OSDN Git Service

android-x86/frameworks-av.git
12 years agoam 1f8c6212: Merge "visualizer: reset buffer if playback is idle" into ics-mr1
Eric Laurent [Fri, 11 Nov 2011 19:50:28 +0000 (11:50 -0800)]
am 1f8c6212: Merge "visualizer: reset buffer if playback is idle" into ics-mr1

* commit '1f8c621279ddb163ee3d63bf32d79955679db0c5':
  visualizer: reset buffer if playback is idle

12 years agoam cd556b82: Merge "Instead of asserting, signal a runtime error if the session doesn...
Andreas Huber [Fri, 11 Nov 2011 19:50:23 +0000 (11:50 -0800)]
am cd556b82: Merge "Instead of asserting, signal a runtime error if the session doesn\'t contain" into ics-mr1

* commit 'cd556b82fd12dceb49e9d6543f80212721b220ec':
  Instead of asserting, signal a runtime error if the session doesn't contain

12 years agoam 5d42075e: Merge "DO NOT MERGE: Instead of asserting, remove active streams if...
Andreas Huber [Fri, 11 Nov 2011 19:50:21 +0000 (11:50 -0800)]
am 5d42075e: Merge "DO NOT MERGE: Instead of asserting, remove active streams if their sockets" into ics-mr1

* commit '5d42075e450a49af2bb37ca3a9610921df7607f2':
  DO NOT MERGE: Instead of asserting, remove active streams if their sockets

12 years agoam d9f25bc9: Merge "Disconnect on socket error on the RTSP control connection." into...
Andreas Huber [Fri, 11 Nov 2011 19:50:20 +0000 (11:50 -0800)]
am d9f25bc9: Merge "Disconnect on socket error on the RTSP control connection." into ics-mr1

* commit 'd9f25bc941746f268ab978e6b2af7b7c4ec2ea89':
  Disconnect on socket error on the RTSP control connection.

12 years agoam 08e42967: Merge "Add support for retrieving location information in mp4/3gpp files...
James Dong [Fri, 11 Nov 2011 19:50:14 +0000 (11:50 -0800)]
am 08e42967: Merge "Add support for retrieving location information in mp4/3gpp files" into ics-mr1

* commit '08e42967b9e459b7beb22e859f1f1eba0546741b':
  Add support for retrieving location information in mp4/3gpp files

12 years agoMerge "visualizer: reset buffer if playback is idle" into ics-mr1
Eric Laurent [Fri, 11 Nov 2011 16:17:22 +0000 (08:17 -0800)]
Merge "visualizer: reset buffer if playback is idle" into ics-mr1

12 years agoMerge "Instead of asserting, signal a runtime error if the session doesn't contain...
Andreas Huber [Fri, 11 Nov 2011 15:39:25 +0000 (07:39 -0800)]
Merge "Instead of asserting, signal a runtime error if the session doesn't contain" into ics-mr1

12 years agoMerge "DO NOT MERGE: Instead of asserting, remove active streams if their sockets...
Andreas Huber [Fri, 11 Nov 2011 15:39:18 +0000 (07:39 -0800)]
Merge "DO NOT MERGE: Instead of asserting, remove active streams if their sockets" into ics-mr1

12 years agoMerge "Disconnect on socket error on the RTSP control connection." into ics-mr1
Andreas Huber [Fri, 11 Nov 2011 15:39:11 +0000 (07:39 -0800)]
Merge "Disconnect on socket error on the RTSP control connection." into ics-mr1

12 years agoDisconnect on socket error on the RTSP control connection.
Andreas Huber [Thu, 10 Nov 2011 22:48:20 +0000 (14:48 -0800)]
Disconnect on socket error on the RTSP control connection.

Change-Id: Ib52a69f9b0830b481c6f5c9b1991d1f4cb36ec7b

12 years agoDO NOT MERGE: Instead of asserting, remove active streams if their sockets
Andreas Huber [Wed, 9 Nov 2011 22:26:43 +0000 (14:26 -0800)]
DO NOT MERGE: Instead of asserting, remove active streams if their sockets

return failure

Change-Id: Icb47adfd2fbe0398c473ba66e068186311c9cc79
related-to-bug: 5593654

12 years agoInstead of asserting, signal a runtime error if the session doesn't contain
Andreas Huber [Thu, 10 Nov 2011 20:40:30 +0000 (12:40 -0800)]
Instead of asserting, signal a runtime error if the session doesn't contain

any playable tracks at all.

Change-Id: Ibbbe2fdcd53b7e020da80c84c8229856107a87e6

12 years agoAdd support for retrieving location information in mp4/3gpp files
James Dong [Wed, 9 Nov 2011 08:48:56 +0000 (00:48 -0800)]
Add support for retrieving location information in mp4/3gpp files

Change-Id: Ib8427704ef9ee5a4fa8fe1818c8a62d77b4ea687
related-to-bug: 5560253

12 years agovisualizer: reset buffer if playback is idle
Eric Laurent [Thu, 10 Nov 2011 18:02:18 +0000 (10:02 -0800)]
visualizer: reset buffer if playback is idle

Visualizer capture buffer must be reset if audio framework has stopped
calling process for a given period of time to get read of residual
data from previous captures.

Issue 5571920.

Change-Id: I6e73f971bb812cdbb2979a3b5e763abab07634eb

12 years agoam 2375d163: Merge "Send RTSP control connection keep-alive requests" into ics-mr1
Andreas Huber [Thu, 10 Nov 2011 16:52:19 +0000 (16:52 +0000)]
am 2375d163: Merge "Send RTSP control connection keep-alive requests" into ics-mr1

* commit '2375d16378e6f8b7694c5a2cd9e0622e382e6049':
  Send RTSP control connection keep-alive requests

12 years agoam b42aea5b: Merge "Don\'t believe the PV decoder\'s lies..." into ics-mr1
Andreas Huber [Thu, 10 Nov 2011 16:52:17 +0000 (16:52 +0000)]
am b42aea5b: Merge "Don\'t believe the PV decoder\'s lies..." into ics-mr1

* commit 'b42aea5b4c7cc7b2766f80f5ceda94b324d776e5':
  Don't believe the PV decoder's lies...

12 years agoam 7c063cf9: Merge "There\'s no point in trying to continue to decode after an error...
Andreas Huber [Thu, 10 Nov 2011 16:49:12 +0000 (16:49 +0000)]
am 7c063cf9: Merge "There\'s no point in trying to continue to decode after an error was signalled." into ics-mr1

* commit '7c063cf9679a2ecdda5e1639cfdca3e1084019f9':
  There's no point in trying to continue to decode after an error was signalled.

12 years agoMerge "Send RTSP control connection keep-alive requests" into ics-mr1
Andreas Huber [Thu, 10 Nov 2011 16:48:44 +0000 (08:48 -0800)]
Merge "Send RTSP control connection keep-alive requests" into ics-mr1

12 years agoMerge "Don't believe the PV decoder's lies..." into ics-mr1
Andreas Huber [Thu, 10 Nov 2011 16:48:34 +0000 (08:48 -0800)]
Merge "Don't believe the PV decoder's lies..." into ics-mr1

12 years agoMerge "There's no point in trying to continue to decode after an error was signalled...
Andreas Huber [Thu, 10 Nov 2011 16:47:25 +0000 (08:47 -0800)]
Merge "There's no point in trying to continue to decode after an error was signalled." into ics-mr1

12 years agoam f2c2b1c4: am 19998120: Merge "Bug 5478024 drop after decode if can\'t drop before...
Glenn Kasten [Wed, 9 Nov 2011 19:00:05 +0000 (19:00 +0000)]
am f2c2b1c4: am 19998120: Merge "Bug 5478024 drop after decode if can\'t drop before" into ics-mr0

* commit 'f2c2b1c484b7b175c88c33efbbcb74d93f7eb0a2':
  Bug 5478024 drop after decode if can't drop before

12 years agoam 2b7f91b9: Merge "Fix regression for SoundPool playback" into ics-mr1
Eric Laurent [Wed, 9 Nov 2011 18:59:51 +0000 (18:59 +0000)]
am 2b7f91b9: Merge "Fix regression for SoundPool playback" into ics-mr1

* commit '2b7f91b932be0aa9ce6fa3acda9ff439e94643fc':
  Fix regression for SoundPool playback

12 years agoThere's no point in trying to continue to decode after an error was signalled.
Andreas Huber [Wed, 9 Nov 2011 18:26:06 +0000 (10:26 -0800)]
There's no point in trying to continue to decode after an error was signalled.

Change-Id: I2a2410cac9444eadd85be5d104799703171f1a24
related-to-bug: 5588658

12 years agoDon't believe the PV decoder's lies...
Andreas Huber [Wed, 9 Nov 2011 18:03:07 +0000 (10:03 -0800)]
Don't believe the PV decoder's lies...

Change-Id: I97391235255afb8a4890a4ff52d4e8e9ca954f27

12 years agoam 19998120: Merge "Bug 5478024 drop after decode if can\'t drop before" into ics-mr0
Glenn Kasten [Wed, 9 Nov 2011 16:41:19 +0000 (16:41 +0000)]
am 19998120: Merge "Bug 5478024 drop after decode if can\'t drop before" into ics-mr0

* commit '19998120f53caceb620f138c6eea182f1e693660':
  Bug 5478024 drop after decode if can't drop before

12 years agoFix regression for SoundPool playback
Eric Laurent [Wed, 9 Nov 2011 02:10:16 +0000 (18:10 -0800)]
Fix regression for SoundPool playback

Commit 19ddf0eb introduced a problem with applications (like SoundPool)
relying on an underrun condition to detect end of playback instead of
stopping the track when all data is written.
AudioFlinger would keep waiting for new data in case of partial buffer
filling and never reach the underrun condition.

Added a mechanism to wait no more than once if not enough frames are present
in the track buffer.

Issue 5585490.

Change-Id: I131e605ff6070831a01ddf734e68459e3bf2354b

12 years agoSend RTSP control connection keep-alive requests
Andreas Huber [Tue, 8 Nov 2011 23:31:23 +0000 (15:31 -0800)]
Send RTSP control connection keep-alive requests

default to 60 secs unless overridden by server's session-id response.

Change-Id: I7c3aff5b787dbb57cc0dccf9db3c75e5cf7e778c
related-to-bug: 5562303

12 years agoam 26f70db9: Merge "Remove surface legacy APIs and code." into ics-mr1
Andreas Huber [Tue, 8 Nov 2011 17:58:45 +0000 (17:58 +0000)]
am 26f70db9: Merge "Remove surface legacy APIs and code." into ics-mr1

* commit '26f70db99f483be36caa7a4c84fec5de50bec034':
  Remove surface legacy APIs and code.

12 years agoresolved conflicts for merge of f3a892ab to ics-aah
Eric Laurent [Tue, 8 Nov 2011 17:52:15 +0000 (09:52 -0800)]
resolved conflicts for merge of f3a892ab to ics-aah

Change-Id: Id9c0ef089d4ed71dbf69c5969b36f9ac44a25ad3

12 years agoMerge "Remove surface legacy APIs and code." into ics-mr1
Andreas Huber [Tue, 8 Nov 2011 16:48:31 +0000 (08:48 -0800)]
Merge "Remove surface legacy APIs and code." into ics-mr1

12 years agoMerge "AudioFlinger: mix track only when really ready" into ics-mr1
Eric Laurent [Tue, 8 Nov 2011 16:32:57 +0000 (08:32 -0800)]
Merge "AudioFlinger: mix track only when really ready" into ics-mr1

12 years agoam ee7ebb32: Merge "Fix a hang issue where the first frame can be dropped for timelap...
James Dong [Tue, 8 Nov 2011 06:58:11 +0000 (06:58 +0000)]
am ee7ebb32: Merge "Fix a hang issue where the first frame can be dropped for timelapse video recording." into ics-mr1

* commit 'ee7ebb3257d9337627b175835dc3d08f9b8ab339':
  Fix a hang issue where the first frame can be dropped for timelapse video recording.

12 years agoMerge "Fix a hang issue where the first frame can be dropped for timelapse video...
James Dong [Tue, 8 Nov 2011 06:57:02 +0000 (22:57 -0800)]
Merge "Fix a hang issue where the first frame can be dropped for timelapse video recording." into ics-mr1

12 years agoam 064975c2: Reconcile with ics-mr0-release ics-factoryrom-2-release
The Android Open Source Project [Mon, 7 Nov 2011 17:37:20 +0000 (17:37 +0000)]
am 064975c2: Reconcile with ics-mr0-release ics-factoryrom-2-release

* commit '064975c27cbca412e682fa152ce00af14027ab4f':
  Fix a leak in Parcel::writeBlob.
  DO NOT MERGE Poll input data with a small timeout and don't consume a full core.

12 years agoReconcile with ics-mr0-release ics-factoryrom-2-release
The Android Open Source Project [Mon, 7 Nov 2011 16:22:56 +0000 (08:22 -0800)]
Reconcile with ics-mr0-release ics-factoryrom-2-release

Change-Id: I9de1253d253c7d106862534570cad0c89d2fdaf2

12 years agoBug 5478024 drop after decode if can't drop before
Glenn Kasten [Sat, 5 Nov 2011 01:05:35 +0000 (18:05 -0700)]
Bug 5478024 drop after decode if can't drop before

Change-Id: Iaa64553be1a710da5d42be4e3a1e1d6dc873e371

12 years agoDO NOT MERGE Poll input data with a small timeout and don't consume a full core.
Andreas Huber [Thu, 3 Nov 2011 18:00:21 +0000 (11:00 -0700)]
DO NOT MERGE Poll input data with a small timeout and don't consume a full core.

Change-Id: I3c288698920fe6ead0df24a52330483609821a41
related-to-bug: 5549263

12 years agoam 914972be: Merge "Fix device management in lvm effect bundle wrapper." into ics-mr1
Eric Laurent [Fri, 4 Nov 2011 15:32:13 +0000 (15:32 +0000)]
am 914972be: Merge "Fix device management in lvm effect bundle wrapper." into ics-mr1

* commit '914972be65bef1fd7e886825e7af27c29eff0a1e':
  Fix device management in lvm effect bundle wrapper.

12 years agoam 79178b8b: Merge "Fix problem in lvm effect bundle wrapper." into ics-mr1
Eric Laurent [Fri, 4 Nov 2011 15:29:44 +0000 (15:29 +0000)]
am 79178b8b: Merge "Fix problem in lvm effect bundle wrapper." into ics-mr1

* commit '79178b8b3da8f69f34a72be13f2a374afbec4f40':
  Fix problem in lvm effect bundle wrapper.

12 years agoMerge "Fix device management in lvm effect bundle wrapper." into ics-mr1
Eric Laurent [Fri, 4 Nov 2011 15:29:30 +0000 (08:29 -0700)]
Merge "Fix device management in lvm effect bundle wrapper." into ics-mr1

12 years agoMerge "Fix problem in lvm effect bundle wrapper." into ics-mr1
Eric Laurent [Fri, 4 Nov 2011 15:27:45 +0000 (08:27 -0700)]
Merge "Fix problem in lvm effect bundle wrapper." into ics-mr1

12 years agoam 8242cc49: Reconcile with ics-mr0-release ics-factoryrom-2-release
The Android Open Source Project [Fri, 4 Nov 2011 14:59:40 +0000 (14:59 +0000)]
am 8242cc49: Reconcile with ics-mr0-release ics-factoryrom-2-release

* commit '8242cc49a90aa13a4830a735438b6e3cf73ea0f0':
  Revert "Merge "Avoid duplicate dialogs leading to NPE" into ics-mr0"
  Revert "Merge "BatteryService(jni): properly handle read's return value" into ics-mr0"
  Revert "Merge "Update camera continuous autofocus javadoc." into ics-mr0"
  Revert "Merge "Fix potential segfault in RS watchdog." into ics-mr0"
  Revert "Merge "Improve the slow query instrumentation." into ics-mr0"

12 years agoReconcile with ics-mr0-release ics-factoryrom-2-release
The Android Open Source Project [Fri, 4 Nov 2011 14:56:42 +0000 (07:56 -0700)]
Reconcile with ics-mr0-release ics-factoryrom-2-release

Change-Id: I81cf3b3994cb9a1c18fca3ce0049e09402543086

12 years agoam 43e72b1f: am 32b5a7ee: (-s ours) Merge "DO NOT MERGE Poll input data with a small...
Dave Burke [Fri, 4 Nov 2011 12:03:26 +0000 (12:03 +0000)]
am 43e72b1f: am 32b5a7ee: (-s ours) Merge "DO NOT MERGE Poll input data with a small timeout and don\'t consume a full core." into ics-mr0

* commit '43e72b1f5692dd9e220158001317c2abc0ecafcf':
  DO NOT MERGE Poll input data with a small timeout and don't consume a full core.

12 years agoam 32b5a7ee: (-s ours) Merge "DO NOT MERGE Poll input data with a small timeout and...
Dave Burke [Fri, 4 Nov 2011 11:59:27 +0000 (11:59 +0000)]
am 32b5a7ee: (-s ours) Merge "DO NOT MERGE Poll input data with a small timeout and don\'t consume a full core." into ics-mr0

* commit '32b5a7eeaf758fad0c24131603cdd54cd2883a58':
  DO NOT MERGE Poll input data with a small timeout and don't consume a full core.

12 years agoDO NOT MERGE Poll input data with a small timeout and don't consume a full core.
Andreas Huber [Thu, 3 Nov 2011 18:00:21 +0000 (11:00 -0700)]
DO NOT MERGE Poll input data with a small timeout and don't consume a full core.

Change-Id: I3c288698920fe6ead0df24a52330483609821a41
related-to-bug: 5549263

12 years agoAudioFlinger: mix track only when really ready
Eric Laurent [Thu, 3 Nov 2011 19:16:05 +0000 (12:16 -0700)]
AudioFlinger: mix track only when really ready

The addition of low power audio playback mode made that audio buffer consumption
by audio HAL can now happen in bursts. This makes that requesting audio data
from an AudioTrack for mixing can happen at much shorter intervals than before.
This revealed an existing problem where AudioFlinger would consider a track ready
for mixing although not enough frames were ready to completely fill one output buffer,
thus creating short periods of silence.

The fix consists in waiting for enough frames to be ready in AudioTrack buffer before
declaring a track ready for mixing. This minimum is not applied when the track is stopped
to allow the buffer to be emptied completely.

Change-Id: I6d04f9b65db5af85b0b53f0a5674be7ec02f9e9f

12 years agoam e395f645: Poll input data with a small timeout and don\'t consume a full core.
Andreas Huber [Thu, 3 Nov 2011 18:10:44 +0000 (18:10 +0000)]
am e395f645: Poll input data with a small timeout and don\'t consume a full core.

* commit 'e395f645aa4c30a4fad8beb3da6f8fad72338617':
  Poll input data with a small timeout and don't consume a full core.

12 years agoPoll input data with a small timeout and don't consume a full core.
Andreas Huber [Thu, 3 Nov 2011 18:00:21 +0000 (11:00 -0700)]
Poll input data with a small timeout and don't consume a full core.

Change-Id: I3c288698920fe6ead0df24a52330483609821a41
related-to-bug: 5549263

12 years agoRevert "Merge "Update camera continuous autofocus javadoc." into ics-mr0"
The Android Automerger [Thu, 3 Nov 2011 03:59:04 +0000 (20:59 -0700)]
Revert "Merge "Update camera continuous autofocus javadoc." into ics-mr0"

This reverts commit 4b6353ea0265bfed52d0637abd1b17596ce25ff0, reversing
changes made to af675222f6340a8a9edbe9e8635014a18521e5e0.

12 years agomerge in ics-factoryrom-2-release history after reset to ics-mr0
The Android Automerger [Thu, 3 Nov 2011 03:53:06 +0000 (20:53 -0700)]
merge in ics-factoryrom-2-release history after reset to ics-mr0

12 years agoam 11b49ba9: Reconcile with ics-mr0-release ics-factoryrom-2-release
The Android Open Source Project [Wed, 2 Nov 2011 17:56:48 +0000 (17:56 +0000)]
am 11b49ba9: Reconcile with ics-mr0-release ics-factoryrom-2-release

* commit '11b49ba933bab7ff8262139141eef0e5f3c45a18':
  Revert "Merge "BatteryService(jni): properly handle read's return value" into ics-mr0"
  Revert "Merge "Update camera continuous autofocus javadoc." into ics-mr0"
  Revert "Merge "Fix potential segfault in RS watchdog." into ics-mr0"
  Revert "Merge "Improve the slow query instrumentation." into ics-mr0"
  Revert "Merge "BatteryService(jni): properly handle read's return value" into ics-mr0"

12 years agoReconcile with ics-mr0-release ics-factoryrom-2-release
The Android Open Source Project [Wed, 2 Nov 2011 17:52:48 +0000 (10:52 -0700)]
Reconcile with ics-mr0-release ics-factoryrom-2-release

Change-Id: I493695d0ea182f32bbfe0e148e192ad1f60ff001

12 years agoRevert "Merge "Update camera continuous autofocus javadoc." into ics-mr0"
The Android Automerger [Wed, 2 Nov 2011 17:14:18 +0000 (10:14 -0700)]
Revert "Merge "Update camera continuous autofocus javadoc." into ics-mr0"

This reverts commit 4b6353ea0265bfed52d0637abd1b17596ce25ff0, reversing
changes made to af675222f6340a8a9edbe9e8635014a18521e5e0.

12 years agoFix problem in lvm effect bundle wrapper.
Eric Laurent [Tue, 1 Nov 2011 00:36:03 +0000 (17:36 -0700)]
Fix problem in lvm effect bundle wrapper.

When an effect is disabled, the process function should either
copy or accumulate the content of the input buffer to the output buffer
depending on the behavior requested by the framework.
Current implementation is copying the input buffer unconditionally.

Related to issue 5433942.

Change-Id: Ic488ca97eadcc4c763de570d7e6c6f5b7a979415

12 years agoFix device management in lvm effect bundle wrapper.
Eric Laurent [Tue, 1 Nov 2011 01:48:15 +0000 (18:48 -0700)]
Fix device management in lvm effect bundle wrapper.

The virtualizer and bass boost effects should be disabled when
not playing throught headsets or headphones.

There is a bug in current wrapper implementation that makes that those
effects are not disabled if the effect is enabled by the user after
the headset has been removed (current logic only works if the device selection
occurs while the effect is enabled).

Change-Id: I7c66b15a0339d95fb9ed13c8320d66379725d3b6

12 years agoam 11ba6da4: am 4b6353ea: Merge "Update camera continuous autofocus javadoc." into...
Chih-Chung Chang [Tue, 1 Nov 2011 04:09:00 +0000 (04:09 +0000)]
am 11ba6da4: am 4b6353ea: Merge "Update camera continuous autofocus javadoc." into ics-mr0

* commit '11ba6da460572b9ce9cb9dc0d5a76b710e7efdb3':
  Update camera continuous autofocus javadoc.

12 years agoam 4b6353ea: Merge "Update camera continuous autofocus javadoc." into ics-mr0
Chih-Chung Chang [Tue, 1 Nov 2011 04:05:45 +0000 (04:05 +0000)]
am 4b6353ea: Merge "Update camera continuous autofocus javadoc." into ics-mr0

* commit '4b6353ea0265bfed52d0637abd1b17596ce25ff0':
  Update camera continuous autofocus javadoc.

12 years agoMerge "Update camera continuous autofocus javadoc." into ics-mr0
Chih-Chung Chang [Tue, 1 Nov 2011 04:04:01 +0000 (21:04 -0700)]
Merge "Update camera continuous autofocus javadoc." into ics-mr0

12 years agomerged by hand (services/java/com/android/server/PowerManagerService.java needs to...
Ed Heyl [Mon, 31 Oct 2011 13:06:27 +0000 (06:06 -0700)]
merged by hand (services/java/com/android/server/PowerManagerService.java needs to be reviewed)

Change-Id: I86d1111d86cd1646ebc8a88d58aa393089e9f928

12 years agokeep previous history after reset to mr1 plus aah changes (ics-aah-wip)
Ed Heyl [Fri, 28 Oct 2011 19:05:53 +0000 (19:05 +0000)]
keep previous history after reset to mr1 plus aah changes (ics-aah-wip)

12 years agoRemove surface legacy APIs and code.
Andreas Huber [Tue, 25 Oct 2011 20:45:00 +0000 (13:45 -0700)]
Remove surface legacy APIs and code.

All surfaces are now supported through surface textures.

Change-Id: I95dd823e7099c0c32a48a1121624149dcc29d9c6

12 years agoMerge "Properly report that we're no longer playing after receiving event" into ics-mr1
Andreas Huber [Fri, 28 Oct 2011 16:13:54 +0000 (09:13 -0700)]
Merge "Properly report that we're no longer playing after receiving event" into ics-mr1

12 years agoDo not inline the audio resampler assembly functions
Jason Simmons [Thu, 27 Oct 2011 23:38:29 +0000 (16:38 -0700)]
Do not inline the audio resampler assembly functions

The assembly expects arguments to live at fixed offsets from the stack pointer
which are invalid if the code is inlined.

Change-Id: Ie93e93c5c69774079112345754fbc85896fc2f64

12 years agoConfigure the A@H TX media player with one string
Jason Simmons [Fri, 16 Sep 2011 20:46:40 +0000 (13:46 -0700)]
Configure the A@H TX media player with one string

Music2 would prefer a configuration interface that takes a single
blob of data passed in from the RPC client and then passes it
opaquely to the media player.  This eliminates the need for Music2
to rev their application if we add more configuration options to
the TX player.

Change-Id: Ia2daea3ff2502f91f8a5a82b898752c435279a0f

12 years agoEnsure that the TRTP retry buffer has contiguous sequence numbers
Jason Simmons [Thu, 15 Sep 2011 21:19:36 +0000 (14:19 -0700)]
Ensure that the TRTP retry buffer has contiguous sequence numbers

Previously, sequence numbers for audio packets were assigned by the
TX player before packets were queued to the sender.  This caused a
race between assignment of sequence numbers on audio packets and
sequence numbers on heartbeat packets.  A heartbeat could get queued
and added to the retry buffer before an audio packet with an earlier
sequence number got queued.

This CL centralizes packet sequence number assignment and insertion
into the retry buffer inside AAH_TXSender::doSendPacket_l.  It also
makes explicit what operations can be done on a TRTPPacket before
and after packing.

Change-Id: I6d02eae81061983e4def4f1b3dd7c1625467b151

12 years agoHold the AAHDecoderPump ThreadWrapper in a ref-counting pointer
Jason Simmons [Mon, 12 Sep 2011 23:58:07 +0000 (16:58 -0700)]
Hold the AAHDecoderPump ThreadWrapper in a ref-counting pointer

Change-Id: Iff840dbd9e1f79a07e62c1481b2b0797f30247cb

12 years agoMediaPlayer changes for Tungsten/Android@Home.
Mike J. Chen [Mon, 15 Aug 2011 20:24:13 +0000 (13:24 -0700)]
MediaPlayer changes for Tungsten/Android@Home.

Squashed merge from master-tungsten of the following changes:

commit e8b156c6b224f0fe5b6a81d5f92bade36275027c
Author: Jason Simmons <jsimmons@google.com>
Date:   Fri Aug 12 10:28:48 2011 -0700

    Add a null check to handle failure of android::createPlayer

    Change-Id: Iacbbc90369a825838d4374b1f774c8880b25753b

commit 43be3231034ff8537fdd84422a7954780038671f
Author: John Grossman <johngro@google.com>
Date:   Mon Jun 27 18:59:12 2011 -0700

    Move libaah_rtp over from the vendor directory.

    Also move factor PipeEvent out into utils.

    Change-Id: Id3877c66efe22d771cf3ef4877107e431b828e37

commit 17526eb3148c9c3d4365b6d5b47e8dc13bca71b6
Author: John Grossman <johngro@google.com>
Date:   Mon Jun 27 17:06:49 2011 -0700

    Name changes for the TRTP Players s/tungsten/aah/g

    Change-Id: I55e9ad13003f6aa6a36955b54426a7efbe31ac51

commit cbf2903ab6893b6e662514e2f6d670e268a419df
Author: John Grossman <johngro@google.com>
Date:   Fri Apr 15 09:27:54 2011 -0700

    Migrate Tungsten code from the HC-Tungsten to the Master-Tungsten branch.

    Change-Id: I95372d913a0761d90168edb4016f5ece0ea74502

Change-Id: I73f3be191cd974933acddf52892e5600d253cc03
Signed-off-by: Mike J. Chen <mjchen@google.com>
Signed-off-by: John Grossman <johngro@google.com>
12 years agoAdd media/libaah_rtp
Mike J. Chen [Mon, 15 Aug 2011 18:26:02 +0000 (11:26 -0700)]
Add media/libaah_rtp

RTP library used to broadcast media from one device to a collection
of listeners.  Handles failures/retries/etc.

This is a squashed merge from master-tungsten of the following changes:

commit e1a5101fe627d71739a7c4263bb3a65c7bc44385
Author: Jason Simmons <jsimmons@google.com>
Date:   Fri Aug 12 13:24:21 2011 -0700

    Hold ThreadWrapper in a ref-counting pointer

    Change-Id: Iaf3343182e37bcc0ca99fbaf8f9bbb8c4984072a

commit 89b90d62e164ff3db27c9cba85255fc476d2dd96
Author: Jason Simmons <jsimmons@google.com>
Date:   Wed Aug 10 13:08:25 2011 -0700

    Update the Tungsten TX player to use HTTPBase

    Change-Id: I9f7ecf1b4b496cec1815284dbcdb958a43284169

commit 43be3231034ff8537fdd84422a7954780038671f
Author: John Grossman <johngro@google.com>
Date:   Mon Jun 27 18:59:12 2011 -0700

    Move libaah_rtp over from the vendor directory.

    Also move factor PipeEvent out into utils.

    Change-Id: Id3877c66efe22d771cf3ef4877107e431b828e37

Change-Id: I5fe1ea941c09204d7b33f15f4e2b2ab320dc468b
Signed-off-by: Mike J. Chen <mjchen@google.com>
Signed-off-by: John Grossman <johngro@google.com>
Signed-off-by: Jason Simmons <jsimmons@google.com>
12 years agoAdd a missing NULL check.
John Grossman [Mon, 29 Aug 2011 17:26:47 +0000 (10:26 -0700)]
Add a missing NULL check.

The get_next_write_timestamp method introduced to the audio HAL is optional.
HALs which do not implement it leave it set to NULL.  Callers (there is
currently only one in the AudioMixer code) need to be certain to check for NULL
before invoking it.

Change-Id: I88ba43bb53bec081c98c9a8842936c4fbfdd44f6

12 years agoMedia framework changes for Tungsten.
Mike J. Chen [Mon, 15 Aug 2011 20:28:26 +0000 (13:28 -0700)]
Media framework changes for Tungsten.

Squashed merge from master-tungsten of the following changes:

commit 73d09e18c4557e583a1684d44d598a1a02fd0cf2
Author: John Grossman <johngro@google.com>
Date:   Mon Jun 20 13:57:44 2011 -0700

    Remove TungstenMisc and rename LinearTransform

    Change-Id: Ie8aa3e24e09fdbf6ef8996c26deb9c5640e20d1b

commit 3114aabe76ad733b59929d87e49c68229f5ae2e8
Author: John Grossman <johngro@google.com>
Date:   Fri Jun 3 10:47:16 2011 -0700

    Name changes and spelling fixes.

    + Replace the term TungstenTime with the Eugene-approved term CommonTime.
    + Fix a spelling error in a comment I noticed.

    Change-Id: I8c10d618206826d16055f78c7724e24443bb03fd

commit cbf2903ab6893b6e662514e2f6d670e268a419df
Author: John Grossman <johngro@google.com>
Date:   Fri Apr 15 09:27:54 2011 -0700

    Migrate Tungsten code from the HC-Tungsten to the Master-Tungsten branch.

    Change-Id: I95372d913a0761d90168edb4016f5ece0ea74502

commit bc7c46aa629f9883e959ef23de8da297f9eb508b
Author: Jason Simmons <jsimmons@google.com>
Date:   Mon Jun 20 13:59:17 2011 -0700

    Create a separate class for timed AudioTracks

commit 43be3231034ff8537fdd84422a7954780038671f
Author: John Grossman <johngro@google.com>
Date:   Mon Jun 27 18:59:12 2011 -0700

    Move libaah_rtp over from the vendor directory.

    Also move factor PipeEvent out into utils.

    Change-Id: Id3877c66efe22d771cf3ef4877107e431b828e37

commit 17526eb3148c9c3d4365b6d5b47e8dc13bca71b6
Author: John Grossman <johngro@google.com>
Date:   Mon Jun 27 17:06:49 2011 -0700

    Name changes for the TRTP Players s/tungsten/aah/g

    Change-Id: I55e9ad13003f6aa6a36955b54426a7efbe31ac51

commit 423fc1bfc0fda799c421a650c83c4b9293b1a08c
Author: Jason Simmons <jsimmons@google.com>
Date:   Mon Jun 20 17:56:09 2011 -0700

    More timed AudioFlinger changes requested by code review:
    * change trimTimedBufferQueue to trimTimedBufferQueue_l
    * create one timed audio buffer heap per client process instead of one per track
    * grow the silence buffer on demand
    * some error handling fixes in timed getNextBuffer
    * calculate the next output PTS in all mixer and track hooks

    Change-Id: Ifc51a08b55029b7c48902ab2f22933ad7bafe1ad

commit a148e2674b1d3cb73289b82b85c333f0a66824a9
Author: John Grossman <johngro@google.com>
Date:   Mon Jun 20 17:02:24 2011 -0700

    Move the A@H time service into frameworks/base

    Change-Id: I5c570cde70e8931e205516cb33517585804ce841

commit dfa438fa49bdaeeb2ec5fd0d17b30d881608b6b1
Author: John Grossman <johngro@google.com>
Date:   Mon Jun 20 11:55:36 2011 -0700

    Fix the build after Mike's code moving.

    Change-Id: Ia883643ded252168bcc5a70584ab6ce97bb05266

commit 04489474ec8e73efe1bf52918831f41659033162
Author: John Grossman <johngro@google.com>
Date:   Fri Jun 17 14:19:50 2011 -0700

    Refactor the local/common clock services.

    This change is one of a set of 5 changes made to different repositories.  Look
    for this comment in all of them.

    Refactor the local/common clock services in tungsten to match android best
    practice.  Notable changes include

    + The kernel no longer knows anything about common time.  Common time has been
      moved completely up into user land.  This has an impact on the accuracy of the
      timesync debugging code, and the netfilter assisted approach to network based
      timesync is going to have to be modified.
    + The timesync driver used by A@H is now just local time driver.
    + The kernel no longer needs access to the linear transform math code, and it
      has been removed.
    + A new HAL has been introduced to expose the concept of local time to the
      system.
    + A non-slewable stub implementation of the local time HAL based on
      CLOCK_MONOTONIC has been added.
    + The TungstenTime library has been eliminated.  Its functionality has been
      distributed among the common time binder service, the local time hal and the
      linear transform utility code.
    + All clients of the old TungstenTime library have been changed to be clients of
      the binder service, the hal and the utility code.
    + The reset_tt utilities have been removed, they no longer have a purpose in the
      system.
    + more progress has been made in eliminating the word "tungsten" from the code.

    Things left to do include
    + Finish getting rid of tungsten from the time service.
    + Move the time service into the framework; AudioFlinger's new timed mode
      depends on it and the service cannot continue to live in vendor tungsten.

    Change-Id: I999b6cfb4a9d267818a86d747c35eecfc6693101

commit d48194545eed1116a84d81e2fb53315d2b0701a7
Author: Jason Simmons <jsimmons@google.com>
Date:   Thu Jun 16 14:22:46 2011 -0700

    Change the interface of the AudioMixer and AudioBufferProvider to accept a presentation timestamp

    Change-Id: Ice2df5628d45a7f77100e7008103b35b3d3160a4

commit 02561419db82b01ffb28df38000716c612988427
Author: John Grossman <johngro@google.com>
Date:   Tue May 10 14:00:21 2011 -0700

    Put in a hack for controling master volume in the policy manager.
    Fix initial master volume reporting.

    Change-Id: Ia6caf2bbc6083c5f99fab852baa40fff10fc5fc7

commit 549cdc3ba115dc654cdade261fb055c72c6cdb79
Author: John Grossman <johngro@google.com>
Date:   Wed May 4 11:46:17 2011 -0700

    Make certain the logic for computing the output stream mixing point is hardened
    against underflow and overflow when input and output sample rates don't match.

    Change-Id: I5ebea07c9938107b435bec7413418622767e4e16

commit 8043d8ed63f51e76d452d22be7d453d4a7794530
Author: Jason Simmons <jsimmons@google.com>
Date:   Wed Apr 27 18:06:27 2011 -0700

    Add the patch for timed audio support to the mono resampler

    Change-Id: I526f34ae9d1e8e3b0ed2fb05af3d024d5c5fe711

commit 2be89486ef23f0b0b0cc2dc25a4c0ee691043f00
Author: John Grossman <johngro@google.com>
Date:   Wed Apr 27 10:38:57 2011 -0700

    Extend the AudioHWInterface to allow it to specify the initial master volume used by AudioFlinger.

    Change-Id: I8823330801c927494cf7ca31a6b8f9264fbfbb26

commit ff89a4d5e37e6a05a2b03f79ab4e97833dd66393
Author: John Grossman <johngro@google.com>
Date:   Wed Apr 27 09:07:14 2011 -0700

    Fix an issue with inconsistent volume reporting.

    Changed masterVolume() to return the same value as the last call
    to setMasterVolume when the HW layer is implementing master
    volume control.  The masterVolume/setMasterVolume API seems to be
    an idea which was abandonded a long time ago; as of today the
    system only ever sets it to 1.0 at startup and then never changes
    it.  Until we can figure out how the concept of external
    amplifier gain control fits into the Android audio framework,
    Tungsten is exposing this API via a hack-tastic invoke back door
    in the TungstenRXPlayer and needs the getter/setter results to be
    consistent.

    Change-Id: I2ac730fa8fc9ee28c88f1a8e6f2e493eb5b65544

commit 086511b2d19cceb976747ac23e12b73fc7c28bea
Author: Jason Simmons <jsimmons@google.com>
Date:   Mon Apr 25 16:07:19 2011 -0700

    Add handling of timed audio tracks in the generic resampling mixer

    Change-Id: Ic3be1d21b1117f1b233808be543c28a0dcec4792

Change-Id: I6ec5d2bca9b8ebc0acd395a7dd92e1a48fcdfa9b
Signed-off-by: Mike J. Chen <mjchen@google.com>
Signed-off-by: John Grossman <johngro@google.com>
Signed-off-by: Jason Simmons <jsimmons@google.com>
12 years agoam 08479cee: Merge "Stagefright: ANW::connect in MediaPlayerService" into ics-mr0
Jamie Gennis [Fri, 28 Oct 2011 02:14:01 +0000 (02:14 +0000)]
am 08479cee: Merge "Stagefright: ANW::connect in MediaPlayerService" into ics-mr0

* commit '08479ceeba56c460fb52f60a24df27776f1936c3':
  Stagefright: ANW::connect in MediaPlayerService

12 years agoDo not inline the audio resampler assembly functions
Jason Simmons [Thu, 27 Oct 2011 23:38:29 +0000 (16:38 -0700)]
Do not inline the audio resampler assembly functions

The assembly expects arguments to live at fixed offsets from the stack pointer
which are invalid if the code is inlined.

Change-Id: I423b3993ac9cd8d62129718d84fd23efbbac174a

12 years agoMerge "Report call status from AudioSource::start() to file writer" into ics-mr1
James Dong [Thu, 27 Oct 2011 22:24:28 +0000 (15:24 -0700)]
Merge "Report call status from AudioSource::start() to file writer" into ics-mr1

12 years agoUpdate camera continuous autofocus javadoc.
Wu-cheng Li [Thu, 27 Oct 2011 10:07:01 +0000 (18:07 +0800)]
Update camera continuous autofocus javadoc.

Suppose applications call autoFocus in CAF picture mode. If
CAF is in the middle of scanning, the picture is very likely to
be blurry. Change focus callback to return when the scanning
finishes.

bug:5514415
Change-Id: Ibcb8f92a5263d7dbd7cce54df3617fb21c6255d4

12 years agoMerge "Fix a reference to a stale pointer in AwesomePlayer." into ics-mr1
Andreas Huber [Thu, 27 Oct 2011 16:07:23 +0000 (09:07 -0700)]
Merge "Fix a reference to a stale pointer in AwesomePlayer." into ics-mr1

12 years agoFix a hang issue where the first frame can be dropped for timelapse video recording.
James Dong [Thu, 27 Oct 2011 06:47:55 +0000 (23:47 -0700)]
Fix a hang issue where the first frame can be dropped for timelapse video recording.

Change-Id: I04754f1005a983953a80a659ff13f8762d0e120c
related-to-bug: 5523502

12 years agoReport call status from AudioSource::start() to file writer
James Dong [Thu, 27 Oct 2011 06:32:13 +0000 (23:32 -0700)]
Report call status from AudioSource::start() to file writer

Change-Id: I5109dc87a92170dede4eaa58b039efe6d8b7c330
related-to-bug: 5514288

12 years agoStagefright: ANW::connect in MediaPlayerService
Jamie Gennis [Thu, 27 Oct 2011 01:36:31 +0000 (18:36 -0700)]
Stagefright: ANW::connect in MediaPlayerService

This change moves the ANativeWindow connect and disconnect logic from
MediaPlayer to MediaPlayerService::Client.

Bug: 5502654
Change-Id: Ifc43b98b01ad8f35d62d7ece43110724ec7fda3d

12 years agoProperly report that we're no longer playing after receiving event
Andreas Huber [Wed, 26 Oct 2011 22:23:31 +0000 (15:23 -0700)]
Properly report that we're no longer playing after receiving event

MEDIA_PLAYBACK_COMPLETE or MEDIA_ERROR.

Change-Id: I03eeba751d3728adb33d9c66de4c7444a91696f3

12 years agoundo reset to ics-mr1 until we have a better method
Ed Heyl [Wed, 26 Oct 2011 21:53:24 +0000 (21:53 +0000)]
undo reset to ics-mr1 until we have a better method

12 years agoam b7b4eaaa: Merge "Stagefright: push blank buffers when tearing down" into ics-mr0
Jamie Gennis [Wed, 26 Oct 2011 20:34:38 +0000 (20:34 +0000)]
am b7b4eaaa: Merge "Stagefright: push blank buffers when tearing down" into ics-mr0

* commit 'b7b4eaaa05284634d50d581d924a250f1194f66d':
  Stagefright: push blank buffers when tearing down

12 years agoFix a reference to a stale pointer in AwesomePlayer.
Andreas Huber [Wed, 26 Oct 2011 18:22:40 +0000 (11:22 -0700)]
Fix a reference to a stale pointer in AwesomePlayer.

The pointer returned by MetaData::findCString may become stale if the metadata
object is modified. Make a copy of the returned mime type in order to prevent
MediaPlayer::dump log corruption.

Change-Id: I9077aee36c4316d83b37fb909bf91e6b3f75d5d2

12 years agoreset to ics-mr1, but keep history
Ed Heyl [Wed, 26 Oct 2011 18:50:21 +0000 (18:50 +0000)]
reset to ics-mr1, but keep history

12 years agoam 71f2cf11: (-s ours) Rename (IF_)LOGV(_IF) to (IF_)ALOGV(_IF) DO NOT MERGE
Steve Block [Wed, 26 Oct 2011 11:48:16 +0000 (04:48 -0700)]
am 71f2cf11: (-s ours) Rename (IF_)LOGV(_IF) to (IF_)ALOGV(_IF)  DO NOT MERGE

* commit '71f2cf116aab893e224056c38ab146bd1538dd3e':
  Rename (IF_)LOGV(_IF) to (IF_)ALOGV(_IF)  DO NOT MERGE

12 years agoRename (IF_)LOGV(_IF) to (IF_)ALOGV(_IF) DO NOT MERGE
Steve Block [Thu, 20 Oct 2011 10:56:00 +0000 (11:56 +0100)]
Rename (IF_)LOGV(_IF) to (IF_)ALOGV(_IF)  DO NOT MERGE

See https://android-git.corp.google.com/g/#/c/143865

Bug: 5449033
Change-Id: I0122812ed6ff6f5b59fe4a43ab8bff0577adde0a

12 years agoStagefright: push blank buffers when tearing down
Jamie Gennis [Tue, 25 Oct 2011 21:50:16 +0000 (14:50 -0700)]
Stagefright: push blank buffers when tearing down

This change makes OMXCodec push RGB 565 buffers filled with black to an
ANativeWindow when tearing down after decoding to protected gralloc
buffers.  This allows the OMX tear down to zero out any protected
buffers that were used without the possibility that the buffer is still
being used by SurfaceFlinger or HWComposer.

Bug: 5483222
Change-Id: I8acedd81a7bb67dfdc2fd15733e3375b6ce8d560

12 years agoam 149a1493: am 23d64420: Merge "Stagefright: idle OMX after ANW errors" into ics-mr0
Dave Burke [Thu, 20 Oct 2011 05:40:28 +0000 (22:40 -0700)]
am 149a1493: am 23d64420: Merge "Stagefright: idle OMX after ANW errors" into ics-mr0

* commit '149a14931e5aa4beb8c4263995f01437a8918465':
  Stagefright: idle OMX after ANW errors

12 years agoam 23d64420: Merge "Stagefright: idle OMX after ANW errors" into ics-mr0
Dave Burke [Thu, 20 Oct 2011 05:38:45 +0000 (22:38 -0700)]
am 23d64420: Merge "Stagefright: idle OMX after ANW errors" into ics-mr0

* commit '23d644202a44383bf008ff86f6faa3ea7e447290':
  Stagefright: idle OMX after ANW errors

12 years agoMerge "Stagefright: idle OMX after ANW errors" into ics-mr0
Dave Burke [Thu, 20 Oct 2011 05:37:18 +0000 (22:37 -0700)]
Merge "Stagefright: idle OMX after ANW errors" into ics-mr0

12 years agoStagefright: idle OMX after ANW errors
Jamie Gennis [Thu, 20 Oct 2011 04:14:13 +0000 (21:14 -0700)]
Stagefright: idle OMX after ANW errors

This change fixes an issue in Stagefright where the state of an OMXCodec
object can get out of sync with the state of the OMX component.  In
particular, if one of the ANativeWindow functions failed and put the
OMXCodec into the ERROR state, this would cause Stagefright to skip
doing the Executing -> Idle transition.  Without this transition the
freeBuffersOnPort call would never be made, and the MediaBuffers would
end up being leaked (which would also leak the Gralloc buffers they
reference).

Bug: 5333695
Change-Id: I85ea0cf92d18e7ef6d35c7d1e2a7b4e2c9745d34

12 years agoam 38cdb606: am f806871f: Fix 5487571: Make sure shutter sound is played when it...
Chih-Chung Chang [Thu, 20 Oct 2011 03:12:03 +0000 (20:12 -0700)]
am 38cdb606: am f806871f: Fix 5487571: Make sure shutter sound is played when it\'s enforced

* commit '38cdb6063059fe5164f8cf93c21ea000fe24f524':
  Fix 5487571: Make sure shutter sound is played when it's enforced

12 years agoam f806871f: Fix 5487571: Make sure shutter sound is played when it\'s enforced
Chih-Chung Chang [Thu, 20 Oct 2011 03:09:13 +0000 (20:09 -0700)]
am f806871f: Fix 5487571: Make sure shutter sound is played when it\'s enforced

* commit 'f806871f3016a8184de4c9dcfbe0b9069dc0d72c':
  Fix 5487571: Make sure shutter sound is played when it's enforced

12 years agoFix 5487571: Make sure shutter sound is played when it's enforced
Chih-Chung Chang [Thu, 20 Oct 2011 02:47:26 +0000 (10:47 +0800)]
Fix 5487571: Make sure shutter sound is played when it's enforced

Change-Id: I0c7adf452217b8308243836b40b1fba8f2029240

12 years agoam c7d92746: am b3cdadb6: Merge "Check whether media recorder client exists before...
James Dong [Thu, 20 Oct 2011 00:59:48 +0000 (17:59 -0700)]
am c7d92746: am b3cdadb6: Merge "Check whether media recorder client exists before dumping." into ics-mr0

* commit 'c7d927466e6314315743e5c98730b3a79c52e892':
  Check whether media recorder client exists before dumping.

12 years agoam b3cdadb6: Merge "Check whether media recorder client exists before dumping." into...
James Dong [Thu, 20 Oct 2011 00:57:58 +0000 (17:57 -0700)]
am b3cdadb6: Merge "Check whether media recorder client exists before dumping." into ics-mr0

* commit 'b3cdadb639027f62c7c1637ca962a70d2d1f3b4d':
  Check whether media recorder client exists before dumping.

12 years agoMerge "Check whether media recorder client exists before dumping." into ics-mr0
James Dong [Thu, 20 Oct 2011 00:55:45 +0000 (17:55 -0700)]
Merge "Check whether media recorder client exists before dumping." into ics-mr0

12 years agoam ee00c054: am 3c86134a: Merge "Fix issue 381905: BassBoostTest CTS tests fail....
Eric Laurent [Wed, 19 Oct 2011 21:04:30 +0000 (14:04 -0700)]
am ee00c054: am 3c86134a: Merge "Fix issue 381905: BassBoostTest CTS tests fail..." into ics-mr0

* commit 'ee00c0541e8f066550dcbd5bc8af257fce8d7944':
  Fix issue 381905: BassBoostTest CTS tests fail...

12 years agoam 3c86134a: Merge "Fix issue 381905: BassBoostTest CTS tests fail..." into ics-mr0
Eric Laurent [Wed, 19 Oct 2011 21:02:10 +0000 (14:02 -0700)]
am 3c86134a: Merge "Fix issue 381905: BassBoostTest CTS tests fail..." into ics-mr0

* commit '3c86134a50618605c86eb9f5f120dbf97826e3e1':
  Fix issue 381905: BassBoostTest CTS tests fail...

12 years agoFix issue 381905: BassBoostTest CTS tests fail...
Eric Laurent [Wed, 19 Oct 2011 18:44:54 +0000 (11:44 -0700)]
Fix issue 381905: BassBoostTest CTS tests fail...

When AudioEffectTest is executed, an Equalizer is created
and enabled on a MediaPlayer session. Effects on the output
mix are therefore suspended.
Then the MediaPlayer is released with the effect still enabled.
In this case, Audioflinger::purgeStaleEffects_l() fails to restore
the suspended effects when the effect attached to the released audio session
is removed.
When subsequent tests are executed on output mix effects, these effects cannot be
enabled as they are still suspended.

Fixed purgeStaleEffects_l() to restore suspended effects if the effect removed is enabled.

Also fixed EffectHandle::disconnect() to only restore suspended effects if the disconnected
handle actually has control over the effect.

Change-Id: I67232e7c34680b0cc01abfd57d5d510a524e5d4f