Bug 201121
| Summary: | Full battery reported as 1% | ||
|---|---|---|---|
| Product: | Drivers | Reporter: | Trent Gamblin (trentfromcanmore) |
| Component: | Input Devices | Assignee: | drivers_input-devices |
| Status: | NEW --- | ||
| Severity: | normal | CC: | developer, luya, trent |
| Priority: | P1 | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Kernel Version: | 4.15 | Subsystem: | |
| Regression: | No | Bisected commit-id: | |
| Attachments: |
lsusb -vd output
Patch to ignore battery on this device |
||
Given by the name, it looks like the information for each device is incomplete Device: ELAN0732:00 04F3:264C Kernel: /dev/input/event5 Group: 7 Seat: seat0, default Size: 332x188mm Capabilities: touch Tap-to-click: n/a Tap-and-drag: n/a Tap drag lock: n/a Left-handed: n/a Nat.scrolling: n/a Middle emulation: n/a Calibration: identity matrix Scroll methods: none Click methods: none Disable-w-typing: n/a Accel profiles: n/a Rotation: n/a Device: ELAN0732:00 04F3:264C Pen Kernel: /dev/input/event6 Group: 7 Seat: seat0, default Size: 346x193mm Capabilities: tablet Tap-to-click: n/a Tap-and-drag: n/a Tap drag lock: n/a Left-handed: n/a Nat.scrolling: n/a Middle emulation: n/a Calibration: n/a Scroll methods: none Click methods: none Disable-w-typing: n/a Accel profiles: none Rotation: n/a Causing desktop environment like Gnome Shell complaining Nov 23 08:05:01 gnome-shell[1760]: Could not get tablet information for 'ELAN0732:00 04F3:264C Pen': (null) Elantech will need to provide the missing detailed information for both stylus and touchscreen. I wrote a patch which ignores the battery on this device, stopping the error message. Is this useful? Otherwise I can try to contact Elantech but I'm unsure if that will help. Created attachment 279653 [details]
Patch to ignore battery on this device
Here is the patch, I have tested it on 4.19 but not 4.20 but I could do so if it's useful.
(In reply to Trent Gamblin from comment #2) > I wrote a patch which ignores the battery on this device, stopping the error > message. Is this useful? Otherwise I can try to contact Elantech but I'm > unsure if that will help. That patch just removes the error message but also the battery status of the active stylus within the desktop environment. Essentially, we are still missing the detailed information like human readable name of the device i.e. HP Pen (or Elan Pen) from the config-table inside the kernel rather than LAN0732:00 04F3:264C Pen in this example. Looking at the comment from https://bugzilla.kernel.org/show_bug.cgi?id=198715#c93 , the suggestion will be to assign the product to ACPI and the component to Config-Table from this report. P.S: if possible, add the info of the device for linuxwacom (https://github.com/linuxwacom/libwacom/) using their template on https://github.com/linuxwacom/libwacom/blob/master/data/wacom.example I've been over this bug with the linuxwacom guys and they determined that the stylus reports the battery status incorrectly, which is why I wrote this patch. Having the correct stylus name would be nice but I'm not sure it will solve anything. Here is the thread with linuxwacom: https://github.com/linuxwacom/input-wacom/issues/37 Thanks for the thread. From details, the bug is specific to the Elan active stylus itself not the Dell as suggested by the patch because it also impacts HP touchscreen which I own. As suggested on #4, let's switch the Product to ACPI and the Component to Config-Table. I think there is a link with that dmesg extract from boot on my HP Envy laptop: [...snip...] [ 5.710951] ACPI Error: Field [D128] at bit offset/length 128/1024 exceeds size of target Buffer (160 bits) (20180810/dsopcode-201) [ 5.710966] ACPI Error: Method parse/execution failed \HWMC, AE_AML_BUFFER_LIMIT (20180810/psparse-516) [ 5.710981] ACPI Error: Method parse/execution failed \_SB.WMID.WMAA, AE_AML_BUFFER_LIMIT (20180810/psparse-516) [ 5.711040] ACPI Error: Field [D128] at bit offset/length 128/1024 exceeds size of target Buffer (160 bits) (20180810/dsopcode-201) [ 5.711044] ACPI Error: Method parse/execution failed \HWMC, AE_AML_BUFFER_LIMIT (20180810/psparse-516) [ 5.711051] ACPI Error: Method parse/execution failed \_SB.WMID.WMAA, AE_AML_BUFFER_LIMIT (20180810/psparse-516) [ 5.711109] ACPI Error: Field [D128] at bit offset/length 128/1024 exceeds size of target Buffer (160 bits) (20180810/dsopcode-201) [ 5.711114] ACPI Error: Method parse/execution failed \HWMC, AE_AML_BUFFER_LIMIT (20180810/psparse-516) [ 5.711121] ACPI Error: Method parse/execution failed \_SB.WMID.WMAA, AE_AML_BUFFER_LIMIT (20180810/psparse-516) [ 5.713899] ACPI Error: Field [D128] at bit offset/length 128/1024 exceeds size of target Buffer (160 bits) (20180810/dsopcode-201) [ 5.713909] ACPI Error: Method parse/execution failed \HWMC, AE_AML_BUFFER_LIMIT (20180810/psparse-516) [ 5.713919] ACPI Error: Method parse/execution failed \_SB.WMID.WMAA, AE_AML_BUFFER_LIMIT (20180810/psparse-516) [ 5.713979] ACPI Error: Field [D128] at bit offset/length 128/1024 exceeds size of target Buffer (160 bits) (20180810/dsopcode-201) [ 5.713984] ACPI Error: Method parse/execution failed \HWMC, AE_AML_BUFFER_LIMIT (20180810/psparse-516) [ 5.713991] ACPI Error: Method parse/execution failed \_SB.WMID.WMAA, AE_AML_BUFFER_LIMIT (20180810/psparse-516) If you have similar dmesg from above on your Dell from boot, then the issue is possibly the ACPI that needs a fix. I have no ACPI errors in my dmesg, might be unrelated. OK, that ACPI eror is out of the equation.
Run the upower -m to monitor the battery activity in order to find the device.
Then run upower -i /org/freedesktop/UPower/devices/keyboard_hid_0018o04F3o264Cx0001_battery as an example and post the result.
upower -i /org/freedesktop/UPower/devices/keyboard_hid_0018o04F3o264Cx0001_battery
native-path: hid-0018:04F3:264C.0001-battery
model: ELAN0732:00 04F3:264C
power supply: no
updated: Tue 27 Nov 2018 08:13:45 AM PST (16 seconds ago)
has history: yes
has statistics: yes
keyboard
present: yes
rechargeable: yes
state: discharging
warning-level: critical
percentage: 1%
icon-name: 'battery-caution-symbolic'
Just that information revealed upower detected the pen as keyboard_hid displaying the identical percentage of the stylus. It shows we are missing the acpi config-table to identity ELAN0732:00 04F3:xxxx Pen (where xxxx is the hexadecimal suffix name of the device, in my case 264C).
Based on that, we need to find the config-table listing the device.
I suggest changing the title like "Elan active stylus mistakenly reported 1% battery". What we want is the Elan style to correctly report the battery status, not hide it. It's worth officially noting that this does affect some other (possibly all?) Elan devices such as ELAN0732:00 04F3:2538 seen in some HP Envy x360s Right. The title should be like "ELAN0732:00 04F3:xxxx data displays wrong battery information for Elan active pen". After some investigations, the battery information is located on /sys/class/power_supply/ In my case, /sys/class/power_supply/hid-0018:04F3:264C.0001-battery/ containing the following: ls -l /sys/class/power_supply/hid-0018\:04F3\:264C.0001-battery/ total 0 -r--r--r--. 1 root root 4096 Nov 27 21:11 capacity lrwxrwxrwx. 1 root root 0 Nov 27 21:40 device -> ../../../0018:04F3:264C.0001 -r--r--r--. 1 root root 4096 Nov 27 21:11 model_name -r--r--r--. 1 root root 4096 Nov 27 21:40 online drwxr-xr-x. 2 root root 0 Nov 27 21:40 power lrwxrwxrwx. 1 root root 0 Nov 27 21:40 powers -> ../../../0018:04F3:264C.0001 -r--r--r--. 1 root root 4096 Nov 27 21:40 present -r--r--r--. 1 root root 4096 Nov 27 21:11 scope -r--r--r--. 1 root root 4096 Nov 27 21:11 status lrwxrwxrwx. 1 root root 0 Nov 27 21:11 subsystem -> ../../../../../../../../class/power_supply -r--r--r--. 1 root root 4096 Nov 27 21:11 type -rw-r--r--. 1 root root 4096 Nov 27 13:11 uevent cat /sys/class/power_supply/hid-0018\:04F3\:264C.0001-battery/uevent POWER_SUPPLY_NAME=hid-0018:04F3:264C.0001-battery POWER_SUPPLY_PRESENT=1 POWER_SUPPLY_ONLINE=1 POWER_SUPPLY_CAPACITY=1 POWER_SUPPLY_MODEL_NAME=ELAN0732:00 04F3:264C POWER_SUPPLY_STATUS=Discharging POWER_SUPPLY_SCOPE=Device Further details show the pen located on /sys/class/power_supply/hid-0018\:04F3\:264C.0001-battery/device/input/input15/ and the name cat /sys/class/power_supply/hid-0018\:04F3\:264C.0001-battery/device/input/input15/name ELAN0732:00 04F3:264C Pen So if I understand, the kernel recognized the Pen has a power supply but does not know how to properly get its battery status. Hopefully one of ACPI kernel developers can further investigate the issue. With no other available alternative, it looks like hiding the false 1% report is a safe option. I tested Windows 10 showing no battery status for the pen. Since Comment 11 indicates this affects other device ids as well, I’m not sure what the correct approach is. They could both be added to the patch but if there are several devices it could be a mess. |
Created attachment 278523 [details] lsusb -vd output With an ELAN touchscreen and active stylus, the stylus reports 1% battery at all times. I've used several batteries including 2 brand new ones. I noticed in the USB HID driver the possibility of ignoring the battery on devices by adding IDs to the source, perhaps it's a good idea with this device. To double check it I ordered a second stylus and they both do the same thing. When I use the stylus it pops up a notification saying the battery is low. On Windows, I don't get any notification nor does there seem to be any way to check the battery level, so I think it might not be a valid value at all. The troubleshooting manual from Dell for the stylus simply says if the stylus isn't working anymore, replace the battery. I'm using a Dell Inspiron 5379 laptop. lsusb -vd output is attached.