New issue
Advanced search Search tips

Issue 884348 link

Starred by 1 user
Status: WontFix
Owner: ----
Closed: Sep 14
Cc:
omrilio@chromium.org
derat@chromium.org
sammiequon@chromium.org
mccanny@chromium.org
tbroch@chromium.org
afakhry@chromium.org
zork@chromium.org
adurbin@chromium.org
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 1
Type: Bug



Sign in to add a comment

Meowth stuck in clamshell and can't enter tablet mode

Project Member Reported by x...@chromium.org, Sep 14 
Chrome Version: 71.0.3543.0 
OS: Chrome

What steps will reproduce the problem?
It happens occasionally. There is no reliable repro step. Omri showed me the device that was in bad state. According to Omri, he attached a normal USB keyboard and attached the default keyboard to Mowoth at some point, and disconnected them one by one later, and the device stuck in clamshell mode forever. Attaching a keyboard or reboot doesn't make it enter tablet mode.

What is the expected result?
It should be able to enter/exit tablet mode.


I added some log to it and found it's because PowerManagerClient sent an incorrect TabletMode state to TabletModeController, see code here https://cs.chromium.org/chromium/src/ash/wm/tablet_mode/tablet_mode_controller.cc?rcl=1cd8b9f92fe68dd8cc064aac3270f86e0d890ee5&l=295. The |mode| variable that passed from PowerManagerClient equals to owerManagerClient::TabletMode::OFF after rebooting the device or restarting the user session. 

I then later attached/detached the default keyboard to it for several times and it suddenly can enter/exit tablet mode again. However, although the keyboard is attached, it still cannot be used.


I attached all the logs here, the log contains the problematic reboot is the latest one, and happens around 12:08pm or 12:09pm Sep 14. 

derat@, since you're familiar with PowerManagerClient, do you have any thoughts about why it might happen?
log.tar.gz
4.3 MB Download

Comment 1 by derat@chromium.org, Sep 14

Cc: tbroch@chromium.org
Components: OS>Firmware>EC
PowerManagerClient just passes on the D-Bus signals that it receives from powerd, and powerd just passes on the input events that it receives from the tablet mode switch. If you're seeing incorrect events, someone who works on the EC may need to take a look.

Comment 2 by x...@chromium.org, Sep 14 

Thanks derat@. Yes I can consistently repro this issue on the problematic device (just reboot the device, it's always there). But tried on other meowth device (same chrome version, different platform version though) and didn't repro it. 

Will install a newer image to the device and see if it goes away.

Comment 3 by x...@chromium.org, Sep 14 

For the problematic device, yes, it always receives the incorrect tablet mode value.

Comment 4 by x...@chromium.org, Sep 14

Cc: afakhry@chromium.org
Status: WontFix (was: Untriaged)
Installed the latest image (with platform version: 11065.0.0) and now I can get the correct tablet mode value from PowerManagerClient, and attaching/detaching the default keyboard can exit/enter tablet mode. However, even the keyboard is attached, it's still can't be used. But I think that's a different issue, maybe just a device specific problem.

And I also noticed that the problematic device that Omri gave to me contained a customized locally built image (platform: 11035.0 by afakhry@). So it seems to me either it regressed but got fixed later in the firmware, or it's the problem with the locally built image. Thus close this issue.

Comment 5 by tbroch@chromium.org, Sep 14 

Some other interesting issues w/ tablet interactions here as well: b/73133611 and b/112537098
►

Sign in to add a comment