears that the kernel (in my case 3.7.0-7-generic) reports the device with the subsystem 'power_supply' with type 'Battery':

Dec 31 12:39:33 krijg kernel: [117903.350857] power_supply hid-00:07:61:F7:60:01-battery: driver failed to report `capacity' property: -5
Dec 31 12:39:38 krijg kernel: [117908.343434] power_supply hid-00:07:61:F7:60:01-battery: driver failed to report `capacity' property: -5
Dec 31 12:39:38 krijg kernel: [117908.364035] hid-generic 0005:046D:B309.0015: unknown main item tag 0x0
Dec 31 12:39:38 krijg kernel: [117908.404078] input: Logitech diNovo Edge as /devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.6/2-1.6:1.0/bluetooth/hci0/hci0:35/input32
Dec 31 12:39:38 krijg kernel: [117908.404346] hid-generic 0005:046D:B309.0015: input,hidraw0: BLUETOOTH HID v1.1b Mouse [Logitech diNovo Edge] on C4:85:08:47:23:C6

Reading the source code for ./linux/up-device-supply.c, it seems that upower doesn't expect CSR devices to be in the kernel, and therefore setting the kind as UP_DEVICE_KIND_BATTERY.

A udev rule like the following doesn't convince upower otherwise:

KERNEL=="hid-00:07:61:F7:60:01-battery", SUBSYSTEM=="power_supply", ENV{UPOWER_VENDOR}="Logitech, Inc.", ENV{UPOWER_PRODUCT}="diNovo Edge", ENV{UPOWER_BATTERY_TYPE}="keyboard"

[Christer (christer-madsen)]

For me, it affects me with the Lenovo Bluetooth Laser Mouse connected. There's no dongle, only connected to the internal bluetooth of my ThinkPad (it's a broadcom chip, but I don't think it's important). And sometimes, Ubuntu thinks that the battery is going dead, and shuts off!!! It's REALLY frustrating and unneccesary, because the laptop battery actually has 4-5 hours left of charge.... Also it displays the bluetooth mouse battery in the statusbar instead of the main battery, which is really strange and annoying.. I want Ubuntu to notify when BT mouse battery is getting low, but not turn off the computer by mistake :/ You know?

[Christer (christer-madsen)]

I'm having the EXACT same problem as David Morris three comments above!

[Ralph Meijer (ralphmeijer)]

@PnG: while this is marked as a duplicate of #1066208, I very much doubt it is, given its fix and my comments above (#8). The Logitech diNovo Edge is recognised as neither a keyboard nor a mouse in Raring's daily (12.10.6daily13.01.25-0ubuntu1) but as a laptop battery that is not a 'supply'. This device is primarily a keyboard with a small touch pad.

Besides that, upowerd doesn't seem to get the actual values right either. Fully charged it barely reaches 20%.

[Gregoire Gentil (gregoire-gentil)]

Here is an ugly patch that works for me (Logitech K810):

--- a/src/linux/up-device-supply.c 2013-06-08 15:07:46.161694077 -0700
+++ b/src/linux/up-device-supply.c 2013-06-08 15:07:05.665695138 -0700
@@ -867,6 +867,8 @@
   } else if (g_ascii_strcasecmp (device_type, "battery") == 0) {
    type = UP_DEVICE_KIND_BATTERY;
    supply->priv->is_power_supply = TRUE;
+gchar *device_name = up_device_supply_get_string (native_path, "model_name");
+if (device_name && ! g_ascii_strcasecmp (device_name, "Logitech K810")) goto out;
   } else if (g_ascii_strcasecmp (device_type, "USB") == 0) {

    /* use a heuristic to find the device type */

[Olivier Bilodeau (plaxx)]

Since this report is identified as a dupe of #1066208 and this issue is marked as fix released, I opened a new issue only regarding the keyboard battery being marked as a laptop battery: #1195840

[Olivier Bilodeau (plaxx)]

And of course I meant lp:1195840