Camera: Minor Improvements and Fixes #2

dropAccessUnit is uninitialized. The do-while loop which checks for this variable has a continue and hence this variable was being used without initialization. The change initializes this variable to default value. CRs-Fixed: 1087183 Change-Id: I301eb165b51f773b5062283041d6749ccaff9e3e

-if sample rate is not specificed during construction of audio record it updates sampling rate with default value based on the source. but divide by zero execption is happeing since it uses same unspeficied sampling rate to calculate latency. -Use updated default sample rate to calcuate latency vaulue. CRs-Fixed: 1090804 Change-Id: Icb1e717d869de7bcd4b52194f3ee4e8b0b318da4

…av-aosp.lnx.2.0-dev

Changes to set protected flag only for secure contents. Change-Id: I18812ae216c30031fb0ffa42f63ce89b1cd6ab4d

Changes to enable isDrm flag for DRM playback so as to enable setPlaybackStatus. Change-Id: I824f59471286e5454fab7630a45bdbc2430711ac

Bug: 31781965 Change-Id: I1c439a0d0f6aa0057b3c651499f28426e1e1f5e4 (cherry picked from commit 9a2732b)

This was meant for buffers shared cross-process, but we are not gaining anything from this check even if it was at the correct place. Bug: 32436178 Change-Id: I6919e8ac6e35092273e171f49f6711ba577ba2e6 (cherry picked from commit 58388aa)

…dexes Bug: 32247948 Bug: 32438598 Bug: 32436341 Test: use POC on bug or cts security test Change-Id: I91bd6aadb6c7410163e03101f365db767f4cd2a3 (cherry picked from commit 0872b65cff9129633471945431b9a5a28418049c) (cherry picked from commit e981cca)

several points in stagefrights mp3 album art code used strlen() to parse user-supplied strings that may be unterminated, resulting in reading beyond the end of a buffer. This changes the code to use strnlen() for 8-bit encodings and strengthens the parsing of 16-bit encodings similarly. It also reworks how we watch for the end-of-buffer to avoid all over-reads. Bug: 32377688 Test: crafted mp3's w/ good/bad cover art. See what showed in play music Change-Id: Ia9f526d71b21ef6a61acacf616b573753cd21df6 (cherry picked from commit fa0806b)

Test: Custom test. Bug: 32438594 Bug: 32624850 Bug: 32635664 Change-Id: I9b1315e2c02f11bea395bfdcf5c1ccddccbad8a6 (cherry picked from commit 3d34cc7)

Bug: 32577290 Change-Id: I9bcc9422ae7dd3ae4a38df330c9dcd7ac4941ec8 (cherry picked from commit 7fdd364)

Bug: 32448258 Bug: 32095626 Test: Use POC bug or cts security test Change-Id: I69f24eac5866f8d9090fc4c0ebe58c2c297b63df (cherry picked from commit 0118340)

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 1087183 I301eb165b51f773b5062283041d6749ccaff9e3e NuPlayer: Uninitialized bool variable in fetchInputData Change-Id: I5e54964b854979cf9f865dfd5b948e0bfccabfd5 CRs-Fixed: 1087183

Uses user defined permission "com.oma.drm.permission.ACCESS_OMA_DRM" to access protected api call from Oma Drm util framework. This changes is required for DRM Image support in Android. Change-Id: I6811d97e45e2bc37356be9061e1a5a5557e19e06 CRs-Fixed: 1088307

….2.0-dev

Change-Id: I7d5faf8a5ab78eaea184544472486a41a1f1e5dd

…etrieve layer info""

…deo sample""

…emote"

Currently service allows only a single client to be attached to a Module. This limits only a single client can use sound trigger at a time. Add changes to attach multiple clients to a given Module through ModuleClient inteface so that multiple clients can in parallel use sound trigger on a given Module. ModuleClient class is introduced as a client interface to a Module. Service provides a unique instance of ModuleClient to each client being attached and adds this client to the module clients list. Bug:32030191 Change-Id: I0214395ed545690bb9f95365c7b8d415fcf60c9a

…rm engine" into av-aosp.lnx.2.0-dev

…0-dev

…p.lnx.2.0-dev

Change-Id: I12811040dbe6c3dd0abeac07cedc51893f1b4a94

Test: POC, Cts Effect, BassBoost, EnvReverb, Equalizer, Test: LoudnessEnhancer, PresetReverb, Virtualizer, Visualizer Bug: 32220769 Change-Id: Iea96ba0daf71691ee8954cca4ba1c10fe827626e (cherry picked from commit dd79ccd)

Bug: 32705438 Bug: 32703959 Test: cts security test Change-Id: I8900c92fa55b56c4c2c9d721efdbabe6bfc8a4a4 (cherry picked from commit e275907)

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 1090804 Icb1e717d869de7bcd4b52194f3ee4e8b0b318da4 libmedia : Fix for AudioRecord CTS test case 1088307 I6811d97e45e2bc37356be9061e1a5a5557e19e06 OMA DRM : Give permission to access drm service api from 1088780 I824f59471286e5454fab7630a45bdbc2430711ac frameworks/av: Changes to enable DRM flag 1088780 I18812ae216c30031fb0ffa42f63ce89b1cd6ab4d frameworks/av: Set Protected flag only for secure Change-Id: Idcd5a7bfde181fa220d9d11976bd750c00c812dc CRs-Fixed: 1088780, 1088307, 1090804

Proceed with AudioRecord::read only if object state is active. An active AudioRecord instance can become inactive if restoreRecord fails (for example due to concurrency with other AudioRecord instances) CRs-Fixed: 1101547 Bug: 26229717 Change-Id: I2a1d0c32b29e0da3e0c457e43f006388fc7b2433

force audio/video both to start from video last preceding IDR position after suspend/resume or seek operation, to make playback smooth Change-Id: I868da2db100459f2971dcfa2fe9c54295f8c31ec CRs-Fixed: 1077339

This reverts commit 5e24b9c. Change-Id: I282ff86dacedfa3a10e81a86b3aa25e7f5da76b1

Allow effect chain on direct output if it's pcm offload use case. Change-Id: I62335ad16091bc6d318b4ca0c014bd6bf160c18d CRs-Fixed: 1102537

Use the AUDIO_OUTPUT_FLAG_DIRECT only for creating the track and do not add this in the audio_flags of AudioTrack Change-Id: Ic85ad5aef91e147a209d1c7a3a7ef0728f3d4207

…to av-aosp.lnx.2.0-dev

…into av-aosp.lnx.2.0-dev

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 1088780 I282ff86dacedfa3a10e81a86b3aa25e7f5da76b1 Revert "frameworks/av: Set Protected flag only for secur 1098700 I0214395ed545690bb9f95365c7b8d415fcf60c9a soundtrigger: Support multiple clients attaching to a mo 1077339 I868da2db100459f2971dcfa2fe9c54295f8c31ec HLS: force audio/video both to start from IDR position Change-Id: Icee6017e8c0cae800073056a2b9e65fbbffefb47 CRs-Fixed: 1098700, 1088780, 1077339

…to av-aosp.lnx.2.0-dev

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 1100184 Ic85ad5aef91e147a209d1c7a3a7ef0728f3d4207 audio: Don't update the audio_output flag with track fla Change-Id: I526bbe57d272a4b5dda46047146cb4df25594e3c CRs-Fixed: 1100184

- On split a2dp audio is received on remote device for 10secs after triggering pause of offload session - Offload session is tear down only after 10secs which resulted silence audio stream to BT till pause timeout - Reduce pausetimeout to 3 secs to inline split & non-split a2dp behaviour Change-Id: If7698c36ad04e3ddd2da3dfbd1816be9c6c36870

….lnx.2.0-dev

…" into av-aosp.lnx.2.0-dev

Bug: 32448258 Bug: 32095626 Test: Use POC bug or cts security test Change-Id: I69f24eac5866f8d9090fc4c0ebe58c2c297b63df (cherry picked from commit 0118340) (cherry picked from commit 20cd1eaf8094dd9745ce95bd10dfb25525fce1a4)

Bug: 32577290 Change-Id: I9bcc9422ae7dd3ae4a38df330c9dcd7ac4941ec8 (cherry picked from commit 7fdd364) (cherry picked from commit 16e7fa2883a045106deedeb837fa25ef34f4931b)

Test: Custom test. Bug: 32438594 Bug: 32624850 Bug: 32635664 Change-Id: I9b1315e2c02f11bea395bfdcf5c1ccddccbad8a6 (cherry picked from commit 3d34cc7) (cherry picked from commit a1712f7f89f2a7158669c0d5e84016297ee0ef26)

several points in stagefrights mp3 album art code used strlen() to parse user-supplied strings that may be unterminated, resulting in reading beyond the end of a buffer. This changes the code to use strnlen() for 8-bit encodings and strengthens the parsing of 16-bit encodings similarly. It also reworks how we watch for the end-of-buffer to avoid all over-reads. Bug: 32377688 Test: crafted mp3's w/ good/bad cover art. See what showed in play music Change-Id: Ia9f526d71b21ef6a61acacf616b573753cd21df6 (cherry picked from commit fa0806b) (cherry picked from commit cfc7e9af62c1815c8018d93316de6a788b902032)

…dexes Bug: 32247948 Bug: 32438598 Bug: 32436341 Test: use POC on bug or cts security test Change-Id: I91bd6aadb6c7410163e03101f365db767f4cd2a3 (cherry picked from commit 0872b65cff9129633471945431b9a5a28418049c) (cherry picked from commit e981cca) (cherry picked from commit b57c79b5d7ba86f3603f35b60576dbf45b8d0c7b)

This was meant for buffers shared cross-process, but we are not gaining anything from this check even if it was at the correct place. Bug: 32436178 Change-Id: I6919e8ac6e35092273e171f49f6711ba577ba2e6 (cherry picked from commit 58388aa) (cherry picked from commit 12ac1ae407b29533ea05f771607ba885b3f72801)

Bug: 31781965 Change-Id: I1c439a0d0f6aa0057b3c651499f28426e1e1f5e4 (cherry picked from commit 9a2732b) (cherry picked from commit 6e6e99247dd3349e8c23175e1ae024adb8c5ec9d)

Add null check for the source buffer prior to memcpy. Change-Id: I5971e0c01fd4821078780c9623154e396f679357 CRs-Fixed: 1096369

Writer stops the video and audio tracks synchronously. In case video pipeline is running slow, video will be stopped first, resulting in shorter video track in the saved recording. To match durations of video and audio: [a] reverse the order of stopping tracks to stop audio first since audio source stops faster than video. [b] defer stopping video till the video frame-time reaches the time of the last processed audio frame-time. If this event does not happen within 1 second, stop video anyways CRs-Fixed: 1087319 Change-Id: I98c62f927160de10ab2e9068741bd62734ead25c

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 1102537 I62335ad16091bc6d318b4ca0c014bd6bf160c18d audio: allow effect chain on direct output if it's pcm o 1101547 I2a1d0c32b29e0da3e0c457e43f006388fc7b2433 AudioRecord: Check object state on entering read 1090760 If7698c36ad04e3ddd2da3dfbd1816be9c6c36870 audio: Reduce offload pause timeout to 3 secs Change-Id: I3fa29b113e652796c0bb4c90e7d4dd5a51d85830 CRs-Fixed: 1101547, 1102537, 1090760

When audio flinger removes an AudioTrack from the active list in case of underrun, framecount gets reset to 0 in audio track client during obtain buffer. Server side patch track uses same frame count(0) to invoke obtain buffer during retry post restart and it lead to crash. Keep a copy of buffer frame count in patch track obtain buffer and use it during retry. CRs-Fixed: 1091467 Change-Id: I29404941aae904a8c713c40838476c30680a30e5

…e indexes" into av-aosp.lnx.2.0-dev

* changes: Effects: Check get parameter command size Make VBRISeeker more robust

…into av-aosp.lnx.2.0-dev

…tive indexes" into av-aosp.lnx.2.0-dev

…ffer" into av-aosp.lnx.2.0-dev

…-aosp.lnx.2.0-dev

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 1100989 I9b1315e2c02f11bea395bfdcf5c1ccddccbad8a6 Effects: Check get parameter command size 1100993 I69f24eac5866f8d9090fc4c0ebe58c2c297b63df Fix security vulnerability: Effect command might allow n 1096369 I5971e0c01fd4821078780c9623154e396f679357 libstagefright: Add NULL check during memcpy for MediaCo 1100991 I91bd6aadb6c7410163e03101f365db767f4cd2a3 Fix security vulnerability: Equalizer command might allo 1100988 Ia9f526d71b21ef6a61acacf616b573753cd21df6 DO NOT MERGE: defensive parsing of mp3 album art informa 1100996 I6919e8ac6e35092273e171f49f6711ba577ba2e6 stagefright: remove allottedSize equality check in IOMX: 1100979 I9bcc9422ae7dd3ae4a38df330c9dcd7ac4941ec8 Make VBRISeeker more robust 1100986 I1c439a0d0f6aa0057b3c651499f28426e1e1f5e4 Visualizer: Check capture size and latency parameters Change-Id: I84a02ba7a366609d227894a656dfa53b82d744bb CRs-Fixed: 1100988, 1100989, 1096369, 1100993, 1100991, 1100996, 1100986, 1100979

Change-Id: Ic5ad9990def89e812e19c512399e96ee10972d88

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- I69f24eac5866f8d9090fc4c0ebe58c2c297b63df Fix security vulnerability: Effect command might allow n I9b1315e2c02f11bea395bfdcf5c1ccddccbad8a6 Effects: Check get parameter command size I6919e8ac6e35092273e171f49f6711ba577ba2e6 stagefright: remove allottedSize equality check in IOMX: I69eb18abe191ae0f1e76f8b62daaf5717cce9669 Camera API1: don't pick wrong preview fps value Ia9f526d71b21ef6a61acacf616b573753cd21df6 DO NOT MERGE: defensive parsing of mp3 album art informa I9bcc9422ae7dd3ae4a38df330c9dcd7ac4941ec8 Make VBRISeeker more robust I1c439a0d0f6aa0057b3c651499f28426e1e1f5e4 Visualizer: Check capture size and latency parameters I91bd6aadb6c7410163e03101f365db767f4cd2a3 Fix security vulnerability: Equalizer command might allo Change-Id: Ib0defe68120a7c147d1e7fb8360ebc011bc990a2

There could be a corner case when direct track paused just after started before rendering anything and HAL is still in standby. So this active track can be added to tracksToRemove by two times, then it is started again before threadloop_removeTracks() called. And finally when threadloop_removeTracks() is invoked, it will trigger two times stopOutput and makes the active count of this stream to 0. Remove the redundant adding where it can also be added when track is paused and all frames are presented out. Change-Id: I9f7575a677a25a348dce750f8b338f45646d309c CRs-Fixed: 1091752

…spended Fastmixer will enter COLD_IDLE state after the output is suspended. If it still has active tracks and may be removed later, the state command sent to fastmixer could be blocked and thus block subsequent calls to this output. Put fastmixer into MIX state instead of COLD_IDLE when it's suspended but still has active tracks to be processed. Change-Id: I7f24f1a3e5b1651001fcef3ba3d00a8fb67e2b79 CRs-Fixed: 1093067

There is a loss of a/v sync during camcorder recording with QC HW AAC encoder when pause/resume operations are done continuously. This is because the encoder does not consider the input timestamp to calculate the timestamp of an encoded frame. For each output buffer, the timestamp is just the number of frames encoded so far times the frame duration. However, as the input timestamp is used by video encoder, there is a drift between the audio and video timestamps after resume which results in a/v sync loss. Avoid the default 30ms offset in the total paused duration when QC AAC encoder is used to ensure that the timestamps of audio and video are in sync. Also, calculate the TotalPausedDuration value considering the capturerate and framerate. Change-Id: I8420e44ab96484f0d6301c366a24eefc8efeaf0f CRs-Fixed: 1070540 Change-Id: Ia201f3df011c61ce8ed47c3ccb7a2e6a457c2cef

Direct output can be paused when track underrun is observed for the first time, so this can interrupt the playback even when there is PCM data cached in the driver. Also the track may be ready before the next prepareTracks, the playback should be continued. To avoid the unnecessary pause, only do HAL pausing when track is going to be removed due to buffer timeout. Change-Id: I3707d883d0daa4ffd2baa9bd69aebf4bf9ddef4e CRs-Fixed: 1094472

getTrackMetaData will return NULL for failure. If we pass the NULL, mediaserver application will crash. Check trackMeta and go to next index if it is NULL. CRs-Fixed: 1115425 Change-Id: I55a73cb0f327b6af31d30ed1b3d27c83578762d3

…to av-aosp.lnx.2.0-dev

…ck" into av-aosp.lnx.2.0-dev

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 1070540 Ia201f3df011c61ce8ed47c3ccb7a2e6a457c2cef StagefrightRecorder: avoid 30ms offset with QC AAC encod 1091467 I29404941aae904a8c713c40838476c30680a30e5 audio flinger: Fix native crash during underruns for pat 1093067 I7f24f1a3e5b1651001fcef3ba3d00a8fb67e2b79 audioflinger: put fastmixer with active tracks into mix 1094472 I3707d883d0daa4ffd2baa9bd69aebf4bf9ddef4e audioflinger: pause direct output when buffer timeout 1091752 I9f7575a677a25a348dce750f8b338f45646d309c audioflinger: remove redundant adding to tracksToRemove Change-Id: If6ae67ee40d6f103c00726a6df651e9732872dae CRs-Fixed: 1070540, 1091467, 1094472, 1091752, 1093067

-define mime types for QT format Change-Id: I4170a4019211a4a973ee4bcc79af80585082b34d

… time" into av-aosp.lnx.2.0-dev

When EIS-3.0-enabled session is stopped, video frames worth 1 sec (up to 30 frames) have to be encoded and muxed. This operation can take up to 1 sec and increase the stop- latency for back-to-back recording. Optimize this by notifying perf-hints to video and camera. CRs-Fixed: 1087319 Change-Id: Ic05edd3e5bfd5b65fec8f48ad45a6e15cd2a437e

unsigned int was getting assigned to a negative value, which in turn was leading to a crash in htonl. Fixing integer overflow in writeCttsBox. Change-Id: I2c4e8efcd1a1b215c412c9e26637f0b9b563a781

Existing architecture chooses CameraSourceTimeLapse for High Speed Recording where timestamps are manipulated based on frame rate. For HSR, Capture rate and Frame rate are equal and no need of timestamps manipulation. Hence create CameraSource. Change-Id: I7f420f5b15fb3c05bb7f918430ca9b7a630ed18e CRs-Fixed: 1091389

When releasing input, check whether the audio session retrieved is null rather than checking the value of index which was already verified earlier on. CRs-Fixed: 1114286 Change-Id: I31901b9ca9de9f3a561c9dbcc0a1d39ecbd06744

…-dev

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 1091389 I7f420f5b15fb3c05bb7f918430ca9b7a630ed18e Stagefright: Create CameraSource for HSR 1087319 Ic05edd3e5bfd5b65fec8f48ad45a6e15cd2a437e stagefright: Optimize stop-delay for EIS enabled recordi 1113404 I2c4e8efcd1a1b215c412c9e26637f0b9b563a781 stagefright: MPEG4Writer: fix Integer overflow 1087319 I98c62f927160de10ab2e9068741bd62734ead25c stagefright: MPEG4Writer: defer video stop to match last Change-Id: I82b1f19fb0cb8d8c9daf000b63510420864e789a CRs-Fixed: 1087319, 1113404, 1091389

Instantiate vraudioservice from within audioserver Change-Id: Icb8f0a892fb0e29dd69a19709aff80471422e37a

Currently mAnchorTimeMediaUs is not getting updated in case of new audio time is not greter than kMinimumAudioClockUpdatePeriodUs, but due to this for very small duration clips the progress bar does not go till end, as the mMaxTimeMediaUs of mediaClock is not updated correctly on newAudiotime. Make sure the mAnchorTimeMediaUs updated irrespective of new audio time is greater than kMinimumAudioClockUpdatePeriodUs or not so that mediaClock returns the correct timestamp having correct mMaxTimeMediaUs Change-Id: If67cd144ea633222f62ae60a7251cc32ac80df33

…put" into av-aosp.lnx.2.0-dev

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 1114286 I31901b9ca9de9f3a561c9dbcc0a1d39ecbd06744 AudioPolicyMng: check for null audioSession when realeas Change-Id: Ia3a0b9328c9d8d7f8e70a909123df3cb32bcc099 CRs-Fixed: 1114286

…2.0-dev

…-dev

Null check on mDevice before calling stopPreviewL. Its already checked in ::disconnect(). Add the check in ::stopPreview() as well Change-Id: I489303c21176c108c92538f8cf30720fc4312082

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 1090786 I4170a4019211a4a973ee4bcc79af80585082b34d Video: add support for playback of QT clips 1089231 I489303c21176c108c92538f8cf30720fc4312082 Camera: Add null check on mDevice to avoid crash 1115425 I55a73cb0f327b6af31d30ed1b3d27c83578762d3 libstagefright: Check trackMeta for NULL Change-Id: I687cdd27b58bce3a35c0d35783c4b31cbd42b6e8 CRs-Fixed: 1115425, 1089231, 1090786

Test: POC, Cts Effect, BassBoost, EnvReverb, Equalizer, Test: LoudnessEnhancer, PresetReverb, Virtualizer, Visualizer Bug: 32220769 Change-Id: Iea96ba0daf71691ee8954cca4ba1c10fe827626e (cherry picked from commit dd79ccd) (cherry picked from commit a155de4)

Bug: 32705438 Bug: 32703959 Test: cts security test Change-Id: I8900c92fa55b56c4c2c9d721efdbabe6bfc8a4a4 (cherry picked from commit e275907) (cherry picked from commit b0bcddb)

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 1111399 Icb8f0a892fb0e29dd69a19709aff80471422e37a vraudioservice: Instantiation Change-Id: Iab88f0f6d24d88bfc988baee28e599a302cf63f4 CRs-Fixed: 1111399

This reverts commit 052c495. The original change is device specific removing from common frameworks location Change-Id: Ie80e8aae4afb9f3f2655680600553cf26d6f63f3

Due to I2c21d4c1788777c2d0d77227bb872701b35c4ff6, any background thread only allows RT bandwidth of just 10msec. This causes RT throttling error when CAM thread is being raised to RT task status because the msm-adhoc bus mutex is an RT mutex. To avoid this issue, we disable the background scheduling of cameraserver caused by binder call. Change-Id: I16ddbd832a606d136575c3d343c63ebc8d161c5b

…rding For HSR recording, there is a loss of a/v sync when pause/resume operations are done continuously. mCaptureFpsEnable is true for HSR and 30ms is deducted from the total paused duration which causes a/v sync issues when QC AAC encoder is used (check Ia201f3df011c61ce8ed47c3ccb7a2e6a457c2cef). Remove the check for mCaptureFpsEnable to avoid the 30ms offset in total paused duration to fix the issue. CRs-Fixed: 2006474 Change-Id: I98dc605a3c535ee0ff757deb1caea7b9e29a3297

Change-Id: Iec5cbbd5d129e1e3342c6ad166b45eff7ee92aef

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- I8900c92fa55b56c4c2c9d721efdbabe6bfc8a4a4 Fix security vulnerability: potential OOB write in audio 1108482 2004819 Ie80e8aae4afb9f3f2655680600553cf26d6f63f3 Revert "update stune group for cameraserver to top-app" 1111985 I8900c92fa55b56c4c2c9d721efdbabe6bfc8a4a4 Fix security vulnerability: potential OOB write in audio 1111988 Iea96ba0daf71691ee8954cca4ba1c10fe827626e Effect: Use local cached data for Effect commit Iea96ba0daf71691ee8954cca4ba1c10fe827626e Effect: Use local cached data for Effect commit Change-Id: I4f02c49c9e7bcf104475a44022b8f3e5fc3fe410 CRs-Fixed: 1111988, 2004819, 1111985, 1108482

In batch-mode, multiple buffers are sent together as a batch from camera to encoder. After pause/resume, pause duration gets subtracted from the time stamps before queueing to the encoder. Batch mode needs updated base time stamp to calculate the timestamps of the buffers in the batch. Hence update the time stamp offsetted with pause duration in buffer meta data. CRs-Fixed: 1106432 Change-Id: I6cdfff4f0f7a25c7b4156382a224688bdda03d28

Avoid waiting for video track to catch up to last audio timestamp for video only recording. Change-Id: I370c61f563a4eaf931e9f43fc139df329fac6fb1

…SR recording"

Enable bytestream mode in extractor only when playback is through MediaPlayer APIs instead of enabling it by default. CRs-Fixed: 2005045 Change-Id: I74c2236fa70ae60da8fbcb1fbabb5a66b43b272c

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 1106432 I6cdfff4f0f7a25c7b4156382a224688bdda03d28 stagefright: correct timestamp in buffer metadata 2006474 I98dc605a3c535ee0ff757deb1caea7b9e29a3297 StagefrightRecorder: avoid 30ms offset in pause duration 2007049 I370c61f563a4eaf931e9f43fc139df329fac6fb1 video: Skip wait for audio for video only recording 1115250 I16ddbd832a606d136575c3d343c63ebc8d161c5b cameraserver: disable background scheduling of cameraser Change-Id: Id5b6857a4430e313f733f8b574fba8424f371862 CRs-Fixed: 2007049, 2006474, 1106432, 1115250

Due to I2c21d4c1788777c2d0d77227bb872701b35c4ff6, any background thread only allows RT bandwidth of just 10msec. This causes RT throttling error when CAM thread is being raised to RT task status because the msm-adhoc bus mutex is an RT mutex. To avoid this issue, we disable the background scheduling of cameraserver caused by binder call. Change-Id: I16ddbd832a606d136575c3d343c63ebc8d161c5b

Avoid waiting for video track to catch up to last audio timestamp for video only recording. Change-Id: I370c61f563a4eaf931e9f43fc139df329fac6fb1

The interface getCurrentPosition() should be called without mutex held from any threads now, since fillAudioBuffer() has already locked the mutex at the begining, recursive access can cause deadlock. Use the interface of MediaClock instead can avoid this. CRs-Fixed: 2014325 Change-Id: Iefdd061d745f548a207dd96f15c323fa7900d40d

Disable sync frame decoder mode for closest seek mode, to ensure that the target frame can be decoded. Change-Id: Ib39fee7bb796c967639aa85be46fd913796041f6

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 2005045 I74c2236fa70ae60da8fbcb1fbabb5a66b43b272c frameworks/av: enable bytestream mode in extractor selec Change-Id: I57112f7ec9548cc0884030f2fa36497e2a8eae90 CRs-Fixed: 2005045

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 2006334 If67cd144ea633222f62ae60a7251cc32ac80df33 NuPlayerRenderer: update mAnchorTimeMediaUs on new audio Change-Id: Ia6790ecc52d5bde230ca6d94ac2ce9e030b8902f CRs-Fixed: 2006334

NULL check on mDevice before calling cancelAutoFocus. Also adding the same check in disconnect. CRs-Fixed: 2000357 1082919 Change-Id: If57e12913bebfca6ab9b871a0bdcc24d663e7107

Remove extra memcpy which was added due to improper cleanup of merge conflicts while aosp upgrade Change-Id: Ibd61ea227138449ce5e4e0ec548863b02e1f4ac5

to avoid leaving them in a partially initialized state. Bug: 33137046 Test: ran CTS tests Change-Id: I1f5c070233c5917d85da9e930e01a3fc51a0a0ec (cherry picked from commit a9660fe) (cherry picked from commit bc62c086e9ba7530723dc8874b83159f4d77d976)

- Reorganize handle effect creation code to make sure the effect engine is created with both thread and effect chain mutex held. - Reorganize handle disconnect code to make sure the effect engine is released with both thread and effect chain mutex held. - Protect IEffect interface methods in EffectHande with a Mutex. - Only pin effect if the session was acquired first. - Do not use strong pointer to EffectModule in EffectHandles: only the EffectChain has a single strong reference to the EffectModule. Bug: 32707507 Change-Id: Ia1098cba2cd32cc2d1c9dfdff4adc2388dfed80e (cherry picked from commit b378b73) (cherry picked from commit 22e26d8ee73488c58ba3e7928e5da155151abfd0)

After AOSP security fix Ia1098cba2cd32cc2d1c9dfdff4adc2388dfed80e need to make sure that effects are applied on PCM offloaded streams. Change-Id: I8bb8c0e1c1fb70f1795cf8b0708a99508d8b564a

Bug: 33661708 Bug: 32707507 Bug: 32095713 Test: run CTS AudioEffectTest#test5_0Command, Custom binder test Change-Id: I03f674f126c191143bd8bdfe236f793e975826a5 (cherry picked from commit 31a4598) (cherry picked from commit 8415635765380be496da9b4578d8f134a527d86b)

Avoid a CHECK and make it the decoder's repsonsibility to handle a malformed bistream gracefully. Bug: 34509901 Bug: 33137046 Test: StagefrightTest#testStagefright_bug_27855419_CVE_2016_2463 Change-Id: I2d94f8da63d65a86a9c711c45546e4c695e0f3b4 (cherry picked from commit 91fe76a) (cherry picked from commit 371d728a9eae3da669567075b7075ba15e611ded)

These functions also exist as arm assembly files, where the overflows just wrap around, and this makes their plain C equivalents behave the same. Bug: 32577290 Bug: 33071964 Test: ran PoC for bug 32577290 using plain C source code Change-Id: I73c2609589e7a89d36f6c44391548312259daf14 (cherry picked from commit cbf5e69)

Change-Id: I135187e2915d3757beaa75b04982c43bcd7424de

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 2012879 I1f5c070233c5917d85da9e930e01a3fc51a0a0ec Don't initialize sync sample parameters until the end 2012475 Ia1098cba2cd32cc2d1c9dfdff4adc2388dfed80e DO NOT MERGE - improve audio effect framwework thread sa 2000357 1082919 If57e12913bebfca6ab9b871a0bdcc24d663e7107 Camera: Add NULL check on mDevice to avoid crash. Change-Id: I2b4a88e0534338c71b89eecf950f63e6bd217bc2 CRs-Fixed: 2012879, 1082919, 2012475, 2000357

- All the required SoundTrigger models of a module client are not getting unloaded and removed as part of detach. - mModels.size() is evaluated for every iteration which gives current size of vector which results in exit of loop without checking for all models in case elements are removed in loop. removeItemsAt would modify the vector indices for each iteration and this would result in skipping access of some models. - Iterate from end of models to check and unload. CRs-Fixed: 2012759 Change-Id: Iead537989c4d0fc3873c08d1f1b51bb22f814da7

HDMI/Display Port does not behave like a fixed volume device during playback of compressed streams. This is because getDeviceForStream(AUDIO_STREAM_MUSIC) returns HDMI + Speaker which is not treated as a fixed volume device. getDeviceForStream returns HDMI + Speaker as getDeviceForStrategy returns Speaker for STRATEGY_ACCESSIBILITY which is appended to the list although the stream is not active. Add only those devices which have outputs with the current stream active to fix the issue. CRs-Fixed: 2014016 Change-Id: Ia7bf565ed51011da040564c6628f67a671355013

Invalidate direct PCM offloaded tracks when mono is enabled. With this change, the AudioTrack client/Media Player must recreate the tracks as non-offloaded PCM. Change-Id: I30c883f88c262eb8a4888014e47558503048c139 CRs-Fixed: 2013919

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 2012759 Iead537989c4d0fc3873c08d1f1b51bb22f814da7 soundtrigger: Modify conditional check in detach 2014016 Ia7bf565ed51011da040564c6628f67a671355013 audiopolicy: return proper devices from getDevicesForStr Change-Id: Ieefac47ecfb0d4647ca54ad354f470397b6686a2 CRs-Fixed: 2012759, 2014016

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 2012900 I73c2609589e7a89d36f6c44391548312259daf14 Turn off overflow protection for various math functions 2012475 I8bb8c0e1c1fb70f1795cf8b0708a99508d8b564a Audio: Make sure effects are enabled for PCM offload 2013919 I30c883f88c262eb8a4888014e47558503048c139 audiopolicy: Invalidate direct PCM tracks when mono is s 2013005 I2d94f8da63d65a86a9c711c45546e4c695e0f3b4 avc_utils: skip empty NALs from malformed bistreams 2012475 I03f674f126c191143bd8bdfe236f793e975826a5 DO NOT MERGE - audioflinger: fix recursive mutex lock in Change-Id: I3def80d222c4e448c67640aae49751fbb133eca9 CRs-Fixed: 2012475, 2013005, 2013919, 2012900

Writer thread in encoder usecase gets stuck in destructor if recorder is stopped right after being paused. This is because encoder source is paused when recoder pauses, then omx encoder data flow stops. This in turn causes MediaCodecSource::read() to be blocked infinitely. The proper sequence to exit the main thread of encoding writer is to stop the source firstly before waiting for thread termination. Change-Id: Ic70bdc2ad06d3e03f6a2966c7c7cc2557a214349 CRs-Fixed: 2012015

Change-Id: I5445604ad073e798667d9bff362e88fe47bec988

This reverts commit c6abb7e1916601e7a9b97a1124acbdcd91ebe22e as there is a CTS regression (AudioManagerTest#testVolume) due to it. getDevicesForStream() can get called for an inactive stream, so the device for that stream/strategy as per PolicyEngine should be returned instead of AUDIO_DEVICE_NONE. CRs-Fixed: 2025755 Change-Id: I1a03dcf74a91ee33a843a923fa77170dfdbd6ff9

…tream""

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 2014325 Iefdd061d745f548a207dd96f15c323fa7900d40d NuPlayer: fix possible deadlock due to recursive mutex a 2025755 I1a03dcf74a91ee33a843a923fa77170dfdbd6ff9 Revert "audiopolicy: return proper devices from getDevic 2012015 Ic70bdc2ad06d3e03f6a2966c7c7cc2557a214349 stagefright: stop audio encoding before releasing writer Change-Id: Ia3df3d354b5aaaa308966b98d3f07e82d295671a CRs-Fixed: 2014325, 2012015, 2025755

…av-aosp.lnx.2.0-dev

…icts Change-Id: I13dadbd25b42ffd0e869c7d62c6d63d5e1165e2f

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- Idae8742fb4b18b2ef4842ec66284268fc650df22 Camera: fix camera facing override logic I69f24eac5866f8d9090fc4c0ebe58c2c297b63df Fix security vulnerability: Effect command might allow n Ia1098cba2cd32cc2d1c9dfdff4adc2388dfed80e DO NOT MERGE ANYWHERE - improve audio effect framwework I9469d6f9929ca67f2f12e658382bc76e7236b256 media: Adjust the first video/audio sample time in slow I33d4bff6b2e19a9fc4284a85a446804878d3a410 EffectBundle: check nb channels to write speaker angles Idbaf221fa2283b33e83f399562a3323dd095cc2c DO NOT MERGE: defensive parsing of mp3 album art informa I479d51e88d3180461cb6ea5540974671cfd84201 DO NOT MERGE: defensive parsing of mp3 album art informa I3356eb615b0e79272d71d72578d363671038c6dd codecs: handle onReset() for a few encoders I03f674f126c191143bd8bdfe236f793e975826a5 DO NOT MERGE - audioflinger: fix recursive mutex lock in Ia1098cba2cd32cc2d1c9dfdff4adc2388dfed80e DO NOT MERGE - improve audio effect framwework thread sa I3356eb615b0e79272d71d72578d363671038c6dd DO NOT MERGE codecs: handle onReset() for a few encoders Ic76eb4241ed74957414aef2990be08cd77a9f5a9 Don't send short URB packet when sending MtpDataPacket. I15d4f138218f6f01a020c4fc568a5b4831cb92a7 AudioRecord: fix AudioRecordThread crash upon exit I479d51e88d3180461cb6ea5540974671cfd84201 DO NOT MERGE: defensive parsing of mp3 album art informa I9b6708bef087e3f4a97868b2d8f7a9b1558233d0 APM::updateCallRouting: ignore for stub output hal I4e3a37aabc9387432671c1c0c469241142612cc4 Don't CHECK when buffer is too large I4c0bc29f061f1b66710c8188a7916bfaf089d23f Don't try to connect camera service if it is disabled. Iea96ba0daf71691ee8954cca4ba1c10fe827626e Effect: Use local cached data for Effect commit Ia1098cba2cd32cc2d1c9dfdff4adc2388dfed80e DO NOT MERGE - improve audio effect framwework thread sa Ia9f526d71b21ef6a61acacf616b573753cd21df6 DO NOT MERGE: defensive parsing of mp3 album art informa I91bd6aadb6c7410163e03101f365db767f4cd2a3 Fix security vulnerability: Equalizer command might allo I2966bed35d0319e3f26e3d4b1b8dc08006a22348 CameraBase: Don't return an sp<> by reference DO NOT MERGE Revert "Log audio information to ensure com I6919e8ac6e35092273e171f49f6711ba577ba2e6 stagefright: remove allottedSize equality check in IOMX: I1ed62b97d3d39e5d7f0e7c407c6cc482b8968373 Camera API1: relax the fps check with some margin I69eb18abe191ae0f1e76f8b62daaf5717cce9669 Camera API1: don't pick wrong preview fps value Ia85519766e19a6e37237166f309750b3e8323c4e Fix overflow check and check read result I9b1315e2c02f11bea395bfdcf5c1ccddccbad8a6 Effects: Check get parameter command size I477820e03a417fe1968450fe783ea7183aeadfd9 NuPlayerDecoder: initialize dropAccessUnit to true I23f5ccba8f211e01d9a3a741c8ea537b55aab4e2 stagefright: parseApp check data boundary conditions Ibad6fc202692cd3480ae726627252afdead083f3 Log audio information to ensure complete delivery 2019252 Ibd61ea227138449ce5e4e0ec548863b02e1f4ac5 stagefright: Remove extra memcpy Id77d023299a851ac6bf8f90c4eea90fe125ea3f5 MediaBufferGroup: release buffers in add_buffer when gro Iccd13967305f1d0a62b89da87070591bfb973bbf Camera API1: filter out the unsupported preview fps rang I9f9c6a372a039226d61f3651be3af207fed63e60 IOMX: convert ANWB to Gralloc meta if using useBuffer in Id6e7cabdd9e2f0f5d44f4ce435121f84d891026a If a device does not have a camera, do not set a camera I2966bed35d0319e3f26e3d4b1b8dc08006a22348 CameraBase: Don't return an sp<> by reference I1c439a0d0f6aa0057b3c651499f28426e1e1f5e4 DO NOT MERGE: Visualizer: Check capture size and latency I436e4a1ee998ce4581eb540bea8b66154960a4e7 DO NOT MERGE SimpleDecodingSource: fix handling of Vorbi Ia1098cba2cd32cc2d1c9dfdff4adc2388dfed80e DO NOT MERGE - improve audio effect framwework thread sa Iec98b856e8d4ce888f8a1cd2568a3dee871effc5 provideKeyResponse to send an event to test AMediaDrm_se I71d17b1586284feb782d14d77d58a8b0614435dd MediaRecorderClient: Use checkService instead of getServ I8900c92fa55b56c4c2c9d721efdbabe6bfc8a4a4 Fix security vulnerability: potential OOB write in audio Idb2a8653744535f711139c87f39b259daddbf337 DO NOT MERGE Move legacy widevine classic global lock fr Ia1098cba2cd32cc2d1c9dfdff4adc2388dfed80e DO NOT MERGE - improve audio effect framwework thread sa I2d94f8da63d65a86a9c711c45546e4c695e0f3b4 avc_utils: skip empty NALs from malformed bistreams I2d94f8da63d65a86a9c711c45546e4c695e0f3b4 avc_utils: skip empty NALs from malformed bistreams 2018475 Ib39fee7bb796c967639aa85be46fd913796041f6 video: Disable sync frame decoder mode for closest seek I1c439a0d0f6aa0057b3c651499f28426e1e1f5e4 Visualizer: Check capture size and latency parameters I73c2609589e7a89d36f6c44391548312259daf14 Turn off overflow protection for various math functions I9bcc9422ae7dd3ae4a38df330c9dcd7ac4941ec8 Make VBRISeeker more robust Ia1098cba2cd32cc2d1c9dfdff4adc2388dfed80e DO NOT MERGE - improve audio effect framwework thread sa I73c2609589e7a89d36f6c44391548312259daf14 Turn off overflow protection for various math functions I2bbbd8a4e92a4c926eff538a67d9719c1b6ec7d0 Fix findId spinning if session id is not found. *** DO N Icd28d465dca3f850938c13058c916efc477b0d2b Camera: Clear not-yet-submitted request in cleanupFailed I0a80bdf67bf2913522821ac705f3dc548d3edead Stop sending object handle with sendObject request. I35bee7c718e01b086d7e05deda13b38083f509f5 Avoid crash for stss sync sample number 0 Ib46cd1b5a3e1707c343d395f235bedb82764e184 DO NOT MERGE: defensive parsing of mp3 album art informa I1f5c070233c5917d85da9e930e01a3fc51a0a0ec Don't initialize sync sample parameters until the end I0a81c26d22fee36485b21c285dcc91fbd518e1dd Revert "Turn off overflow protection for various math fu Change-Id: I954b9cf5c6593200970d92f4f7b4c9cd57300dbb CRs-Fixed: 2018475, 2019252

This reverts commit c6abb7e1916601e7a9b97a1124acbdcd91ebe22e as there is a CTS regression (AudioManagerTest#testVolume) due to it. getDevicesForStream() can get called for an inactive stream, so the device for that stream/strategy as per PolicyEngine should be returned instead of AUDIO_DEVICE_NONE. CRs-Fixed: 2025755 Change-Id: I1a03dcf74a91ee33a843a923fa77170dfdbd6ff9

Change-Id: I1d8f1d9b8991dabb6635951076311f78d2fcf603

A2DP output device is selected when SCO is active and A2DP suspended. Add support to avoid selection of A2DP device in such scenario. Modify conditional checks to ensure mA2dpSuspended is set to true or false when A2DP playback is on primary output as well. Retain original code to avoid suspend/restore output in case A2DP is on primary output. Change-Id: I55da447636f8901979fc0fce6858a31cdf525b7b

Earlier during pause resume operations the pause duration was multiplied with the ratio of capture FPS and the desired frame rate of the recorded video. This logic led to situations where if the pause duration was less than the above quantity, on subtracting, the value would come negative. This led to elongation of the length of the video beyond the desired length. Also the earlier logic led to miscalculation of the timeoffset that needs to be subtracted from the timestamp from camera. This led to elongation of time length as the value subtracted was much lesser than what was required. Change-Id: I98cdb14bb2b9c86013df9b2c8f2e558f184b633e

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 2032606 I55da447636f8901979fc0fce6858a31cdf525b7b audiopolicy: Do not select A2DP device if suspended Change-Id: I0b848ead74c2ccc0eade68c609b26e9f7c604546 CRs-Fixed: 2032606

release_l() is invoked from effect destructor which doesn't hold effect chain mutex, which results into multi-thread contention. Holding effect mutex when releasing effect engine to improve thread safety. Change-Id: I2056693f79bcb3df9a5a9dac1b700d50ed535372 CRs-Fixed: 2026604

Assignment of IMediaSource::mMetaData of type sp<T> in IMediaSource::getFormat() is not thread-safe and is found to be causing stability issues. Synchronize access to meta smart pointer to prevent potential race condition. CRs-Fixed: 2038194 Change-Id: I8a94d82d69f35307e5fab8174d752a847f47a2f2

Avoid waiting for video track to catch up to last audio timestamp in conditions when maxfileduration or maxfilesize limit is reached or if writer encounters any malformed error. Also, update the wait timeout interval to 500msec instead of 1sec as camera can hold maximum 15 frames when EIS is enabled. Change-Id: Idb3330137009be05a2bb4131a192c8f483546b6f

The MediaCodec member associated with an NuPlayerDecoder is cleared if configure fails. A pending seek operation in NuPlayer can issue a ResumeDecoderAction. Add a NULL check before accessing the MediaCodec instance to handle this resume. Change-Id: Idc6c15168ec922ca1bd6d867cbe59847647c6f26 CRs-Fixed: 913215

Add code to validate the size parameter passed to HecvParameterSets::addNalUnit(). Previously vulnerable to decrementing an unsigned past 0, yielding a huge result value. Bug: 35467107 Test: ran POC, no crash, emitted new "bad length" log entry Change-Id: Ia169b9edc1e0f7c5302e3c68aa90a54e8863d79e (cherry picked from commit e0dcf097cc029d056926029a29419e1650cbdf1b) (cherry picked from commit 1552726c8a9a7579fcb00f7849d8a1d59e4f571c)

Bug: 30895578 Bug: 34970788 Change-Id: I4cba36bbe3502678210e5925181683df9726b431 (cherry picked from commit 328cd66cc72ba7bc5452ed5a93f29ddcd73aa9f9) (cherry picked from commit 8ee699977c155c6d67e4abd8b4259b5447148a80) (cherry picked from commit 0be0ed5a7c9d288f324675e80b0fca657570a8bf)

Bug: 35763994 Test: ran CTS with and without fix Change-Id: If835e97ce578d4fa567e33e349e48fb7b2559e0e (cherry picked from commit 8538a603ef992e75f29336499cb783f3ec19f18c) (cherry picked from commit 76a66d3e9cea6ce8f4d5bf71ed8a0ea6f113a355)

Bug: 34618607 Change-Id: I84f0ef948414d0b2d54e8948b6c30b8ae4da2b36 (cherry picked from commit d1c19c57f66d91ea8033c8fa6510a8760a6e663b) (cherry picked from commit 94a2371d16769e8136a83215a1af34034ad7752f)

Bug: 35269635 Test: decoded PoC with and without patch Change-Id: I636a14360c7801cc5bca63c9cb44d1d235df8fd8 (cherry picked from commit 2ad2a92318a3b9daf78ebcdc597085adbf32600d) (cherry picked from commit ddc86fc36a42a441f592cb848a06331b0861c3b1)

Original code blindly copied some header information into the user-supplied buffer without checking for sufficient space. The code does check when it gets to filling the data -- it's just the header copies that weren't checked. Bug: 34617444 Test: ran POC before/after Change-Id: I6e80ec90616f6cd02bb8316cd2d6e309b7e4729d (cherry picked from commit 6231243626b8b9c57593b1f0ee417f2c4af4c0aa) (cherry picked from commit f8fc1e85e2007be748b4d3342a8de36602208e2a)

Test: Native POC Bug: 34749571 Change-Id: I7529658e52ac7e64d162eb5338f10fb25eaa8fe7 (cherry picked from commit 1883f69de5f2c4e71df58d5b71d7c39f9779b50c) (cherry picked from commit eaa3969f757291f151efedde17ec529b8659024d) (cherry picked from commit 51e72602eb3b7ee9aedc63e6cf7ca58958775481)

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 2016461 I98cdb14bb2b9c86013df9b2c8f2e558f184b633e media : Changing time stamp manipulation in HFR recordin 2026604 I2056693f79bcb3df9a5a9dac1b700d50ed535372 audio: hold effect mutex when releasing effect engine 913215 Idc6c15168ec922ca1bd6d867cbe59847647c6f26 NuPlayer: Decoder: Check for valid codec on resume Change-Id: Ib650ae581f8cee81b0007e9dcea2d08bf8eeb818 CRs-Fixed: 2016461, 913215, 2026604

Don't assume 16 bit format for audio clip when calculating frame size. Otherwise data that is not frame aligned will be sent to NuPlayer, resulting in dropped data. Instead pass frame size based on audio format. CRs-Fixed: 2020023 Change-Id: I33097c6928ba355c6a3a6f4428c3957067f2397e

Change-Id: Ib379672ea4b31f21b4038770702b7568d5afb366

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 2037739 I4cba36bbe3502678210e5925181683df9726b431 FLACExtractor: copy protect mWriteBuffer Change-Id: Ic763b7fd32bd3b986744b8114057232dc6d93b9d CRs-Fixed: 2037739

As sample rate of input frame may change dynamically, it is possible that the input frame size requested may grow given the output frame size is fixed. If we don't enlarge the input buffer in this case, array out of boundary occurs and corrupts the heap. Fix is to resize resampler's inframe buffer independent of out buffer. CRs-Fixed: 2044381 Change-Id: I43f70dbb3eaadabc8ed06817a04165feab320b3c

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 2037750 I7529658e52ac7e64d162eb5338f10fb25eaa8fe7 AudioFlinger: Check framecount overflow when creating tr 2037747 I6e80ec90616f6cd02bb8316cd2d6e309b7e4729d Add bounds check in SoftAACEncoder2::onQueueFilled() 2037764 Ia169b9edc1e0f7c5302e3c68aa90a54e8863d79e Validate lengths in HEVC metadata parsing 2037758 I636a14360c7801cc5bca63c9cb44d1d235df8fd8 Fix NPDs in h263 decoder 2037763 If835e97ce578d4fa567e33e349e48fb7b2559e0e Fix integer overflow and divide-by-zero 2037735 I84f0ef948414d0b2d54e8948b6c30b8ae4da2b36 Fix out of bounds access Change-Id: I9aef42d7b9d5117a77cc605fa063c75917906d36 CRs-Fixed: 2037758, 2037735, 2037750, 2037747, 2037763, 2037764

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 2020023 I33097c6928ba355c6a3a6f4428c3957067f2397e stagefright: don't assume 16 bit format 2038194 I8a94d82d69f35307e5fab8174d752a847f47a2f2 libmedia: synchronize access to meta Change-Id: Iaf100b4a5f98d2f5b2621bfe946a8549f588e897 CRs-Fixed: 2038194, 2020023

Timeoffset miscalculation may lead to retrograde timestamps which will raise a error in MPEG4Writer. This change prevents such occurences in HFR by making sure that the successive buffers have a timestamp difference of 33 milliseconds. The previous changes for timestamp manipulation have been reverted. Change-Id: I3faf7294f743b1031ccc6624c3348f7e12b339b8

…ffer"

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 2044381 I43f70dbb3eaadabc8ed06817a04165feab320b3c audio: resize resampler's inframe buffer independent of 2039323 Idb3330137009be05a2bb4131a192c8f483546b6f stagefright: MPEG4Writer: Skip wait for audio in few con Change-Id: Iff276bd4be03c76dc47eaab95b164d92b7a1e4ff CRs-Fixed: 2039323, 2044381

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 2040285 I3faf7294f743b1031ccc6624c3348f7e12b339b8 media: Modify timestamps for HFR use case Change-Id: I70a1c385930425a4f5163fc2ca0038af80f2001b CRs-Fixed: 2040285

Change-Id: I9022a217501020e5b89986ede074b97a355cbeb7

Change-Id: I4c35df0b0a8581124243fd2d8a5cb4adbc2d9829

Issue: For quadracfa capture, Blob/YUV output streams need to be configured with custom dimensions which will not be available in advertised stream configurations map. Fix: Skip the stream size check for whitelisted apps to allow configuration of streams with custom dimensions. Setprop to be used: adb shell setprop persist.camera.cfa.packagelist <pack1,pack2> CRs-Fixed: 2075934 Change-Id: Id94b40c94f42bf4579dc6d8bb6273003312ea669

…tion Bug: 35467458 Change-Id: Ia76c8cec8ad2abb95ca29b2a89075f7acab4b174 (cherry picked from commit e25320450ca414dd859fd5d80bcd18ecc91630de) (cherry picked from commit 07954224344fc938f157b3a2af9682690831e234)

Soft avc encoder cached buffer sizes across a point where the sizes could be reset by an app, allowing crafted requests to hit outside of the current buffer. This remedies that by forcing buffer reallocation with new sizes whenever the encoder state at such 'reset' points. Bug: 35421151 Test: run POC with no crash Change-Id: I8c689846142264f7b6a277332260a6bd8a2bd92d (cherry picked from commit 463452a94907278ca5e4c5632e50bbe59e861544) (cherry picked from commit 5eb118f8299129e323fa402e0605abe696bdc678)

A malformed ogg file might lace together a very large packet, which could lead to out of memory conditions. Limit the packet size to avoid this. Bug: 36592202 Change-Id: I8650b3ec54a0de9ec302a7cbac296bb85efcfb3d (cherry picked from commit bf928560aca13c5a615cb3ffc3b6aad16cdf3824) (cherry picked from commit 4efebc8b90d2c6c98d67c9dc0348a713c2a0700f)

Bug: 37239013 Change-Id: Ic33e0f7ed946d0729efa46f69aff1a5d35e81b1e (cherry picked from commit ab34612a31e82b713ca0ac043e14f68f3788fbda) (cherry picked from commit 4cd260f382f51b309be9aa48725cc709f5b034f2)

so don't use those char* after having updated the MetaData Bug: 36576151 Change-Id: I1430f3151cb929e436fbdac566cf91fc4164b934 (cherry picked from commit 423e74647d1b8fd8b2905356af1fc2b1c67a0f3f) (cherry picked from commit a520c5d445b7cb6ac4c2a09b86d2e860710a32ae)

CRs Change ID Subject -------------------------------------------------------------------------------------------------------------- 2066798 Ia76c8cec8ad2abb95ca29b2a89075f7acab4b174 Don't allow using or allocating a buffer after the first 2066777 I1430f3151cb929e436fbdac566cf91fc4164b934 Modifying MetaData invalidates previous char* 2066784 I8c689846142264f7b6a277332260a6bd8a2bd92d Prevent OOB write in soft_avc encoder 2082635 Id94b40c94f42bf4579dc6d8bb6273003312ea669 Camera: Skip stream size check for whitelisted apps. 2066789 I8650b3ec54a0de9ec302a7cbac296bb85efcfb3d Limit ogg packet size 2066797 Ic33e0f7ed946d0729efa46f69aff1a5d35e81b1e Fix memory leak in error case Change-Id: I12d57473c7d6dfe6ef4d69b44afca0cd35baeedf CRs-Fixed: 2066777, 2066797, 2066789, 2066798, 2066784, 2082635

Change-Id: Ib563f1b5b3c36fb4e1bb144ae0fc714d0a0d28f0

…g/external/gigabyte/platform/frameworks/av into n "LA.UM.5.8.r1-02700-8x98.0" Conflicts: media/libstagefright/MediaCodecSource.cpp services/audiopolicy/managerdefault/AudioPolicyManager.cpp

Camera: Minor Improvements and Fixes #2

Are you sure you want to change the base?

Camera: Minor Improvements and Fixes #2

Commits on Nov 15, 2016

Commits on Nov 28, 2016

Commits on Nov 30, 2016

Commits on Dec 1, 2016

Commits on Dec 3, 2016

Commits on Dec 6, 2016

Commits on Dec 7, 2016

Commits on Dec 8, 2016

Commits on Dec 9, 2016

Commits on Dec 19, 2016

Commits on Dec 20, 2016

Commits on Dec 21, 2016

Commits on Dec 22, 2016

Commits on Dec 26, 2016

Commits on Dec 29, 2016

Commits on Jan 2, 2017

Commits on Jan 3, 2017

Commits on Jan 4, 2017

Commits on Jan 11, 2017

Commits on Jan 12, 2017

Commits on Jan 17, 2017

Commits on Jan 18, 2017

Commits on Jan 19, 2017

Commits on Jan 20, 2017

Commits on Jan 24, 2017

Commits on Jan 25, 2017

Commits on Jan 26, 2017

Commits on Jan 27, 2017

Commits on Jan 28, 2017

Commits on Jan 30, 2017

Commits on Jan 31, 2017

Commits on Feb 1, 2017

Commits on Feb 2, 2017

Commits on Feb 3, 2017

Commits on Feb 6, 2017

Commits on Feb 7, 2017

Commits on Feb 8, 2017

Commits on Feb 9, 2017

Commits on Feb 13, 2017

Commits on Feb 15, 2017

Commits on Feb 16, 2017

Commits on Feb 17, 2017

Commits on Feb 22, 2017

Commits on Feb 23, 2017

Commits on Feb 25, 2017

Commits on Mar 1, 2017

Commits on Mar 6, 2017

Commits on Mar 7, 2017

Commits on Mar 8, 2017

Commits on Mar 12, 2017

Commits on Mar 13, 2017

Commits on Mar 14, 2017

Commits on Mar 16, 2017

Commits on Mar 17, 2017

Commits on Mar 18, 2017

Commits on Mar 20, 2017

Commits on Mar 22, 2017

Commits on Mar 24, 2017

Commits on Mar 26, 2017

Commits on Mar 27, 2017

Commits on Mar 30, 2017

Commits on Apr 2, 2017

Commits on Apr 4, 2017

Commits on Apr 5, 2017

Commits on Apr 6, 2017

Commits on Apr 7, 2017

Commits on Apr 11, 2017

Commits on Apr 19, 2017

Commits on Apr 24, 2017

Commits on Apr 28, 2017

Commits on May 3, 2017

Commits on May 4, 2017

Commits on May 5, 2017

Commits on May 8, 2017

Commits on May 9, 2017

Commits on May 10, 2017

Commits on May 11, 2017