[vegeta] Phone app makes screen stay black during call (so you can't hang up)

I just tried to repro this on my vegeta (r5 also), and I can't

This appears to be the first step where things go wrong:

- Watch the screen turning black

On my handset the screen goes grey, and the dialler draws its UI. At no point during the initial setup of the call is the screen black/blank.

I tried accepting incoming calls in both the locked/screen off state and the screen on/unlocked state.

John, unfortunately I must confirm my original bug report.

I have a stock Ubuntu system as delivered by bq, and updated OTA as soon as delivered, now being the the latest update delivered by bq (or by you for bq devices?).

1.) The screen dims down and turns black immediately after initiating a call (either placing it myself or accepting an incoming call).

2.) a) When I do *not* bring the phone to my ear (= when the proximity sensor is not triggered, apparently) I can turn on the screen again by pressing the lock/unlock button at the top right side of my bq Aquaris E5 HD, and it *stays on*.

2.) b) To the contrary, when I *do* bring the phone to my ear for proceeding with the phone call the screen turns off immediately again after going alight for half a second when I press the lock/unlock button at the top right side of the phone.

--> 2.b.) is in fact the very troublesome situation! 2.a. is uncomfortable, but it's not actually a hurdle for stopping the phone call.

So, one difference may be the number of SIMs in use, and which ones the incoming calls arrive on.

in #1, I had one SIM installed, on the UK O2 network.

I have 2 SIMs in the phone (one active (slot #1), Switzerland, the other with the network out of reach (slot #2), Austrian network).

For the phone calls I'm using slot #1 at the moment.

another proximity, notifications, screen control symptom

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

@peter-bittner, when you say 'out of reach', do you think the SIM will be attempting to roam, or is roaming disabled in some way (sim, network, handset)?

Is this relevant information for this bug?

Roaming is disabled on the Austrian SIM. I just leave the SIM card in the phone when I'm not in Austria (or Italy), which is most of the time. (In fact, when I'm on Swiss territory I'd like to manually deactivate the second SIM. It rarely makes sense to have both cards active at the same time. I may file an enhancement request on this one day.)

So, the two cards are active, but only the primary is really "dialled in" on a network.

To be precise: Roaming is deactivated on the SIM, really the provider (Drei = 3 = Hutchinson 3G Austria).

I'm gessing the dialler/phone app is getting into some sort of confusion because of the states of the two sims. So, yes, I hope it is relevant.

The fact that the phone app comes up blank suggests the issue starts at the beginning of the call, not at the end :-)

u-s-c has no role in sim presence, however powerd is involved in both screen blanking and radio aiui

any chance recent fixes addressed this?

possibly
however, we have no vegatas among the team

@john-mc I'm going to transfer this to you as it sounds like possible sim related issue with powerd ?
just chat with me on irc if you think this is wrong or there's some way for my team to help.
we're familiar with the screen blanking, but not the radio interaction of powerd

Just wanted to confirm that this is still causing me pain daily on my bq Aquaris E5 (OTA-7). I can not at all (or hardly) use any phone call features (e.g. transfering to the loudspeaker to do hands-free talking is an acrobatic act, because the screen insists in turning black immediately again, e.g. I'm pressing the lock button on the right upper side of the phone).

Are there any plans on fixing this issue? This is causing me pain every single day. And it's causing me waste money too, because I'm having a hard time hanging up with every single call I take. Not nice.

@peter unfortunately we cannot reproduce it and have no other reports, so lets gather some info if we can

Does this happen on every call?
If you remove the second sim does it still occur?

please attach the file /var/log/syslog after this happens
same for /var/log/upstart/powerd.log

if possible during the call, type
powerd-cli list
and provide the output here.

I submitted Bug #1523308 which is probably a duplicate of this. Happened earlier today on my MX4 (one SIM card), first time I've noticed it. There's a syslog attached to that bug. I don't seem to have a powerd log though.

Bug #1532785 is probably another duplicate of this bug. Reported for two bq Aquaris 4.5 devices with OTA-8 (too).

Sorry, OTA 8.5 that is ...

I am thinking this is the same as bug #1531158

@Pat According to the reporter of bug #1532785 it's a different one. See his comment on the mailing list: https://lists.launchpad.net/ubuntu-phone/msg17611.html

Still an issue after upgrade to OTA 9.

Still a constant issue on OTA-9.1 (vegeta). I can't do a single call that would allow me to hang up like a normal person. People look at me when I'm fighting with my phone trying to hang up for almost a minute. No kidding.

Is this issue still monitored? Do you need additional information to reproduce to behavior? `powerd-cli list` maybe?

Output below is during incoming phone call, after having answered, moved phone to ear so that screen went black, removed from ear and screen stayed black:

phablet@ubuntu-phablet:~$ sudo powerd-cli list
System State Requests:
  Name: active, Owner: :1.81, State: 1

phablet@ubuntu-phablet:~$ tail -f /var/log/syslog
--> http://paste.ubuntu.com/15330960/

While I cannot yet reproduce it I have a germ of a theory. From the logs I see that both Peter and Mathew have autobrightness enabled. I also see that the powerd logic is explicitly setting the brightness to zero seemingly independent of the proximity reading. So perhaps we have a race between proximity, screen on and off, and autobrightness which is confusing powerd into a bad state.

@peter please try turning Adjust brightness automatically to off in the battery indicator and see if you still get the issue.

Mathew's log

Dec 7 09:41:58 ubuntu-phablet powerd[720]: set brightness 55
Dec 7 09:41:58 ubuntu-phablet powerd[720]: light_dev: setting brightness to 55
Dec 7 09:42:00 ubuntu-phablet powerd[720]: light_dev: setting brightness to 15
Dec 7 09:42:03 ubuntu-phablet powerd[720]: light_dev: setting brightness to 0

Peter's log
Mar 8 23:44:58 ubuntu-phablet powerd[757]: set brightness 12
Mar 8 23:44:58 ubuntu-phablet powerd[757]: light_dev: setting brightness to 12
Mar 8 23:44:58 ubuntu-phablet powerd[757]: turn_display_on(0)
Mar 8 23:44:58 ubuntu-phablet powerd[757]: light_dev: setting brightness to 0

Mar 8 23:44:59 ubuntu-phablet powerd[757]: light_dev: setting brightness to 12
Mar 8 23:45:00 ubuntu-phablet powerd[757]: we get signal from :1.40: DisplayPowerStateChange
Mar 8 23:45:00 ubuntu-phablet powerd[757]: Received DisplayPowerStateChange: state=1 flags=2
Mar 8 23:45:00 ubuntu-phablet powerd[757]: set brightness 12
Mar 8 23:45:01 ubuntu-phablet powerd[757]: light_dev: setting brightness to 0

Mar 8 23:45:02 ubuntu-phablet powerd[757]: turn_display_on(0)
Mar 8 23:45:02 ubuntu-phablet powerd[757]: turn_display_on(0)
Mar 8 23:45:02 ubuntu-phablet powerd[757]: light_dev: setting brightness to 0

Actually the only way to get turn_display_on(0) and set brightness to 0 is from the proximity sensor. Still a race perhaps...

another long shot,inquiries about using a screen protector at BQ support
http://www.mibqyyo.com/comunidad/discussion/61136/sensor-proximity-problems-in-a-call/p1
http://www.mibqyyo.com/comunidad/discussion/67626/proximity-sensor-doesnt-work-correctly/p1

Nice idea the one with the autobrightness. Unfortunately, it doesn't make a difference. That the problem is some kind of (almost regular) race condition could be very probable. I just had one single occasion when the screen, well, still stayed black after removing the phone from my ear but stayed alight after pressing the hardware lock button. So I could hang up "like a normal person". The phone call before that and after that, though, still had the screen turn back black immediately after pressing the hardware lock screen and going alight for half a second.

Same as before. But I'll keep my autobrightness turned off for now and observe if anything is different than before. Doesn't look very probable though.

I'm not using and have never used any screen protector. Instead I have my phone protected in a FREITAG phone sleeve [1], but clearly I need to take it out before answering or placing a phone call.

[1] http://www.freitag.ch/medias/sys_master/8821260648478/FREITAG-F28-iPhone-5-5S-Cover-Case-Gaine-Einzigartig-Speziell-Modern-Sleeve_big.jpg

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

I dont use screen protector but having issues with BQ 5HD , going black during call. Tried to reset the phone the other day. But the problem is still there. I will try to turn off autobrightness and see how that goes.

just a quick note, this bug is not being ignored, it seems this is E4.5/E5 specific. engineers only have E4, so we're trying to get an E5 into the hands of our engineers. the info and logs you've provided are great, unfortunately we're at the point we need some hands on debug.

Hi Peter,

So I just tried to replicate this in my E5 but haven't been able to.

How frequently do you encounter this issue? Like every 10 times you receive a call, how many of those do you encounter the issue reported?

@albaguirre I encounter this with every single call, no kidding. No matter whether I place that call or I receive it. Maybe once every month it happens that the screen stays alight (as described above), but not even then it's "normal" behavior as you see it on feature phones on competing smart phone platforms.

I'm sorry to say that. And I'm sure I'm not the only one. Just probably more noisy.

Hi Peter,

I have a suspicion your proximity sensor may not be working correctly. Could you install this ProximityCheck app in your phone: https://uappexplorer.com/app/proximitycheck.albaguirre and record a video of what you see?

Thanks Alberto, no looks like the sensor works as it should. Says "UNKNOWN" (grey) in the beginning, then "NEAR" (orange) when my hand or head is close, and "FAR" (green) when I'm further away than 10 cm or so.

@albaguirre It might really be the proximity sensor. But in a strange way. Kind of, "it depends on the mood of the device, and when in phone calls it's almost always in a bad mood."

I've noticed this today when my son wanted to find out what the ProximityCheck app is about. I tried to demo it to him, and the sensor switched to "NEAR" (orange) and didn't switch back to "FAR" anymore. I tried again turning the sensor off and on again in the app, and the erroneous behavior remained. I closed the app and tried again, and again it remained orange after moving it to the ear and taking it away again. In another try the app kind of flickered, mostly orange but green inbetween for fractions of a second even with nothing close the phone.

Most interesting though, after some time when trying another time the app finally showed normal behavior, with the proximity sensor telling "NEAR" with an object right in front of the phone and "FAR" with the same object further away from it. Very strange. So, is this maybe a hardware defect? In a series of devices of the same model?

@Peter, that would explain the behavior then.

"So, is this maybe a hardware defect? In a series of devices of the same model?"

That's a possibility; A look at the proximity device driver and/or sensor stack wouldn't hurt though...

I found a solution that worked on my phone BQ HD5 , Looks to me like a hardware issue. My lcd was sligthly losen in the top right corner next to the sensor. When i pressed in the screen in the corner i felt some movement, what i did was that i gently pulled to screen up a little and putted a little bit of glue inbetween the LCD screen and the frame just up in the corner. Now there is no movement when pressing the LCD and the proxmity sensors works fine again.

Hope this is useful.

Who is still affected by this bug? (I've been asked by bq about this. If you're still affected by the bug can you comment on this ticket, please?)

right now (7.6.2016) my wife faced this problem again on her BQ E4.5 OTA-11;

I'm now facing this issue constantly, with every single call. When I run ProximityCheck app (see above) it shows me that the proximity sensor is activated when I put my hand or head near the phone, and never deactivated again when removing it or doing whatever interaction with the phone.

@peter given its that reproducible now lets get a clean syslog while you are using the check app
and once you reboot does the symptom persist?

@Pat I've rebooted the phone (3 times), activated SSH access (to read the syslog), opened the ProximityCheck app and played with it (press "Start", move the hand near the sensor, move the phone to my head as if I were taking a phone call, etc.). As explained earlier, the app turns orange to indicate the sensor in "NEAR" state, and never turns back to "UNKNOWN" or "FAR" again.

This is the syslog from that activity: http://paste.ubuntu.com/19166245/

From Peter's log
Jul 12 13:32:24 ubuntu-phablet kernel: [ 1222.960067][lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c
Jul 12 13:32:24 ubuntu-phablet kernel: [ 1243.132054][mt-i2c]ERROR,533: WRRD transfer length is not right. trans_len=2, tans_num=2, trans_auxlen=0
Jul 12 13:32:24 ubuntu-phablet kernel: [ 1243.132074][mt-i2c]ERROR,612: _i2c_get_transfer_len fail,return_value=-22
Jul 12 13:32:24 ubuntu-phablet kernel: [ 1243.132089][ALS/PS] tmd2772_init_client_for_cali_call 659 : init dev: -22
Jul 12 13:32:24 ubuntu-phablet kernel: [ 1243.207867][LHJ_PS][tmd2772_ps_calibrate_call][L2619]prox_mean=1023
Jul 12 13:32:24 ubuntu-phablet kernel: [ 1243.207881]prox_mean>800
Jul 12 13:32:25 ubuntu-phablet kernel: [ 1243.732204][LHJ_PS][TMD2772_get_ps_value][L1320]ps=902,[700,800]val=0

...

Jul 12 13:33:05 ubuntu-phablet kernel: [ 1245.126340][lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c[lcm_esd_check]buffer[0]=0x1c
Jul 12 13:33:05 ubuntu-phablet kernel: [ 1284.289510][mt-i2c]ERROR,533: WRRD transfer length is not right. trans_len=2, tans_num=2, trans_auxlen=0
Jul 12 13:33:05 ubuntu-phablet kernel: [ 1284.289539][mt-i2c]ERROR,612: _i2c_get_transfer_len fail,return_value=-22
Jul 12 13:33:05 ubuntu-phablet kernel: [ 1284.289553][ALS/PS] tmd2772_init_client_for_cali_call 659 : init dev: -22
Jul 12 13:33:06 ubuntu-phablet kernel: [ 1284.364629][LHJ_PS][tmd2772_ps_calibrate_call][L2619]prox_mean=864
Jul 12 13:33:06 ubuntu-phablet kernel: [ 1284.364643]prox_mean>800

What I see on proposed

Jul 12 03:56:07 ubuntu-phablet repowerd[1049]: DefaultStateMachine: handle_user_activity_extending_power_state
Jul 12 03:56:07 ubuntu-phablet repowerd[1049]: UnityScreenService: dbus_requestSysState(:1.41,usensord,1)
Jul 12 03:56:07 ubuntu-phablet repowerd[1049]: LibsuspendSuspendControl: disallow_suspend(UnityScreenService)
Jul 12 03:56:07 ubuntu-phablet repowerd[1049]: UnityScreenService: dbus_requestSysState(:1.41,usensord,1) => 8
Jul 12 03:56:07 ubuntu-phablet kernel: [ 901.550636][LHJ_PS][tmd2772_ps_calibrate_call][L2636]prox_mean=52
Jul 12 03:56:07 ubuntu-phablet kernel: [ 901.550649]prox_mean<800
Jul 12 03:56:07 ubuntu-phablet kernel: [ 901.556892][LHJ_PS][TMD2772_get_ps_value][L1320]ps=53,[92,92]val=1
Jul 12 03:56:07 ubuntu-phablet kernel: [ 901.563338][LHJ_PS][TMD2772_get_ps_value][L1320]ps=53,[92,92]val=1
Jul 12 03:56:07 ubuntu-phablet kernel: [ 901.577180][LHJ_PS][TMD2772_get_ps_value][L...

@alf any ideas? I realize not a powerd issue

I wanted to add some information that there is a bug very very similar to this which has been shared upstream. There is a fix from our upstream that is currently in rc-proposed from upstream that will be rolled out as part of the OTA12 update that is meant to help address this issue.

@lpotter, can you take a look at the logs as discussed. Is this a HW issue?

In #46, there is a difference of an ALS entry. (Pat's does not have one, while the one containing
an error has one) but the error is contained in the ALS/PS, while LHJ_PS looks normal. These functions that look like they pertain to calibration.

As in the log from #45 seems to point to ALS/PS reading from mt-i2c, it could originate in the hardware, but the software doesn't recover. Whether that software is the driver, android sensors or powerd or something else would require digging through the code. My guess is the (als) driver doesn't handle this well.

It is interesting that this seems to happen on krillin as well, but apparently not all devices. I'm not familiar enough with these to devices to know if they use the same als/ps chip or share driver code.

FWIW the E4.5 and E5 have the same chip and driver as the design differs only in the screen and camera.

Looking into the code, nothing jumps out at me as being out of the ordinary there. Someone with a device would need to look into this further.

I am wondering how widespread this bug may be? Does it effect only a few phones (as in a bad batch of sensors) or is it more widespread?

The bug affects 11 people atm so hard to say

Just a gentle reminder: I'm still struggling with my bq Aquaris E5 with every single phone call I receive or make. I look like an idiot when I fight with my phone for half a minute each time I need to hang up. It's a real pain. - No improvement also with OTA-13.

@peter sorry we just do not have any way to reproduce it or debug as of yet

also per comments 37 and 38 its possibly the hardware. The proximity management in the repowerd daemon is quite different and improved now so if 13 didn't yield a resolution I am not sure how to address it.

@Pat Thanks for the feedback. I'll try to install Android on the phone, then we'll see whether this is actually a hardware failure.

@Pat @Victor Alright, so it's a hardware issue: I've installed Android and there are the same issues with the screen turning black. Then I'm done with this bug report.

@peter thanks for the follow up, hope you can get a new one or get it fixed

If anyone else still experiences this please comment and change back to new

@Pat I don't intend to make anyone nervous here, but my phone behaves normally now under Android (screen turns black when I move my head towards the phone during a call, turns back on again when I move the phone away from my head). I'd have to observe this for longer maybe, and then reinstall Ubuntu Touch and see if the issue comes back. Not sure what's happening. Anyway, for now I don't care.

I have the same problem as Peter with my krillin (E4.5) and OTA-13. I even experienced it while receiving an incoming call, which it hard to answer. But that happened only once and only after the last update.

A little update on my previous comment. It happened again while receiving a call. I still can't get a grasp on the conditions necessary for this to happen, though.

Hi, same problem here. Calling out and receiving a call, in both cases the screen turns black when connection is made. Pushing on/off button constantly (with quick intervals) turns on screen for very short moment, allowing to disconnect, after many tries.

ProximityCheck, always shows NEAR when starting the application.

Aquaris E5 HD Ubuntu Edition - OTA13

Using the ProximityCheck tool I noticed that sometimes the proximity sensor gets stuck in "NEAR" state, but sometimes it works normally. I still haven't been able to find any logical pattern in this behavior, though.

Interestingly, I'm seeing similar albeit not the same issues with phone calls on Android too (now that I'm using Android on the phone for more than a month). The phone screen stays black at the end of a phone call. Even worse, I can wake it up even with pressing hardware buttons. I need to wait until the other caller hangs up.

This could be related to an Android driver issue for the proximity sensor. Especially, if the same one is used on Ubuntu Touch. Or a mixture of hardware flaw and driver unreliability. As usual, just guessing.

I had this problem on my E5 since a couple of weeks back.

I read brummle's comment #39 about the lcd coming loose and remembered that my lcd has come loose before.

So I handled the phone and the screen clicked in in the top right corner and the problem seems to have gone.