lpm screen doesn't work when charging

It seems that it's an interrupt issue. The LPM mode is doesn't work for periods of time and it was more a coincidence that it happened every try with the charger plugged in the other night. Either close this or I could change it to LPM mode not working every time. I could supply dmesg logs, if that would help.

@Mohive: The only difference between "charging" and "not charging" that I can recall is: in late suspend the proximity sensor is put to a less power hungry mode and "charging" blocks the device from entering late suspend. So, if anything, it should work better with charger attached.

But compared to the "from-pocket", the so called "hover-over" triggering is a bit difficult to explain and getting it to trigger reliably takes some practice - which is why it is not enabled by default. Basically you need to start from "proximity not covered for > 3 seconds", then "cover and uncover the sensor in < 1.5 seconds". The covering needs to be long enough that sensor actually reacts. Something like 1. cover 2. say "hundred-one" 3. uncover.

If you still want to check if there is something wrong, or just for fun, running mce in verbose mode from root shell like this should show the lpm related triggering information:

systemctl stop mce
mce -T -qqqq -l*:tklock_lpmui_probe -l*:tklock_datapipe_proximity_sensor_cb -l*:mdy_display_state_enter

And back to normal operation via:

ctrl-c
systemctl start mce

I have made sure t cover the sensor more than just a short trigger, about 1-2 s. In dmesg I can see "entering sleepmode 0/1" (can't remember exact wording right now). The test sequences were more than 5 s apart. I'll try running mce in verbose mode tonight.