[5.0.0.71] Overlay display blanks screen (e.g. with reminder or calls) only flickers for call on button press, proximity sensor always "covered"

Bug Reports
1

REPRODUCIBILITY: always
OS VERSION: 5.0.0.72
HARDWARE: Xperia X10 II
UI LANGUAGE: German
REGRESSION: yes

DESCRIPTION:

Once an overlay is shown like when a call is active or a reminder for an appointment triggers, the screen stays blank. And it does not accept swipes or presses.

With a reminder, this can be remedied by pressing the button three times.

During a call there is no way out I've found, so it is critical.

PRECONDITIONS:

This started happening after the upgrade to 5.0.0.71.

STEPS TO REPRODUCE:

  1. initiate a call

  2. (alternatively) create an appointment with a reminder and wait until it fires

EXPECTED RESULT:

The overlay with the reminder or call information should be shown and accept swipes or presses.

ACTUAL RESULT:

Blank screen.

MODIFICATIONS:

ADDITIONAL INFORMATION:

The situation stays the same after a reboot.

Running CSD, the proximity sensor tests fails. It shows the proximity sensors always as ā€œcoveredā€ (though there is nothing covering the facing side of the phone).

2

Seems like sensorfwd is stuck. It happens too often on my C2 as well. But it did on my Xperias as well.

3

Seems like sensorfwd is stuck

Anything that can be done to workaround or resolve the issue?

4

On my C2 it seems to help to run CSD and do cover and light intensity test.

If that doesnā€™t help then deamon restart in CLI or some other way.

5

On my C2 it seems to help to run CSD and do cover and light intensity test.

Running Test tool CSD | Sailfish OS Documentation :

  • Light sensor: Pass.

  • Proximity sensor fails, it is always considered ā€œCoveredā€.

Behavior still the same. I wonder though, as before the software update to 5.0.0.71, it never was a problem.

If that doesnā€™t help then daemon restart in CLI or some other way.

As rebooting does not help, Iā€™ve been trying the cli.

after 
  systemctl restart sensorfwd.service

[..]
CGroup: /system.slice/sensorfwd.service
           ā””ā”€25126 /usr/sbin/sensorfwd -c=/etc/sensorfw/primaryuse.conf --syste
md --log-level=warning --no-magnetometer-bg-calibration

Nov 06 15:11:31 Xperia10II-DualSIM systemd[1]: Starting Sensor daemon for senso
r framework...
Nov 06 15:11:31 Xperia10II-DualSIM systemd[1]: Started Sensor daemon for sensor
 framework.
Nov 06 15:11:32 Xperia10II-DualSIM sensorfwd[25126]: se
nsorfw: Hybris sensor manager initialized
Nov 06 15:11:32 Xperia10II-DualSIM sensorfwd[25126]: se
nsorfw: Plugin not available: "wakeupsensor"
Nov 06 15:11:32 Xperia10II-DualSIM sensorfwd[25126]: se
nsorfw: SensorManagerError:  "plugin not available"
Nov 06 15:11:32 Xperia10II-DualSIM sensorfwd[25126]: se
nsorfw: minLimit: 0 0
Nov 06 15:11:32 Xperia10II-DualSIM sensorfwd[25126]: se
nsorfw: maxLimit: 2147483647 2147483647

 cat /etc/sensorfw/primaryuse.conf
[plugins]
accelerometeradaptor = hybrisaccelerometeradaptor
alsadaptor = hybrisalsadaptor
proximityadaptor = hybrisproximityadaptor
magnetometeradaptor = hybrismagnetometeradaptor
gyroscopeadaptor = hybrisgyroscopeadaptor
orientationadaptor = hybrisorientationadaptor
 
[magnetometer]
scale_coefficient = 1
transformation_matrix = "-1,0,0,0,1,0,0,0,1"
needs_calibration = 0
 
[accelerometer]
transformation_matrix = "1,0,0,0,1,0,0,0,1"

But still the same problem. (Thanks for the hints so far.)

6

I had that issue long ago. Donā€™t remember exactly how it turned out. IIRC I restarted deamon when that happened (sensor stuck). Sometimes triple power button push helped temporarily also.

No other ideas than this:

Do you have any dust if you use display protection?

Thereā€™s also a slight possibility of sensor or connector failure.

7

Do you have any dust if you use display protection?

(no display protection and also no visible dust)

Thereā€™s also a slight possibility of sensor or connector failure.

Thought about this as well, but why would that materialise right after the upgrade?

8

Is there a way to downgrade to 5.0.0.68 to see if the behaviour is in the software?

9

Not directly. You can flash older version. If you have it or can get it.

EDIT: And I really doubt that this would help anyway. As that bug was common for some back in a day.

10

And I really doubt that this would help anyway. As that bug was common for some back in a day.

Because the sensor ā€œfailureā€ was within a day of the upgrade, Iā€™d expect that there is a good chance that the upgrade contributed or even is fully responsible for it.

11

Disabled the proximity sensor for now, so that I can at least do phone calls and will see how this goes.

disabling the proximity sensor

(as devel-su)

vi /etc/sensorfw/sensord.conf.d/20-sensors-default.conf
changed the following line

--- out/20-sensors-default.conf
+++ sensord.conf.d/20-sensors-default.conf
@@ -16 +16 @@
-proximitysensor=Feature_ProximitySensor
+proximitysensor=False

and systemctl restart sensorfwd.service

@filip.k thanks a lot for your help, it means a lot to have someone responding. Your hint towards sensorfw and CSD was valuable.

4 Likes
12

Just noticed exactly the same problem, I couldnā€™t exit out of a call. It appears to be ā€˜causedā€™ by having enabled the LPM screen. More precisely, if you have set

mcetool --set-ps-on-demand=disabled

which is necessary for the LPM screen to work, the proximity sensor breaks. But if you change the setting to

mcetool --set-ps-on-demand=enabled

the proximity sensor works as expected again. Strange.