Adds support for Bluetooth headsets to the iOS audio layer

Adds support for Bluetooth headsets to the iOS audio layer.

This patch also also ensures that audio is restored after an incoming
GSM call.

BUG=webrtc:5058, webrtc:5012
TEST=Manual tests using modified AppRTCDemo and three different BT headsets

Committed: https://crrev.com/45c136b579c04c6cbf5f3ea77e7f8171d48a5890
Cr-Commit-Position: refs/heads/master@{#10354}

[henrika_webrtc, 2015-10-13 14:32:55]

henrika@webrtc.org changed reviewers:
+ tkchin@webrtc.org

[henrika_webrtc, 2015-10-13 14:32:56]

PTAL tkchin.

https://codereview.webrtc.org/1401963002/diff/1/webrtc/examples/objc/AppRTCDe...
File webrtc/examples/objc/AppRTCDemo/ios/ARDMainViewController.m (right):

https://codereview.webrtc.org/1401963002/diff/1/webrtc/examples/objc/AppRTCDe...
webrtc/examples/objc/AppRTCDemo/ios/ARDMainViewController.m:30: // [self
showAlertWithMessage:@"We are inactive!"];
Only for testing. Will not be checked in.

[henrika_webrtc, 2015-10-16 07:56:52]

Friendly ping :-)

[tkchin_webrtc, 2015-10-16 16:23:26]

On 2015/10/16 07:56:52, henrika_webrtc wrote:
> Friendly ping :-)

Thanks for ping, I'd missed the initial mail.
just some style things

[tkchin_webrtc, 2015-10-16 16:23:32]

https://codereview.webrtc.org/1401963002/diff/20001/webrtc/examples/objc/AppR...
File webrtc/examples/objc/AppRTCDemo/ios/ARDMainViewController.m (right):

https://codereview.webrtc.org/1401963002/diff/20001/webrtc/examples/objc/AppR...
webrtc/examples/objc/AppRTCDemo/ios/ARDMainViewController.m:30: // [self
showAlertWithMessage:@"We are inactive!"];
I'd leave the original bits in until we actually make it gracefully handle
backgrounding.

https://codereview.webrtc.org/1401963002/diff/20001/webrtc/modules/audio_devi...
File webrtc/modules/audio_device/ios/audio_device_ios.mm (right):

https://codereview.webrtc.org/1401963002/diff/20001/webrtc/modules/audio_devi...
webrtc/modules/audio_device/ios/audio_device_ios.mm:418: // TODO(henrika):
improve this section and try to merge actions with actions
imo we should make changes to depot_tools to allow 100 lines for objc/objc++
files

instead, to make things here more readable, suggest making a var for the block:

void (^interruption_block)(NSNotification *) = ^(NSNotification *) {
 ...
};

id interruption_observer =
    [center addObserverForName:AVAudioSessionInterruptionNotification
                        object:nil
                         queue:
                    usingBlock:interruption_block];

likewise for route_change_observer

https://codereview.webrtc.org/1401963002/diff/20001/webrtc/modules/audio_devi...
webrtc/modules/audio_device/ios/audio_device_ios.mm:426: [notification
userInfo][AVAudioSessionInterruptionTypeKey];
style:
notification.userInfo[AVAudioSessionInterruptionKey]
typeNumber.unsignedIntegerValue

https://codereview.webrtc.org/1401963002/diff/20001/webrtc/modules/audio_devi...
webrtc/modules/audio_device/ios/audio_device_ios.mm:443: // Post interruption
the audio unit render callbacks don't
Is this still true today?

https://codereview.webrtc.org/1401963002/diff/20001/webrtc/modules/audio_devi...
webrtc/modules/audio_device/ios/audio_device_ios.mm:460: NSUInteger reason_value
= [[notification.userInfo
style:
NSNumber *reason_number =
notification.userInfo[AVAudioSessionRouteChangeReasonKey];

Cast to the enum so that switch won't need default:

AVAudioSessionRouteChangeReason reason =
(AVAudioSessionRouteChangeReason)reason_number.unsignedIntegerValue;
switch (reason) {
  ...
}

https://codereview.webrtc.org/1401963002/diff/20001/webrtc/modules/audio_devi...
webrtc/modules/audio_device/ios/audio_device_ios.mm:496:
userInfo][AVAudioSessionRouteChangePreviousRouteKey];
ditto .userInfo

https://codereview.webrtc.org/1401963002/diff/20001/webrtc/modules/audio_devi...
webrtc/modules/audio_device/ios/audio_device_ios.mm:504: [[AVAudioSession
sharedInstance] sampleRate];
ditto dot syntax for properties

https://codereview.webrtc.org/1401963002/diff/20001/webrtc/modules/audio_devi...
webrtc/modules/audio_device/ios/audio_device_ios.mm:791: result =
AudioUnitUninitialize(vpio_unit_);
oh dear, were we really not doing this before? Thx for fixing.

[henrika_webrtc, 2015-10-20 10:18:16]

Thanks! Uploading changes. Will check interruption events next.

https://codereview.webrtc.org/1401963002/diff/20001/webrtc/examples/objc/AppR...
File webrtc/examples/objc/AppRTCDemo/ios/ARDMainViewController.m (right):

https://codereview.webrtc.org/1401963002/diff/20001/webrtc/examples/objc/AppR...
webrtc/examples/objc/AppRTCDemo/ios/ARDMainViewController.m:30: // [self
showAlertWithMessage:@"We are inactive!"];
Will do. Only added for testing purposes to avoid closing the call when I shift
audio device (since it kills AppRTCDemo today).

https://codereview.webrtc.org/1401963002/diff/20001/webrtc/modules/audio_devi...
File webrtc/modules/audio_device/ios/audio_device_ios.mm (right):

https://codereview.webrtc.org/1401963002/diff/20001/webrtc/modules/audio_devi...
webrtc/modules/audio_device/ios/audio_device_ios.mm:418: // TODO(henrika):
improve this section and try to merge actions with actions
No action on adding 100 lines support. 
Thanks for the tip on improving the block structure.

https://codereview.webrtc.org/1401963002/diff/20001/webrtc/modules/audio_devi...
webrtc/modules/audio_device/ios/audio_device_ios.mm:426: [notification
userInfo][AVAudioSessionInterruptionTypeKey];
On 2015/10/16 16:23:32, tkchin_webrtc wrote:
> style:
> notification.userInfo[AVAudioSessionInterruptionKey]
> typeNumber.unsignedIntegerValue

Done.

https://codereview.webrtc.org/1401963002/diff/20001/webrtc/modules/audio_devi...
webrtc/modules/audio_device/ios/audio_device_ios.mm:443: // Post interruption
the audio unit render callbacks don't
See TODO above. I will check and make changes in this CL if required.

https://codereview.webrtc.org/1401963002/diff/20001/webrtc/modules/audio_devi...
webrtc/modules/audio_device/ios/audio_device_ios.mm:460: NSUInteger reason_value
= [[notification.userInfo
On 2015/10/16 16:23:32, tkchin_webrtc wrote:
> style:
> NSNumber *reason_number =
> notification.userInfo[AVAudioSessionRouteChangeReasonKey];
> 
> Cast to the enum so that switch won't need default:
> 
> AVAudioSessionRouteChangeReason reason =
> (AVAudioSessionRouteChangeReason)reason_number.unsignedIntegerValue;
> switch (reason) {
>   ...
> }

Done.

https://codereview.webrtc.org/1401963002/diff/20001/webrtc/modules/audio_devi...
webrtc/modules/audio_device/ios/audio_device_ios.mm:496:
userInfo][AVAudioSessionRouteChangePreviousRouteKey];
On 2015/10/16 16:23:32, tkchin_webrtc wrote:
> ditto .userInfo

Done.

https://codereview.webrtc.org/1401963002/diff/20001/webrtc/modules/audio_devi...
webrtc/modules/audio_device/ios/audio_device_ios.mm:504: [[AVAudioSession
sharedInstance] sampleRate];
On 2015/10/16 16:23:32, tkchin_webrtc wrote:
> ditto dot syntax for properties

Done.

https://codereview.webrtc.org/1401963002/diff/20001/webrtc/modules/audio_devi...
webrtc/modules/audio_device/ios/audio_device_ios.mm:791: result =
AudioUnitUninitialize(vpio_unit_);
You are welcome ;-)

[henrika_webrtc, 2015-10-20 14:25:24]

Now also supports interruption handling. Have verified by making phone call to
ongoing WebRTC call.

PTAL

[tkchin_webrtc, 2015-10-21 02:21:51]

lgtm

https://codereview.webrtc.org/1401963002/diff/80001/webrtc/examples/objc/AppR...
File webrtc/examples/objc/AppRTCDemo/ios/ARDMainViewController.m (right):

https://codereview.webrtc.org/1401963002/diff/80001/webrtc/examples/objc/AppR...
webrtc/examples/objc/AppRTCDemo/ios/ARDMainViewController.m:31: // [self
dismissViewControllerAnimated:NO completion:nil];
uncomment second line? :)

https://codereview.webrtc.org/1401963002/diff/80001/webrtc/modules/audio_devi...
File webrtc/modules/audio_device/ios/audio_device_ios.mm (right):

https://codereview.webrtc.org/1401963002/diff/80001/webrtc/modules/audio_devi...
webrtc/modules/audio_device/ios/audio_device_ios.mm:447:
RTC_DCHECK(CheckAndLogError(success, error));
why a dcheck? This is something that can genuinely fail.

[henrika_webrtc, 2015-10-21 08:28:45]

Thanks!

https://codereview.webrtc.org/1401963002/diff/80001/webrtc/examples/objc/AppR...
File webrtc/examples/objc/AppRTCDemo/ios/ARDMainViewController.m (right):

https://codereview.webrtc.org/1401963002/diff/80001/webrtc/examples/objc/AppR...
webrtc/examples/objc/AppRTCDemo/ios/ARDMainViewController.m:31: // [self
dismissViewControllerAnimated:NO completion:nil];
On 2015/10/21 02:21:51, tkchin_webrtc wrote:
> uncomment second line? :)

Done.

https://codereview.webrtc.org/1401963002/diff/80001/webrtc/modules/audio_devi...
File webrtc/modules/audio_device/ios/audio_device_ios.mm (right):

https://codereview.webrtc.org/1401963002/diff/80001/webrtc/modules/audio_devi...
webrtc/modules/audio_device/ios/audio_device_ios.mm:447:
RTC_DCHECK(CheckAndLogError(success, error));
Removed DCHECK and instead log any error and only start the audio unit if the
session started as it should.

[henrika_webrtc, 2015-10-21 08:35:59]

Just FYI, this compiles locally for me but not on the bots:

FAILED: /b/build/goma/gomacc
../../third_party/llvm-build/Release+Asserts/bin/clang++ -MMD -MF
obj/webrtc/modules/audio_device/ios/audio_device.audio_device_ios.o.d
-DV8_DEPRECATION_WARNINGS -DCLD_VERSION=2 -DDISABLE_NACL -DCHROMIUM_BUILD
-DCR_CLANG_REVISION=247874-1 -DUSE_LIBJPEG_TURBO=1 -DENABLE_CONFIGURATION_POLICY
-DSYSTEM_NATIVELY_SIGNALS_MEMORY_PRESSURE -DDONT_EMBED_BUILD_METADATA
-DDCHECK_ALWAYS_ON=1 -DFIELDTRIAL_TESTING_ENABLED -DDISABLE_FTP_SUPPORT=1
-DV8_USE_EXTERNAL_STARTUP_DATA -DWEBRTC_RESTRICT_LOGGING -DEXPAT_RELATIVE_PATH
-DWEBRTC_ARCH_ARM -DWEBRTC_ARCH_ARM_V7 -DWEBRTC_HAS_NEON -DWEBRTC_POSIX
-DWEBRTC_MAC -DWEBRTC_IOS -DWEBRTC_INCLUDE_INTERNAL_AUDIO_DEVICE -DUSE_LIBPCI=1
-DUSE_OPENSSL=1 -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS
-DDYNAMIC_ANNOTATIONS_ENABLED=1 -DWTF_USE_DYNAMIC_ANNOTATIONS=1 -D_DEBUG -Igen
-I../.. -I../../webrtc/modules/audio_device -I../../webrtc/modules/interface
-I../../webrtc/modules/audio_device/include
-I../../webrtc/modules/audio_device/dummy
-I../../webrtc/modules/audio_device/ios
-I../../webrtc/common_audio/resampler/include
-I../../webrtc/common_audio/signal_processing/include
-I../../webrtc/common_audio/vad/include -isysroot
/Applications/Xcode5.1.1.app/Contents/Developer/Platforms/iPhoneOS.platform/Developer/SDKs/iPhoneOS7.1.sdk
-O0 -fvisibility=hidden -Werror -Wnewline-eof -miphoneos-version-min=7.0 -arch
armv7 -Wall -Wendif-labels -Wextra -Wno-unused-parameter
-Wno-missing-field-initializers -Wno-selector-type-mismatch -Wheader-hygiene
-Wno-char-subscripts -Wno-unneeded-internal-declaration
-Wno-covered-switch-default -Wstring-conversion -Wno-c++11-narrowing
-Wno-deprecated-register -Wno-inconsistent-missing-override
-Wno-shift-negative-value -Wno-bitfield-width -std=c++11 -stdlib=libc++
-fno-rtti -fno-exceptions -fvisibility-inlines-hidden -fno-threadsafe-statics
-fcolor-diagnostics -fstack-protector-all -Wno-undefined-bool-conversion
-Wno-tautological-undefined-compare -fobjc-arc -Wobjc-missing-property-synthesis
-fobjc-call-cxx-cdtors  -c
../../webrtc/modules/audio_device/ios/audio_device_ios.mm -o
obj/webrtc/modules/audio_device/ios/audio_device.audio_device_ios.o
../../webrtc/modules/audio_device/ios/audio_device_ios.mm:510:45:error: property
'sampleRate' not found on object of type 'id'
              AVAudioSession.sharedInstance.sampleRate;

now changed to [AVAudioSession sharedInstance].sampleRate

[henrika_webrtc, 2015-10-21 08:36:15]

The CQ bit was checked by henrika@webrtc.org

[henrika_webrtc, 2015-10-21 08:36:15]

The patchset sent to the CQ was uploaded after l-g-t-m from tkchin@webrtc.org
Link to the patchset: https://codereview.webrtc.org/1401963002/#ps120001 (title:
"now passes trybotsq")

[commit-bot: I haz the power, 2015-10-21 08:36:27]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1401963002/120001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1401963002/120001

[henrika_webrtc, 2015-10-21 08:37:08]

The CQ bit was unchecked by henrika@webrtc.org

[henrika_webrtc, 2015-10-21 08:42:50]

Description was changed from

==========
Adds support for Bluetooth headsets to the iOS audio layer

BUG=webrtc:5058
TEST=Manual tests using modified AppRTCDemo and three different BT headsets
==========

to

==========
Adds support for Bluetooth headsets to the iOS audio layer.

This patch also also ensures that audio is restored after an incoming
GSM call.

BUG=webrtc:5058, webrtc:5012
TEST=Manual tests using modified AppRTCDemo and three different BT headsets
==========

[henrika_webrtc, 2015-10-21 08:47:29]

The CQ bit was checked by henrika@webrtc.org

[henrika_webrtc, 2015-10-21 08:47:30]

The patchset sent to the CQ was uploaded after l-g-t-m from tkchin@webrtc.org
Link to the patchset: https://codereview.webrtc.org/1401963002/#ps160001 (title:
"nit")

[commit-bot: I haz the power, 2015-10-21 08:47:35]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1401963002/160001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1401963002/160001

[commit-bot: I haz the power, 2015-10-21 11:11:58]

Committed patchset #9 (id:160001)

[commit-bot: I haz the power, 2015-10-21 11:12:12]

Patchset 9 (id:??) landed as
https://crrev.com/45c136b579c04c6cbf5f3ea77e7f8171d48a5890
Cr-Commit-Position: refs/heads/master@{#10354}