867912 - [Buri][Audio] Configuring silent mode by long-pressing power button is not persistent after device reboot

Status: VERIFIED FIXED

(Firefox OS Graveyard :: Gaia::System, defect, P1)

Description

[sync-1]

AU_LINUX_GECKO_ICS_STRAWBERRY_V1.01.00.01.019.085
 Firefox os  v1.0.1
 Mozilla build ID:20130422230201
 
 +++ This bug was initially created as a clone of Bug #448319 +++
 
 DEFECT DESCRIPTION:
 [Audio]Silence incoming call function abnormal.
 
  REPRODUCING PROCEDURES:
 1.Set ringer&notification volume to a level(not 0),long press power key to choose "silence incoming calls",status bar will display silence(vibrate)icon.
 But the volume of ringer&notification will not change to 0.->ko1
 2.Restart device,silence(vibrate)icon will disappear from status bar,and if has an incoming call will has ringer.->ko2
 
  EXPECTED BEHAVIOUR:
 ko1.Check with SW,silence incoming call is to disable the ringer not clear volume,but it is not a good behavior that status bar is not synchronized with settings.
 ko2.Keep silence incoming call settings.
 
  ASSOCIATE SPECIFICATION:
 
  TEST PLAN REFERENCE:
 
  TOOLS AND PLATFORMS USED:
 
  USER IMPACT:
 
  REPRODUCING RATE:
 3/3
 
  For FT PR, Please list reference mobile's behavior:
 
 ++++++++++ end of initial bug #448319 description ++++++++++
 
 
 
 CONTACT INFO (Name,Phone number):
 
  DEFECT DESCRIPTION:
 
  REPRODUCING PROCEDURES:
 
  EXPECTED BEHAVIOUR:
 
  ASSOCIATE SPECIFICATION:
 
  TEST PLAN REFERENCE:
 
  TOOLS AND PLATFORMS USED:
 
  USER IMPACT:
 
  REPRODUCING RATE:
 
  For FT PR, Please list reference mobile's behavior:

Comment 1

[Daniel Coloma:dcoloma]

The first scenario is working fine for me, if I call after step 1, there is no sound.

Second scenario is still wrong. Changing the subject to clarify what the bug is.

Comment 2

[Andrew Overholt [:overholt]]

baku/Marco, is this persistence something the audio manager handles?

Comment 3

[(inactive after 6/18) Alive Kuo [:alive]]

1. works for me because 'silent incoming call' isn't implemented as 'set volume to 0'.
I admit we need to refine this function in device menu.

2. looks wired...gaia bug maybe.

Comment 4

[Marco Chen [:mchen]]

Discussed with Alive and he will take care this issue on Gaia side.

Comment 5

[Andrea Marchesini [:baku]]

> 1. works for me because 'silent incoming call' isn't implemented as 'set
> volume to 0'.

How silence mode is implemented? Probably the easier way is to ask AudioChannelService to mute all the elements.

Comment 6

[(inactive after 6/18) Alive Kuo [:alive]]

(In reply to Andrew Overholt [:overholt] from comment #2)
> baku/Marco, is this persistence something the audio manager handles?

Andrew, my proposed change is to do bug 868314: change the behavior of 'silent incoming calls' to 'silent mode' (implemented in bug 819842).

Comment 7

[(inactive after 6/18) Alive Kuo [:alive]]

(In reply to Andrea Marchesini (:baku) from comment #5)
> > 1. works for me because 'silent incoming call' isn't implemented as 'set
> > volume to 0'.
> 
> How silence mode is implemented? Probably the easier way is to ask
> AudioChannelService to mute all the elements.

We just set two channels volume to 0: notification and content as UX suggested.

Comment 8

[(inactive after 6/18) Alive Kuo [:alive]]

(In reply to Alive Kuo [:alive] from comment #7)
> (In reply to Andrea Marchesini (:baku) from comment #5)
> 
> We just set two channels volume to 0: notification and content as UX
> suggested.

Baku, I suggest that leave UX relevant logic in gaia to avoid rapid spec change revealed in gecko, how do you think? I know it isn't a perfect solution from gaia side.

To be clear, in my mind there should be some API for gaia to mute different channel(without setting the volume to 0) but I wonder this is an efficient and bug-free way.

Since now the trigger to silent mode is pressing volume key down to 0, it makes sense to "just set volume to 0".

Comment 9

[Andrea Marchesini [:baku]]

Probably it would be cleaner to use the AudioChannelService. In this way the memory consumption will be much less: the media element has a wakelock enabled when it is playing audio, but if it's muted by the AudioChannelService, this wakelock is disabled. The patch about this is landed yesterday: Bug 859824

Comment 10

[(inactive after 6/18) Alive Kuo [:alive]]

(In reply to Andrea Marchesini (:baku) from comment #9)
> Probably it would be cleaner to use the AudioChannelService. In this way the
> memory consumption will be much less: the media element has a wakelock
> enabled when it is playing audio, but if it's muted by the
> AudioChannelService, this wakelock is disabled. The patch about this is
> landed yesterday: Bug 859824

It's not clear to me what to do. The bug 859824 looks to me that you're going to release some resource when music is "paused". Do you want to actively "pause" every audio when the user enables silent mode? And prevent every audio to start playing?

Comment 11

[Andrea Marchesini [:baku]]

> It's not clear to me what to do. The bug 859824 looks to me that you're
> going to release some resource when music is "paused". Do you want to
> actively "pause" every audio when the user enables silent mode? And prevent
> every audio to start playing?

Not only. That patch releases the wakelock for 2 reasons:

1. when the media element is stopped/paused
2. when the audiochannelservice muted the element (search for SetCanPlay());

We care about this second reason for silence mode.
Otherwise:

1. the user opens the music app. the music app plays audio - audiochanneltype => content.
2. the device is in silence mode, so no audio is actually played
3. the device is locked
4. the device doesn't go in sleep mode because of the wake lock.

This is the reason why I think it's better to use the AudioChannelService to silence any (or some) media element. Then, maybe we can fix this issue for 1.1.

Comment 12

[(inactive after 6/18) Alive Kuo [:alive]]

(In reply to Andrea Marchesini (:baku) from comment #11)
> 2. when the audiochannelservice muted the element (search for SetCanPlay());
> 
> We care about this second reason for silence mode.
> Otherwise:
> 
> 1. the user opens the music app. the music app plays audio -
> audiochanneltype => content.
> 2. the device is in silence mode, so no audio is actually played
> 3. the device is locked
> 4. the device doesn't go in sleep mode because of the wake lock.
> 
> This is the reason why I think it's better to use the AudioChannelService to
> silence any (or some) media element. Then, maybe we can fix this issue for
> 1.1.

From above I guess you want to "avoid redundant power consumption when audio is playing at volume 0", but my problem is 1) how gaia tell you to mute it? 2) couldn't we "mute the channel"  when you see channel volume is 0 now?

Comment 13

[Marco Chen [:mchen]]

(In reply to Andrea Marchesini (:baku) from comment #11)
About use AudioChannelService to pause media element for silent mode, I think that is not what we defined for silent mode.

From user's point of view, I would expect my music app is on playing even on silent mode. Because silent indicates to mute audio only not pause audio.

Comment 14

[(inactive after 6/18) Alive Kuo [:alive]]

I am going to provide a gaia side solution at first since :baku is busy now.

Comment 15

[(inactive after 6/18) Alive Kuo [:alive]]

Casey, if we're going to modify the silent mode function behavior on the device menu,
what's the string of "silent" and the opposite one?
Currently it's "silent incoming calls" and "ring incoming calls".

Comment 16

[(inactive after 6/18) Alive Kuo [:alive]]

BTW, The root cause of why the settings reverses after rebooting is:
https://github.com/mozilla-b2g/gaia/blob/master/apps/system/js/sleep_menu.js#L40

The handler here changes |ring.enabled| once the notification channel volume is changed.
Sigh, a really incorrect implementation.

Comment 17

[(inactive after 6/18) Alive Kuo [:alive]]

Two ways to fix this bug in my mind now:
A)
1. Remove |ring.enable| settings key from the apps reference it.
2. Use 'audio.volume.notification'+'vibration.enabled' to be the key to trigger silent mode.
3. For statusbar/lockscreen: 
   show mute icon when: audio.volume.notification == 0 && vibration.enabled == false
   show vibration icon when: audio.volume.notification == 0 && vibration.enabled == true
   show nothing when: audio.volume.notification != 0 && vibration.enabled == false
4. For dialer/keyboard:
   Doesn't call play() when audio.volume.notification == 0

B)
1. Rename |ring.enabled| with |silent.enabled|
2. For sound manager:
   Save current notification channel volume, let audio.volume.notification == 0 when silent.enabled == true
   Restore saved notification channel volume
3. For statusbar/lockscreen/dialer/keyboard: just replace |ring.enabled| with |!silent.enabled|

I will try B) first.

Comment 18

[Stephany Wilkes]

Flagging Rob in the hope that we have an existing pattern that speaks to this one.

Comment 19

[(inactive after 6/18) Alive Kuo [:alive]]

WIP:

https://github.com/alivedise/gaia/commit/432cdadf38f8ec5b57ef10c6471c34d6c84ce20a

This one is pretty complex...:)

Todo:
* solve the problem silent mode caused by two channels.

Comment 20

[(inactive after 6/18) Alive Kuo [:alive]]

Attachment: Refined multi-channel silent mode

Patch v1:

1. Remove `ring.enabled` because it conflicts with `audio.volume.notification`
2. Use `audio.volume.notification` to indicate silent mode
3. Multi channel silent mode:
   (1) Enter silent mode from Device menu: Silent notification channel, cache it.
   (2) Leave silent mode from Device menu: Restore cached notification channel. If there's no cached one, restore to 1.
   (3) Set audio.volume.notification->0 from Settings app: 
       (A) With vibration on: Vibration icon shows, not silent mode.
       (B) With vibration off: Enter notification silent mode, no cached volume.
   (4) Enter silent mode from volume rockers:
       (A) volumedown to 0 first time: vibration on, not silent mode.
       (B) volumedown to 0 second time and afterwards: Both Notification and Content silent mode. Cache content volume, no cached notification volume.

4. Fix a bug introduced from 819842 content silent mode.

IMO there's no silver bullet to multiple channel silent mode...this is already the bugfree way I figure out to maintain the silent state from multi input(Settings, hardware key, virtual key).

Comment 21

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

set milestone to partner bug

Comment 22

[Lawrence Mandel [:lmandel] (use needinfo)]

Daniel - If I understand this bug correctly, there is a simple workaround (set phone back to vibrate after reboot). Comment 19 also states that the WIP patch is pretty complex and, therefore, will likely add significant risk of regression. Does this bug really need to block the release?

Comment 23

[(inactive after 6/18) Alive Kuo [:alive]]

master
https://github.com/mozilla-b2g/gaia/commit/8b2e95522652dedc68052152ea6e49faba1cc013

Drivers: it's up to drivers to decide this is tef+ or not, but the code has already been granted so I'm closing this now.
IMO the old function is just a failed implementation.

Comment 24

[(inactive after 6/18) Alive Kuo [:alive]]

I am filing a bug for unit test on sound manager in https://bugzilla.mozilla.org/show_bug.cgi?id=870654

Comment 25

[John Ford [:jhford] CET/CEST Berlin Time]

This bug was partially uplifted.

Uplifted 8b2e95522652dedc68052152ea6e49faba1cc013 to:
v1-train: 1ef0c50111dbab687634b468d6914e781c600349

Commit 8b2e95522652dedc68052152ea6e49faba1cc013 didn't uplift to branch v1.0.1

Comment 26

[(inactive after 6/18) Alive Kuo [:alive]]

We're waiting for bug 819842 to be tef+.
If that doesn't happen, need some more additional work here to do patch on v1.0.1
because the patch here is highly relevant to bug 819842's patch now.

The addition work is:
In device menu, remove all 'ring.enabled', use 'audio.volume.notification'.
In sound manager, add enterSilentMode/leaveSilentMode for notification channel only.

Comment 27

[Daniel Coloma:dcoloma]

Bug 819842 is now tef+ and can be uplifted to v1.0.1, John, can you try to uplift 819842 one and later on this to v1.0.1?

Comment 28

[Kai-Chih Hu [:khu]]

Talked with Alive and he will uplift it to v1.0.1.

Comment 29

[(inactive after 6/18) Alive Kuo [:alive]]

v1.0.1
https://github.com/mozilla-b2g/gaia/commit/aa90119cd24fe54a3bb1228a7323cc12bd021db6

Comment 30

[xiaokang.chen]

(In reply to Alive Kuo [:alive] from comment #29)
> v1.0.1
> https://github.com/mozilla-b2g/gaia/commit/
> aa90119cd24fe54a3bb1228a7323cc12bd021db6

Hi Alive, 
I have add somecomment in https://github.com/mozilla-b2g/gaia/pull/9638/files, could you have a look?

Comment 31

[xiaokang.chen]

(In reply to Alive Kuo [:alive] from comment #29)
> v1.0.1
> https://github.com/mozilla-b2g/gaia/commit/
> aa90119cd24fe54a3bb1228a7323cc12bd021db6

Hi Alive, 
I have add somecomment in https://github.com/mozilla-b2g/gaia/pull/9638/files, could you have a look?

Comment 32

[xiaokang.chen]

(In reply to Alive Kuo [:alive] from comment #29)
> v1.0.1
> https://github.com/mozilla-b2g/gaia/commit/
> aa90119cd24fe54a3bb1228a7323cc12bd021db6

Hi Alive, 
I have add somecomment in https://github.com/mozilla-b2g/gaia/pull/9638/files, could you have a look?

Comment 33

[(inactive after 6/18) Alive Kuo [:alive]]

https://github.com/mozilla-b2g/gaia/commit/19f77d82fb60b4d2999202a59fa84a2062bf3ea3
Fix this.

(In reply to xiaokang.chen from comment #32)
> (In reply to Alive Kuo [:alive] from comment #29)
> > v1.0.1
> > https://github.com/mozilla-b2g/gaia/commit/
> > aa90119cd24fe54a3bb1228a7323cc12bd021db6
> 
> Hi Alive, 
> I have add somecomment in
> https://github.com/mozilla-b2g/gaia/pull/9638/files, could you have a look?

Comment 34

[mdaniloff]

This issue no longer reproduces on both V1.1 and V1.0.1. User is able to disable the ringer and there is no call ringer volume heard during an incoming call.

Leo Mozz build V1.1: 20130530070208
Gecko: http://hg.mozilla.org/releases/mozilla-b2g18/rev/09ac1fd2959c
Gaia: 1cca9324d4444ad28c6fa99875e17abf7e8230be
Version 18.0

Buri Mozz V1.0.1 build: 20130530070213
Gecko: http://hg.mozilla.org/releases/mozilla-b2g18_v1_0_1/rev/11b55d3ada71
Gaia: ac293ce59acc3bede083fad1b973794fa8bf0253
Version 18.0

Comment 35

[sync-1]

Comment from Mozilla:This issue no longer reproduces on both V1.1 and V1.0.1. User is able to disable the ringer and there is no call ringer volume heard during an incoming call.
 
 Leo Mozz build V1.1: 20130530070208
 Gecko: http://hg.mozilla.org/releases/mozilla-b2g18/rev/09ac1fd2959c
 Gaia: 1cca9324d4444ad28c6fa99875e17abf7e8230be
 Version 18.0
 
 Buri Mozz V1.0.1 build: 20130530070213
 Gecko: http://hg.mozilla.org/releases/mozilla-b2g18_v1_0_1/rev/11b55d3ada71
 Gaia: ac293ce59acc3bede083fad1b973794fa8bf0253
 Version 18.0