817939 - [FM Radio] Turn on the FM Radio and then sleep the phone for a while, the UI will inconsistent to interaction

Status: RESOLVED FIXED

(Core :: DOM: Device Interfaces, defect, P2)

Description

[John Shih]

## Environment :
Unagi phone, build 2012-12-03
Build info: 
gaia master : 5d150b2b10472478e8841730abd9ae9597595206
mozilla-beta : 1e56f66d7ee9

## Repro :
1. Launch the FM Radio with headphone plugged-in
2. Check can listen to the Radio through the headphone
3. Sleep the phone for about 2~ minutes
4. Wake the phone and unlock the lockscreen
5. Back to FM Radio, hit the next/previous button to switch the channel

## Expected:
* After searching for a few seconds, radio will switch to next/previous channel and displays the frequency on screen

## Actual:
* Hear the radio is change to different channel
* The frequency displayed on screen won't change
* The searching animation is keep active
* Hit the next/previous buttons several times will see the frequency change inconsistently (Please the the video in attachment)

Comment 1

[John Shih]

Due to the file is too large and the problem in youtube (can't upload video), I'll upload later

Comment 2

[John Shih]

http://www.youtube.com/watch?v=Wze8KCaSQBQ&feature=youtu.be
The video

Comment 3

[Steve Chung [:steveck]]

Hi John,
I could not reproduce it with 12/09 unagi build.
Could you please verify with this version? thanks.

Comment 4

[StevenLee[:slee]]

Hi Pin,
Can you look this bug?

Comment 5

[Pin Zhang [:pzhang] (inactive)]

(In reply to StevenLee from comment #4)
> Hi Pin,
> Can you look this bug?

I can't reproduce this issue either.

Comment 6

[John Shih]

yeah, seems like this has been fixed by some modification
closing

2012-12-11 Unagi
gaia master : d6aaf482785f51eed6d9e38080d140a078e902c2
mozilla-beta : 32929d7d0a1f

Comment 7

[John Shih]

since can't find the core reason for this bug, and it still can be reproduced sometimes so reopen the bug
I'll spend time to find more information about this bug

Comment 8

[Steve Chung [:steveck]]

Hi John,
Since I still could not reproduce this issue, I added some debug message in FM DOM api to verify the root cause that is seek success event missed from beck-end or error while Service.DOMRequest.fireSuccess. I can flash the debug version gecko for you, thanks.

Comment 9

[Al Tsai [:atsai]]

Try it w/ 12-16 Engineer Nightly build on unagi.
Cannot reproduce it.

Comment 10

[John Shih]

I found it can not be reproduced after the device just flashed, but after reboot (several times) it can still happen

Comment 11

[Steve Chung [:steveck]]

Thanks for your help, Al.

Hi John,
You mean it may still happen if reboot several times. Could it always be reproduced after that?

Comment 12

[Al Tsai [:atsai]]

I can reproduce it after reboot my unagi

Comment 13

[Dietrich Ayala (:dietrich)]

Driver retriage: Steve, can you work directly with Al to figure out how to reproduce this? If it's reproducible and still bad, please re-nominate for blocking.

Comment 14

[Al Tsai [:atsai]]

Currently, the only way to reproduce it is to reboot the machine and do the steps in description. It happens sometimes. We'll keep following the issues and try to give the steps for always reproduce it.

Comment 16

[Vivien Nicolas (:vingtetun) (:21) - (NOT reading bugmails, needinfo? please)]

(In reply to Dietrich Ayala (:dietrich) from comment #13)
> Driver retriage: Steve, can you work directly with Al to figure out how to
> reproduce this? If it's reproducible and still bad, please re-nominate for
> blocking.

I can reproduce and the app seems broken when it happens. Nominating again.

Comment 17

[Steve Chung [:steveck]]

(In reply to Vivien Nicolas (:vingtetun) from comment #16)
> 
> I can reproduce and the app seems broken when it happens. Nominating again.

Hi Vivien, May I know which build are you testing now?
I'm testing the latest b2g18-unagi_betatest(1/6), but it still works for me. Is the reproduce step as simple as Al said(just reboot the device) or need any other action to make it reproducible? And how the reproduce rate? Thanks.

Comment 18

[Pin Zhang [:pzhang] (inactive)]

I followed the steps to reproduce the problem, and the searching indicator is always shown, but the frequency is changed consistently.

Comment 19

[Pin Zhang [:pzhang] (inactive)]

(In reply to Pin Zhang [:pzhang] from comment #18)
> I followed the steps to reproduce the problem, and the searching indicator
> is always shown, but the frequency is changed consistently.
I can't reproduce it now, @@
Hi Vivien, is it always producible on your device? And what's the commit numbers of both the Gecko and Gaia?

Comment 20

[Vivien Nicolas (:vingtetun) (:21) - (NOT reading bugmails, needinfo? please)]

Gecko changeset: http://hg.mozilla.org/releases/mozilla-b2g18/rev/d7dbdec9352e
Gaia changeset: https://github.com/mozilla-b2g/gaia/commit/194062b4e8b17a38247d5417094ab9743bd862b2

Comment 21

[Pin Zhang [:pzhang] (inactive)]

If I opened console to see log messages, i.e `adb logcat`, the problem can not be reproduced; and if I unplugged the USB cable, it's reproducible on Vivien's device.
It should be the time issue that we encountered before, and I think I need Steven's help.

Comment 22

[Pin Zhang [:pzhang] (inactive)]

(In reply to Pin Zhang [:pzhang] from comment #21)
> If I opened console to see log messages, i.e `adb logcat`, the problem can
> not be reproduced; and if I unplugged the USB cable, it's reproducible on
> Vivien's device.
> It should be the time issue that we encountered before, and I think I need
> Steven's help.

Hi Steven, if the problem happens, the seek complete event is not fired, that's why the searching indicator is always shown, can you take a look at thia?

Comment 23

[Vivien Nicolas (:vingtetun) (:21) - (NOT reading bugmails, needinfo? please)]

Sounds like a driver bug.

Comment 24

[Pin Zhang [:pzhang] (inactive)]

(In reply to Pin Zhang [:pzhang] from comment #21)
> If I opened console to see log messages, i.e `adb logcat`, the problem can
> not be reproduced; and if I unplugged the USB cable, it's reproducible on
> Vivien's device.
> It should be the time issue that we encountered before, and I think I need
> Steven's help.
Actually, if I just plug the phone with the USB cable , no need to run `adb logcat` in the console, the problem won't happen.

Comment 25

[Pin Zhang [:pzhang] (inactive)]

The version of the Vivien's Device is :
   Linux version 3.0.8-perf (songsy@ubuntu-cdr) (gcc version 4.4.3 (GCC) ) #1 PREEMPT Tue Oct 16 07:34:52 PDT 2012

The kernel version is not the latest, and the problem can be always reproduced, and it also happens sometimes on the latest version, still don't know the pattern to reproduce it on the latest version.

Comment 26

[John Shih]

I have noticed this for a long time, both happen on the new and old kernel
the only difference is that how hard it can be reproduced

Comment 27

[Joe Cheng [:jcheng] (please needinfo)]

talked to Thinker to take over the bug

Comment 28

[StevenLee[:slee]]

I downgraded to kernel-update-3 and found that the STR is 
1. launch fm radio with headset inserted
2. seek to a station
3. unplug usb cable
4. press power key to enter sleep mode and wait several seconds
5. press power key again and go to FM radio app
6. seek

When the problem happens, I plug usb cable and check my log. I found the driver does not callback data anymore. I think it should be the driver's fault.

Comment 29

[Thinker Li [:sinker]]

Attachment: restart ioctl with EINTR for FMRadio

Comment 30

[Thinker Li [:sinker]]

Attachment: restart ioctl with EINTR for FMRadio

--
Attachment #699208 [details] [diff] - Attachment is obsolete: true

Comment 31

[Pin Zhang [:pzhang] (inactive)]

(In reply to Thinker Li [:sinker] from comment #30)
> Created attachment 699214 [details] [diff] [review]
> restart ioctl with EINTR for FMRadio
> 
> --
> Attachment #699208 [details] [diff] [diff] - Attachment is obsolete: true

I tested the patch, it works fine now.

Comment 32

[Al Tsai [:atsai]]

tested with the patch also, it works fine.

Comment 33

[Justin Lebar (not reading bugmail)]

Comment on attachment 699214 [details] [diff] [review]
restart ioctl with EINTR for FMRadio

Oh, I understand what this does!  :)

Comment 34

[Thinker Li [:sinker]]

(In reply to Justin Lebar [:jlebar] from comment #33)
> Comment on attachment 699214 [details] [diff] [review]
> restart ioctl with EINTR for FMRadio
> 
> Oh, I understand what this does!  :)

cool!! And thank you!

Comment 35

[Thinker Li [:sinker]]

Attachment: restart ioctl with EINTR for FMRadio, v2

--
Attachment #699214 [details] [diff] - Attachment is obsolete: true

Comment 36

[Kan-Ru Chen [:kanru] (UTC+9)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/b89d480e49f0
https://hg.mozilla.org/releases/mozilla-aurora/rev/94030dd8405e
https://hg.mozilla.org/releases/mozilla-b2g18/rev/e21583a00313

Comment 37

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/b89d480e49f0