987545 - When headset is connected, uplink voice is picked up by handset mic instead of headset mic

Status: RESOLVED FIXED

(Firefox OS Graveyard :: AudioChannel, defect, P1)

Description

[Pooja]

Steps to Reproduce:
1. Connect wired headset to the target.
2. Make an MO voice call
3. Speak something into headset mic.
4. Observe that whatever you speak is not picked up by headset mic, instead it is being picked up by handset mic.

Expected Result:
During a voice call with headset inserted to target, what ever is spoken in to headset mic should be picked up and routed through headset mic

Actual Result: 
During a voice call with headset inserted to target, what ever is spoken in to headset mic is not picked up by headset mic, instead it is being picked up by handset mic.

Comment 1

[Preeti Raghunath(:Preeti)]

David,

Please help review and assign an engineer

Comment 2

[David Scravaglieri [:scravag]]

Assigned to Etienne to take a look.

Comment 3

[Matthew Vaughan]

This issue does *not* reproduce for us on the Buri or Helix devices using the 03/26/14 1.4 build. 

However, this issue does seem to reproduce on the Open C device using its 03/26/14 1.3 build. The Open C has presented some issues that we don't see on older devices though, including 2 out of the 3 headsets we used not being detected, so this may not be as reliable a test to go on.

Pooja, would you be able to tell us what device you tested this on and maybe some environmental variables? That would help us greatly. Thank you!

Comment 4

[Inder]

Matthew: this was tested on QRD8x26. Please work with Preeti to get hold of the device if you don't have it already.

Comment 5

[Preeti Raghunath(:Preeti)]

Lets please investigate on the Open C and put your findings down. We can check if the issue is on our end or not.

Comment 6

[Matthew Vaughan]

Attachment: LogCat - Open C using the Sony Ericsson headset during a call

(In reply to Preeti Raghunath(:Preeti) from comment #5)
> Lets please investigate on the Open C and put your findings down. We can
> check if the issue is on our end or not.

I tested 5 different headsets with microphones on the Open C. The results for each headset are:

Alcatel headset = not detected
HTC headset     = not detected
Sony Ericsson MH410 = only the headset earbuds work; handset microphone is used for speaking.
ZTE HMZ8-C4-OMTP    = only the headset earbuds work; handset microphone is used for speaking.
Unknown brand       = only the headset earbuds work; handset microphone is used for speaking.

Device: Open C v1.3 MOZ RIL
BuildID: 20140326040052
Gaia: be872843fa702e25ba113a93ce84fd8b265d87fb
Gecko: db0c8822aeafe9038815bdadf6d4312927cef250
Version: 28.0
Firmware Version: P821A10-ENG_20140321

Please note, all the listed headsets are detected and work correctly on a Buri.

Additionally, I have attached a logcat for the Open C using the Sony Ericsson headset.

Comment 7

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

Rik, do you want to take this before Etienne comes back from PTO? Thanks

Comment 8

[Jason Smith [:jsmith]]

If this is only happening on Jellybean & not ICS, then this isn't a regression.

Comment 9

[Jason Smith [:jsmith]]

If this is present on 1.3 though, then why are we blocking 1.4 specifically here? Moving to 1.3? to discuss this against the first applicable release hit here.

Comment 10

[Etienne Segonzac (:etienne)]

Looks like a lower level issue.

Comment 11

[Preeti Raghunath(:Preeti)]

Based on QA findings per Comment #6, this is easily reproducible and hence blocking the same.

Comment 12

[Jason Smith [:jsmith]]

Eric - Can you find an assignee for this?

Comment 13

[Eric Chou [:ericchou] [:echou]]

Star will look into this.

Comment 14

[Star Cheng [:scheng]]

(In reply to Matthew Vaughan from comment #6)
> Created attachment 8397429 [details]
> LogCat - Open C using the Sony Ericsson headset during a call
> 
> (In reply to Preeti Raghunath(:Preeti) from comment #5)
> > Lets please investigate on the Open C and put your findings down. We can
> > check if the issue is on our end or not.
> 
> I tested 5 different headsets with microphones on the Open C. The results
> for each headset are:
> 
> Alcatel headset = not detected
> HTC headset     = not detected
> Sony Ericsson MH410 = only the headset earbuds work; handset microphone is
> used for speaking.
> ZTE HMZ8-C4-OMTP    = only the headset earbuds work; handset microphone is
> used for speaking.
> Unknown brand       = only the headset earbuds work; handset microphone is
> used for speaking.
> 
> Device: Open C v1.3 MOZ RIL
> BuildID: 20140326040052
> Gaia: be872843fa702e25ba113a93ce84fd8b265d87fb
> Gecko: db0c8822aeafe9038815bdadf6d4312927cef250
> Version: 28.0
> Firmware Version: P821A10-ENG_20140321
> 
> Please note, all the listed headsets are detected and work correctly on a
> Buri.
> 
> Additionally, I have attached a logcat for the Open C using the Sony
> Ericsson headset.


Hi, Matthew

Could you help to dump audio policy log by using executing "adb shell dumpsys > headset.dump" command? And check the value of "Output devices" in AudioPolicyManager Dump section. (the format looks like below:)


AudioPolicyManager Dump: 0xb7f49eb8
 Primary Output: 2
 A2DP device address: 
 SCO device address: 
 USB audio ALSA 
 Output devices: 0000000b   ->    (check this value)
 Input devices: 00000044
 Phone state: 0
 Force use for communications 0
 Force use for media 0
 Force use for record 0
 Force use for dock 0
 Force use for system 0

Comment 15

[Star Cheng [:scheng]]

(In reply to Jason Smith [:jsmith] from comment #8)
> If this is only happening on Jellybean & not ICS, then this isn't a
> regression.

The latest SW version can NOT duplicate it in Nexus4 device. (Nexus4 is JB version)

I tested 4 different headsets with microphones:
Apple headset (for iPhone 4S): works normally
HTC headset (for HTC New One): works normally
Nexus-S headset (for Nexus-S) : works normally
Alcatel headset (for hamachi) : works normally

Comment 16

[Star Cheng [:scheng]]

(In reply to Star Cheng [:scheng] from comment #15)
> (In reply to Jason Smith [:jsmith] from comment #8)
> > If this is only happening on Jellybean & not ICS, then this isn't a
> > regression.
> 
> The latest SW version can NOT duplicate it in Nexus4 device. (Nexus4 is JB
> version)
> 
> I tested 4 different headsets with microphones:
> Apple headset (for iPhone 4S): works normally
> HTC headset (for HTC New One): works normally
> Nexus-S headset (for Nexus-S) : works normally
> Alcatel headset (for hamachi) : works normally


It can *not* duplicate in flame device by using above 4 headsets. The SW version as below:

Gaia      601df8b81e7c98514c55ee84ef2592b22b9c3816
Gecko     a986fdc9ea4ac16a2037b4face64dde3f3b57fda
BuildID   20140327230556
Version   28.0

Comment 17

[Matthew Vaughan]

Attachment: Audio policy dumps zip folder

(In reply to Star Cheng [:scheng] from comment #14)
> Hi, Matthew
> 
> Could you help to dump audio policy log by using executing "adb shell
> dumpsys > headset.dump" command? And check the value of "Output devices" in
> AudioPolicyManager Dump section. (the format looks like below:)
> 
> 
> AudioPolicyManager Dump: 0xb7f49eb8
>  Primary Output: 2
>  A2DP device address: 
>  SCO device address: 
>  USB audio ALSA 
>  Output devices: 0000000b   ->    (check this value)
>  Input devices: 00000044
>  Phone state: 0
>  Force use for communications 0
>  Force use for media 0
>  Force use for record 0
>  Force use for dock 0
>  Force use for system 0

I got an audio policy dump for all 4 of the headsets we have (the Sony headset I used in earlier testing is missing unfortunately). 2 of them were detected and the other 2 were not. The "Output devices" number for the 2 detected headsets matched what you put above (0000000b), but their microphone did not work and instead I had to use the handset's microphone to speak.

Alcatel headset  = not detected
HTC headset      = not detected
ZTE HMZ8-C4-OMTP = only the headset earbuds work; handset microphone is used for speaking.
Unknown brand    = only the headset earbuds work; handset microphone is used for speaking.

Comment 18

[Preeti Raghunath(:Preeti)]

Star

Are you able to repro this on Open C? Seems like QA is. Please provide the next steps here.

Comment 19

[Jason Smith [:jsmith]]

No longer going to block on Mozilla-specific 1.3 blockers unless it hits a special exception list, so moving to 1.4? for triage.

Comment 20

[Star Cheng [:scheng]]

(In reply to Preeti Raghunath(:Preeti) from comment #18)
> Star
> 
> Are you able to repro this on Open C? Seems like QA is. Please provide the
> next steps here.

Yes, I can duplicate it in OpenC device. There are different log between Flame and OpenC.

[OpenC]
D/ALSADevice(  284): mInChannels:1
D/ALSADevice(  284): route: devices 0x8 in mode 2
D/ALSADevice(  284): switchDevice: device 0x8 mode:2
D/ALSADevice(  284): getUCMDevice for output device: devices:8 is input device:0
D/ALSADevice(  284): switchDevice: rxDevice Voice Headphones devices:0x400008
D/ALSADevice(  284): getUCMDevice for input device: devices:400000 is input device:1
D/ALSADevice(  284): switchDevice: txDevice:Handset devices:0x400008

rxDevice : Voice Headphones (0x08) -> AUDIO_DEVICE_OUT_WIRED_HEADPHONE
txDevice : Handset (0x400000 ) -> AUDIO_DEVICE_IN_BUILTIN_MIC

[Flame]
D/ALSADevice(  281): mInChannels:1
D/ALSADevice(  281): route: devices 0x4 in mode 2
D/ALSADevice(  281): switchDevice: device 0x4 mode:2
D/ALSADevice(  281): getUCMDevice for output device: devices:4 is input device:0
D/ALSADevice(  281): switchDevice: rxDevice Voice Headphones devices:0x1000004
D/ALSADevice(  281): getUCMDevice for input device: devices:1000000 is input device:1
D/ALSADevice(  281): switchDevice: txDevice:Headset devices:0x1000004


rxDevice : Voice Headphones (0x04) -> AUDIO_DEVICE_OUT_WIRED_HEADSET
txDevice : Handset (0x1000000) -> AUDIO_DEVICE_IN_WIRED_HEADSET

[AudioSystem.h]
    /* output devices */      
    AUDIO_DEVICE_OUT_EARPIECE                  = 0x1,
    AUDIO_DEVICE_OUT_SPEAKER                   = 0x2,
    AUDIO_DEVICE_OUT_WIRED_HEADSET             = 0x4,
    AUDIO_DEVICE_OUT_WIRED_HEADPHONE           = 0x8,
    AUDIO_DEVICE_OUT_BLUETOOTH_SCO             = 0x10,

    /* input devices */
    AUDIO_DEVICE_IN_COMMUNICATION         = 0x100000,
    AUDIO_DEVICE_IN_AMBIENT               = 0x200000,
    AUDIO_DEVICE_IN_BUILTIN_MIC           = 0x400000,
    AUDIO_DEVICE_IN_BLUETOOTH_SCO_HEADSET = 0x800000,
    AUDIO_DEVICE_IN_WIRED_HEADSET         = 0x1000000,
    AUDIO_DEVICE_IN_AUX_DIGITAL           = 0x2000000,


I will check the headset event from GONK in gecko layer (AudioManager).

Comment 21

[Star Cheng [:scheng]]

I think this is hardware (audio jack) issue. Here are 2 case I tested. 
case 1: fully insert headset to audio jack
case 2: not fully insert headset to audio jack but make sure the headset icon appears in the status bar

[case 1] 
I/Gonk    (  285): [star] updateHeadphoneSwitch sMicrophoneState 1 event status 3
I/AudioManager(  285): [star] HeadphoneSwitchObserver Notify
I/AudioManager(  285): [star] SWITCH_STATE_HEADPHONE


[case 2]
I/Gonk    (  285): [star] updateHeadphoneSwitch sMicrophoneState 0 event status 2
I/AudioManager(  285): [star] HeadphoneSwitchObserver Notify
I/AudioManager(  285): [star] SWITCH_STATE_HEADSET


 
case 2 is normal (rx: headphone + tx: headphone mic )
case 1 is wrong  (rx: headphone + tx: handset mic )

Hi, Michael

I get the log from /proc/kmsg in kernel of the openC device. But it doesn't show any audio jack value. Could you help to check the below value in driver layer? 

input: msm8x10-skuaa-snd-card Headset Jack as /devices/sound.37/sound/card0/input7

Thanks

Comment 22

[Pooja]

(In reply to Star Cheng [:scheng] from comment #14)
> (In reply to Matthew Vaughan from comment #6)
> > Created attachment 8397429 [details]
> > LogCat - Open C using the Sony Ericsson headset during a call
> > 
> > (In reply to Preeti Raghunath(:Preeti) from comment #5)
> > > Lets please investigate on the Open C and put your findings down. We can
> > > check if the issue is on our end or not.
> > 
> > I tested 5 different headsets with microphones on the Open C. The results
> > for each headset are:
> > 
> > Alcatel headset = not detected
> > HTC headset     = not detected
> > Sony Ericsson MH410 = only the headset earbuds work; handset microphone is
> > used for speaking.
> > ZTE HMZ8-C4-OMTP    = only the headset earbuds work; handset microphone is
> > used for speaking.
> > Unknown brand       = only the headset earbuds work; handset microphone is
> > used for speaking.
> > 
> > Device: Open C v1.3 MOZ RIL
> > BuildID: 20140326040052
> > Gaia: be872843fa702e25ba113a93ce84fd8b265d87fb
> > Gecko: db0c8822aeafe9038815bdadf6d4312927cef250
> > Version: 28.0
> > Firmware Version: P821A10-ENG_20140321
> > 
> > Please note, all the listed headsets are detected and work correctly on a
> > Buri.
> > 
> > Additionally, I have attached a logcat for the Open C using the Sony
> > Ericsson headset.
> 
> 
> Hi, Matthew
> 
> Could you help to dump audio policy log by using executing "adb shell
> dumpsys > headset.dump" command? And check the value of "Output devices" in
> AudioPolicyManager Dump section. (the format looks like below:)
> 
> 
> AudioPolicyManager Dump: 0xb7f49eb8
>  Primary Output: 2
>  A2DP device address: 
>  SCO device address: 
>  USB audio ALSA 
>  Output devices: 0000000b   ->    (check this value)
>  Input devices: 00000044
>  Phone state: 0
>  Force use for communications 0
>  Force use for media 0
>  Force use for record 0
>  Force use for dock 0
>  Force use for system 0

Was on holiday so could not replied on time. BTW have attached the logs from our device too, If it helps in anyway

Comment 23

[Pooja]

Attachment: headset.dump

Comment 24

[Michael Vines [:m1] [:evilmachines]]

(In reply to Inder from comment #4)
> Matthew: this was tested on QRD8x26. Please work with Preeti to get hold of
> the device if you don't have it already.

Star, can you try on a QRD8x26 please.

Comment 25

[Star Cheng [:scheng]]

Hi, Michael


I need your help. I got confused about the expression "sMicrophoneState == AKEY_STATE_DOWN". Does it mean the device support microphone? (http://dxr.mozilla.org/mozilla-central/source/widget/gonk/nsAppShell.cpp#449)

If yes, it should be 
        event.status() = sMicrophoneState == AKEY_STATE_DOWN ?
            hal::SWITCH_STATE_HEADSET : hal::SWITCH_STATE_HEADPHONE;

instead of 
        event.status() = sMicrophoneState == AKEY_STATE_DOWN ?
            hal::SWITCH_STATE_HEADPHONE : SWITCH_STATE_HEADSET;

Comment 26

[Star Cheng [:scheng]]

Attachment: bug-987545-fix.patch

Hi, Poojas

Could you check this patch work?
Thanks

Comment 27

[Star Cheng [:scheng]]

(In reply to Star Cheng [:scheng] from comment #26)
> Created attachment 8403871 [details] [diff] [review]
> bug-987545-fix.patch
> 
> 
> Hi, Poojas
> 
> Could you check this patch work?
> Thanks


I tested 3 different headset(Hamachi hedset/iPhone 4S headset/Nexus S headset) to QRD8x26, but no one is detected as headset(with microphone). These headsets can be detected as headset in OpenC & Flame (MSM8810 platform). Can you help to check this patch by using correct headset to QRD8x26.


[1] insert headset
I/Gonk    (  252): [star] notifySwitch
I/Gonk    (  252): [star] needSwitchUpdate 1 sHeadphoneState 1 sMicrophoneState 0
I/Gonk    (  252): [star] updateHeadphoneSwitch sMicrophoneState 0 event status 3  -> SWITCH_STATE_HEADPHONE

[2]remove headset
I/Gonk    (  252): [star] notifySwitch
I/Gonk    (  252): [star] needSwitchUpdate 1 sHeadphoneState 0 sMicrophoneState 0
I/Gonk    (  252): [star] updateHeadphoneSwitch SWITCH_STATE_OFF

Comment 28

[Pooja]

(In reply to Star Cheng [:scheng] from comment #26)
> Created attachment 8403871 [details] [diff] [review]
> bug-987545-fix.patch
> 
> 
> Hi, Poojas
> 
> Could you check this patch work?
> Thanks

HI star,

Thanks for the fix. Its working . Issue is no more reproducible, We tried with US Headset

Comment 29

[Star Cheng [:scheng]]

Attachment: bug-987545-fix-patch-reviewed

1.TPBL result: https://tbpl.mozilla.org/?tree=Try&rev=09caedef6880
2.add reviewer name in patch file (Michanel, thanks you Y^.^y)

Comment 30

[Star Cheng [:scheng]]

(In reply to Star Cheng [:scheng] from comment #29)
> Created attachment 8405115 [details]
> bug-987545-fix-patch-reviewed
> 
> 
> 1.TPBL result: https://tbpl.mozilla.org/?tree=Try&rev=09caedef6880
> 2.add reviewer name in patch file (Michanel, thanks you Y^.^y)

Typo!! Michael, I am sorry! Orz....

Comment 31

[Star Cheng [:scheng]]

Attachment: bug-987545-fix-patch_v2-reviewed

fix worong summary.

Comment 32

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/integration/b2g-inbound/rev/c516a7214b17

Comment 33

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/c516a7214b17

Comment 34

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/releases/mozilla-aurora/rev/c4ccc01c351c

Comment 35

[Brogan Zumwalt [Inactive]]

A new test case needs to be written to address bug.

Comment 36

[KTucker [:KTucker][Inactive 3/4/2016]]

Test case added to moztrap:

https://moztrap.mozilla.org/manage/case/14334/