bcbcb1bdda4f1970830fa4924e3a5f24171ae005 |
|
13-Dec-2017 |
Mikhail Naganov <mnaganov@google.com> |
audiopolicy: Add more tests, fix issues found Added tests that ensure correct initialization of APM. Fixed issues: -- UB in VolumeCurvesForStream::volIndexToDb due to not checking that the pointer is non-null; -- AudioInputDescriptor::open and SwAudioOutputDescriptor::open were not checking that the returned IO handle is valid, this was causing an infinite loop in AudioPolicyManager::selectOutputForMusicEffects, added an assertion. -- Memory leak of VolumeCurves collection in APM. Test: audiopolicy_tests Change-Id: Ia4ecca1dd03b74d7f93720f042da05d5a0c74a6b
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
733ce945741b1fc12cee92d8133c3136b4e6e0e0 |
|
07-Dec-2017 |
Eric Laurent <elaurent@google.com> |
audio policy: fix HAL stream activity ref counting commit f05fc906 did not take into account the specificities of duplicated outputs causing a crash when accessing the profile (null for duplicated outputs) and not counting activity resulting from duplicated output activity. Also account for HW source activity when connected to an output mix. Bug: 70319466 Test: CTS tests for AudioTrack and AudioRecord Test: audio smoke tests Change-Id: Id0364170794302e42212a90e1b40e28c1e7193fe
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
447d53631b7560eace6016b2711a75533339cbcc |
|
08-Dec-2017 |
Eric Laurent <elaurent@google.com> |
Revert "Revert "audio policy: fix regression in A2DP playback"" This reverts commit 460451e5d7b9813926fe8fc17283e77b63adaee7.
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
3974e3b22856ef35dd73760056ff38b7a4062176 |
|
08-Dec-2017 |
Eric Laurent <elaurent@google.com> |
Revert "Revert "audio policy: add open and active count for IO profiles"" This reverts commit 7b0e95327267f08e36b38fdd8fb2624e251b8e97.
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
53b810ead5b726994fe6a28c6fa955e7b8799aaa |
|
11-Dec-2017 |
Eric Laurent <elaurent@google.com> |
audio policy: fix device address passed when opening HAL streams Commit fe231127 caused a regression where the device address is not passed properly to audio HAL by openInputStream() or openOutputStream(). The address must be read from available device descriptors, not supported devices in profiles which do not have the current device address for removable devices. Bug: 70321528 Test: audio smoke tests,CTS AudioRecordTest Change-Id: I83211a31f86391b80c3c244b436a2e36923ccea0
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
7b0e95327267f08e36b38fdd8fb2624e251b8e97 |
|
07-Dec-2017 |
Eric Laurent <elaurent@google.com> |
Revert "audio policy: add open and active count for IO profiles" Bug: 70319466 This reverts commit f05fc9061254a3329bf0a3b40e27cac87b51b4cc.
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
460451e5d7b9813926fe8fc17283e77b63adaee7 |
|
07-Dec-2017 |
Eric Laurent <elaurent@google.com> |
Revert "audio policy: fix regression in A2DP playback" Bug: 70319466 This reverts commit 995ba712ed77e36129485f5ac811cf328d0a27f6.
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
995ba712ed77e36129485f5ac811cf328d0a27f6 |
|
06-Dec-2017 |
Eric Laurent <elaurent@google.com> |
audio policy: fix regression in A2DP playback commit f05fc906 introduced a regression preventing playback over A2DP to resume after the second time an A2DP device is connected. Bug: 70245910 Test: Audio smoke test, CTS test for AudioTrack. Change-Id: Ic6b5da96debf62be08ae12efec27ad36854fe3c5
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
f05fc9061254a3329bf0a3b40e27cac87b51b4cc |
|
21-Nov-2017 |
Eric Laurent <elaurent@google.com> |
audio policy: add open and active count for IO profiles Remove the hardcoded limit of one open output stream per IO profile and add a max open count per IO profile. Same thing for maximum number of active input streams per IO profile. Now both the maximum number of opened streams and active streams for a given profile is stored in that profile. For now only defaults are set corresponding to legacy behavior, but these values will be loaded form the audio policy configuration file. Test: AudioTrack and AudioRecord CTS tests Change-Id: I5ebcc7f8a06b3f0a52815d241c561bb65e036026
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
fe23112742eeb6199c3aa1f3dfcf800fc2aa0e26 |
|
18-Nov-2017 |
Eric Laurent <elaurent@google.com> |
audio policy: refactor output and input opening Refactor the way input and output streams are open so that common logic is centralized in AudioOutputDescriptor and AudioInputDescriptor classes. This is in preparation of adding reference counting for ouputs and inputs opened for a given profile. Test: Manual Change-Id: Id806e23077eb41f6398400ab40aeaa694c7d5603
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
bb9480975322b0d1a09f76ef14f3285fc8af4e5f |
|
24-Jan-2017 |
Eric Laurent <elaurent@google.com> |
audio: enable concurrent capture for voice recognition Add AUDIO_SOURCE_VOICE_RECOGNITION to the list of use cases allowed for concurrent capture. As AUDIO_SOURCE_HOTWORD is enabled for concurrent capture it makes sense to enable a client to transition to voice recognition after the hotword has been detected in a concurrent capture situation. Also fixed some concurrent capture scenarii Bug: 22702906 Test: build and check "OK Google" detection Change-Id: Ic6dd41372fe74122395764beca7ca9a65197464b
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
f7c50104b07091880019f4ecc70d977961148268 |
|
01-Oct-2016 |
Eric Laurent <elaurent@google.com> |
resolve merge conflicts of a3de77f to master Change-Id: Idb3346465ee7f47d95fd24e2dafdd66f3759c1be
|
271a93e924f46ccaf3f76a8fc38685afea40d3ee |
|
29-Sep-2016 |
Eric Laurent <elaurent@google.com> |
audio policy: fix capture indication to sound trigger service. Only indicate active capture to sound trigger service when capturing from primay HW module. Bug: 31826453 Bug: 31771879 Change-Id: Idf92393d3e7dbf926b847ab23f670e7ab5dd5248
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
fc732a63bd8c8c62ce75f36f03088ca6727caece |
|
30-Mar-2016 |
Glenn Kasten <gkasten@google.com> |
Merge "Use audio_module_handle_t and audio_patch_handle_t consistently" into nyc-dev am: 54b48c3 am: b9b9fa3 * commit 'b9b9fa3d28a0436bb5e267d8fc3626ebd820fe78': Use audio_module_handle_t and audio_patch_handle_t consistently Change-Id: I100c4795ec6d0d9e9c23d6e850303a23f3bfb603
|
a13cde98a880341f0a56d91da6364b093fb5d24e |
|
29-Mar-2016 |
Glenn Kasten <gkasten@google.com> |
Use audio_module_handle_t and audio_patch_handle_t consistently Bug: 27814144 Change-Id: I54b7a73ec318779fd3addb060998f4d94141152d
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
d51b1b247b120c7e49851d923f3c5aa0aa19a787 |
|
27-Feb-2016 |
Jean-Michel Trivi <jmtrivi@google.com> |
resolve merge conflicts of 1e9a65df13 to master Change-Id: I7c255d7813805b5ffb0d8e5ecd6fc11ee0b80c8a
|
8c7cf3b7d0d7bccb9affbc34ae2ab2d6b332f972 |
|
26-Feb-2016 |
Jean-Michel Trivi <jmtrivi@google.com> |
Send audio patch in recording callback Remove direct access to patch handle in AudioInputDescriptor, replace with set / get functions. Add interface for AudioSession to query configuration from enclosing input descriptor. When input descriptor's patch handle changes during a recording, propagate the change through the recording notification callback. Bug 26798796 Change-Id: Ia1395d44da7ccfe69ca46877638d1eb49b3935ad
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
c8d25af1c3c32b7671dbeb967fb86f24fa5fa313 |
|
24-Feb-2016 |
Jean-Michel Trivi <jmtrivi@google.com> |
Merge "Send client / device format in recording callback" into nyc-dev am: f70077e435 am: bce5dfdac0 * commit 'bce5dfdac0e90701f53e4f8a2c0f5a3c44a0a4ac': Send client / device format in recording callback
|
7281aa9810b33eff47b00104db26c97c77931611 |
|
18-Feb-2016 |
Jean-Michel Trivi <jmtrivi@google.com> |
Send client / device format in recording callback Bug 26798796 Change-Id: Ib4f64159c8371d87da49a8f0cad9317a87d843c4
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
232f26f8b673973ffa463f80b08b731ffacc8785 |
|
18-Feb-2016 |
Eric Laurent <elaurent@google.com> |
DO NOT MERGE - audio policy: disable concurrent capture This reverts commit 4342335fb84bfa4ec5b99e3ef82f186820c2275f. This reverts commit 56afc7a5cc0f0bfc8021f6413b2e2267050bdc8a. This reverts commit 313d1e7dfb4bf8e0a2f09a9adb287062dd64410c. This reverts commit fb66dd9f95a1168698f072c4e5c2a2cf8c49a80f. Bug: 22702906 Change-Id: I9085918f1f7b12e28430958cceb768bb29c12bdb
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
56afc7a5cc0f0bfc8021f6413b2e2267050bdc8a |
|
03-Feb-2016 |
Jean-Michel Trivi <jmtrivi@google.com> |
Fix connection of OUT_REMOTE_SUBMIX device on APM startInput Fix a regression in AudioPolicyManager::startInput() where the OUT_REMOTE_SUBMIX audio device was never made available due to the test on the input descriptor, which was already active by the time it is tested, as the activity on the session has already been updated. Add methods to AudioInputDescriptor and AudioSessionCollection to query the number of active sessions. Use the new method to optimize AudioSessionCollection::hasActiveSession() Bug 26798796 Change-Id: I2de4e47970505e64c4c78a148a80a9f0ec19e821
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
fb66dd9f95a1168698f072c4e5c2a2cf8c49a80f |
|
29-Jan-2016 |
Eric Laurent <elaurent@google.com> |
Revert "Revert "audio policy for concurrent capture."" This reverts commit 84332aaa807037baca05340875f2d94fcc519ac4. Bug:26841909
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
84332aaa807037baca05340875f2d94fcc519ac4 |
|
28-Jan-2016 |
Eric Laurent <elaurent@google.com> |
Revert "audio policy for concurrent capture." This reverts commit 4b2fcd8abf90b1bd95539f9f4ef2ac39ce9e2df1. Bug:26841909 Change-Id: Ifc74b195394b9fb3f73d33455ad061bd896d5331
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
4b2fcd8abf90b1bd95539f9f4ef2ac39ce9e2df1 |
|
16-Jan-2016 |
Eric Laurent <elaurent@google.com> |
audio policy for concurrent capture. First implementation of concurrent capture policy: Initially, only capture for HOTWORD or FM TUNER sources or re-route submix is allowed at the same time as any other capture. Several capture clients can share the same input stream. In this case active clients are sorted by audio source priority level and the input path configuration and routing is driven by the highest priority active source. Bug: 22702906 Change-Id: Icbedfe001c6243ecad7a7b57ab78d893b2b97402
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
112b0af826aeca45855690b9c105b2cdf9938bbe |
|
19-Nov-2015 |
François Gaffie <francois.gaffie@intel.com> |
audio policy: introduction of AudioProfile In order to be able to declare the possible sample specifications for an Audio Port (a given format, a list of sample rates, a list of channels), a new object AudioProfile is introduced. Change-Id: Ie3ca3400f05984e7b4132fc7ba90020eb4d998c7 Signed-off-by: François Gaffie <francois.gaffie@intel.com>
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
599c758b258cc5da0dba9b530425381facc37d77 |
|
08-Dec-2015 |
Eric Laurent <elaurent@google.com> |
audio policy: audio sessions on input descriptors In preparation for concurrent capture, add support for multiple audio sessions per input stream. Each session keeps its own properties, open and active reference counting. No functional change for now: still one session per input and one active input at a time. Bug: 18815985. Bug: 22702906. Change-Id: I915a65989a7fd0d3cbe2fcf5a0aee2ea0df5f4f5
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
65bfe916ca972b95f41d7f6cc2566c2e12a3eada |
|
05-Dec-2015 |
Jean-Michel Trivi <jmtrivi@google.com> |
AudioPolicyManager: AudioInputDescriptor open ref count cleanup Always initialize the mOpenRefCount field. Add functions for management for open ref count. Change-Id: I0bbd021283047abfebbc108ced68c79e29297f25
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
a8ecc2c72ca26389bd6b0162181d60aaeaca8149 |
|
09-Nov-2015 |
François Gaffie <francois.gaffie@intel.com> |
audio policy: remove deserialization from policy objects This patch removes the responsability of the deserialization of all objects managed by the policy by: -adding required accessors -moving deserialization function to Config parser helper -adds a TypeConverter to associate string to / from values of all audio enumeration shared between HALs and policy. Change-Id: I1ce798848f4657b37e47446c9fbdc63f7ed0390e Signed-off-by: François Gaffie <francois.gaffie@intel.com>
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
64265b2fb8f5be63b6c2ad4fcbec9acf74705bc4 |
|
19-Sep-2015 |
Eric Laurent <elaurent@google.com> |
audio policy: fix preemtible capture race Because a preemtible capture session can preempt another one, we end up in an endless loop situation were each session is allowed to restart after being preempted, thus preempting the other one which restarts and so on. To avoid this situation, we store which audio session was preempted when a particular input started and prevent preemption of this active input by this session. We also inherit sessions from the preempted input to avoid a 3 way preemption loop etc... Bug: 24007220. Change-Id: I0eab5299440ef3ab9e987635dc9a300cf42f2c79
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
322b4d25387a04c9afebe998326d005bbdf17ede |
|
04-Apr-2015 |
Eric Laurent <elaurent@google.com> |
audio policy: fix unique audio port ID. Remove mId member from AudioPort as it was shadowed by mId in DeviceDescriptor. Add getters for Id, and HW module name and version to device, output and input descriptors. Fix DeviceDescriptor name initialization. Change-Id: I4a69f385e40330954d9dad5f2926c521f60b2ec1
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
53615e29c99c5e9d2ca77aaefd7bf5c770513120 |
|
19-Mar-2015 |
François Gaffie <francois.gaffie@intel.com> |
Migrate helper functions from managerdefault to common This patch moves from manager to common: -parse helper functions of the policy configuration file -collection helper function on -output / input descriptors -DeviceDescriptor -AudioPatch / Audio Port -IO Profile -HwModule Change-Id: If45e53418db75af1af198f43c4ef27884499055f Signed-off-by: François Gaffie <francois.gaffie@intel.com>
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|
98cc191247388132b6fd8a4ecd07abd6e4c5a0ed |
|
18-Mar-2015 |
François Gaffie <francois.gaffie@intel.com> |
Split ManagerDefault into Manager and ManagerDefinitions This patch splits the managerdefault into a manager and a manager defintion library that contains all pillar elements of a policy manager. It renames the file with the name of the main class they contains. It splits the AudioPort into AudioPort and AudioPatch. Change-Id: I992cf0b8aed895805cc003ba0980d2c9e92c985b Signed-off-by: François Gaffie <francois.gaffie@intel.com>
/frameworks/av/services/audiopolicy/common/managerdefinitions/src/AudioInputDescriptor.cpp
|