| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
| |
* Annotate source buffers with the audio format
* Add support for 32-bit signed PCM offload (zero copy)
Change-Id: Id758830784740c0a038452d383c8ec8e3e4593bb
|
|
|
|
|
|
| |
* Remove unnecessary check condition.
Change-Id: I3267a0c13165d74a2ea90333b42f000b51ace98b
|
|
|
|
|
|
|
|
| |
* This code should have been left inside the conditional
* Also fix a misplaced flag which remained set even if opening
in offload mode failed
Change-Id: Id72c17051db601e37b2289e6d904ce8f75ba6878
|
|
|
|
|
|
|
| |
* Fix PCM offload when passthrough isn't used (OMX decoder)
* Fix resume of PCM offload after pause timeout
Change-Id: I742eafd6ae8656fb214ba6b81cc63af57590c28c
|
|
|
|
|
|
|
|
|
|
| |
Audio EOS won't be notified by renderer if AudioSink is not ready
and first buffer itself has EOS. Playback complete won't happen
due to missing audio EOS. Hence, audio EOS needs to be handled
and notified.
Change-Id: I779c7034d1964485c2b064c0179d3cd341af5a5f
CRs-Fixed: 801121
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
While offload playback is going on, if user seeks
to new position and after playback for some time,
pauses till offload tear-down happens then playback
resumes with old seek position
The book-keeping of seektime for start of playback
mStartupSeekTimeUs is also done in running state,
it should be done only if current state is paused.
Change-Id: I4173a2ee4244bcc6794822dde24c467f0189b84f
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
While offload playback is going on, if user pause
and seeks to new position before offload tear-down,
on resume playback starts from old time-stamp
Once tear-down happens NuPlayer loses updated time-stamp
so while resuming it resumes playback with stale time-stamp.
Make sure the updated time-stamp is maintained in
NuPlayerDriver wich is used while starting the playback
after resume.
Change-Id: I3451051f569264b21a43be81b01798fabed0182e
|
|
|
|
|
|
|
|
|
|
|
| |
* Fix compilation with debug enabled
* Allow FLAC offload to be enabled independent of QCOM_HARDWARE,
tested with the Nexus 5X and is working!
* Disable the FLACExtractor if offload is enabled since it can
only output decoded PCM. This will force FFMPEG to be used,
which does everything we need for FLAC offload
Change-Id: I7d71c153a6a6ea7df8e32bc73f5cbe9f51cdcf64
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Forward-port the open-source code from L which was moved to closed-source
in M. This is being done out of necessity- the architecture chosen by
Qualcomm is not optimal and doesn't work well with a singular
codebase which attempts to service a large number of devices.
* This patch brings in the code to support PCM offload (AudioFlinger
bypass). This allows for playback of high resolution clips without
decimation stages, and enables reduced power consumption for audio
pipelines which take advantage of the Hexagon DSP (effects).
Change-Id: I0ef15fc3df538ab723f3c12ce0ed71d0e607c99e
|
|
|
|
| |
Change-Id: I25708df616fe53d709b80c65ddecbdddef303124
|
|\
| |
| |
| |
| |
| |
| |
| | |
https://android.googlesource.com/platform/frameworks/av into cm-13.0
Android 6.0.1 release 3
Change-Id: I2f2a1fe1b58c828e8341556996211562d6e195ab
|
| |
| |
| |
| |
| | |
Bug: 25070434
Change-Id: Ib0eb6a419683a0e686a4f63b82e9300cb1f69484
|
| |
| |
| |
| |
| | |
Bug: 25070434
Change-Id: Iade3472c496ac42456e42db35e402f7b66416f5b
|
| |
| |
| |
| |
| |
| | |
Bug: 21093153.
Change-Id: I77e03a620b43967991dcee9d6ca01f67c8b7a02e
(cherry picked from commit 15ce09dddf830d6a6ca097c5ec0a7ec7c35b6b3a)
|
| |
| |
| |
| |
| |
| |
| | |
Delay until resume.
Bug: 23669269
Change-Id: I0a805812d80b1aad561425442bca0fc05a05752f
|
| |\ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
AwesomePlayer never did, and neither should NuPlayer.
Bug: 18656047
Change-Id: Ia9abe42cfe6ea643633cec859e9da770092569bb
|
| | |
| | |
| | |
| | |
| | |
| | | |
Bug: 24749881
Change-Id: I0d3a757262446ac702dfebd464d5e59f1a98eba8
(cherry picked from commit 3261f0db54e31e652af9bd3b76a8e85174f5ee87)
|
| |\ \ |
|
| | |/
| | |
| | |
| | |
| | |
| | | |
Bug: 24338126
Change-Id: Ic55d1a551c18255423423079b6c2bf045dea6e44
(cherry picked from commit f3631d2921ee0738f87427848e38873370ef9ec0)
|
| |/
| |
| |
| |
| |
| | |
Bug: 24345295
Change-Id: I868c9c44ea22de98a083432262e485d0f134203f
(cherry picked from commit d005c5ddb4842369979df7b76f1d0f5f1380fcd9)
|
| |
| |
| |
| |
| | |
Bug: 23658148
Change-Id: Ic37cac7b5d166143e0b77e9919b0aaef486e4fdd
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Read performance in StreamingSource became worse from LMR1 because
the design was changed to asynchronous from synchronous.
As a result, the generation of access unit does not make it
in time for some multiplexed contents, and this causes frame drop.
This patch improves read performance by increasing the number of
packets for one read.
Change-Id: I14955da5a9c5415e35905cc95f1471e0811a4991
|
| |
| |
| |
| |
| |
| |
| | |
Acquire lock after DataSource::CreateFromURI() to avoid a deadlock/ANR
as GenericSource::disconnect() also tries to acquire same lock.
Change-Id: I28430800853ed11e0a37a8a38defab72470c3f6d
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
mDataSource can be deleted by client thread while GenericSource
looper is accessing dataSource in sniff functions, which can
lead to race condition. Add local sp<> to hold an additional
reference to dataSource.
Change-Id: I9e51f75f3c2bd703e381230dd93cb4d3f6621dab
CRs-Fixed: 923729
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
DataSource and HTTPSource can be accessed/modified by
GenericSource's looper and Client's thread which can
lead to race conditions when copying the sp<>.
Add a Mutex lock to synchronize such accesses and avoid
race conditions.
CRs-Fixed: 906899
Change-Id: I2fb4b4a7079e638e151f4fe67a780007a4011652
|
| |
| |
| |
| |
| | |
Bug: 25357751
Change-Id: Iffd9af2ddf7626155d98caf086fe209600141585
|
|\ \
| | |
| | |
| | | |
git://codeaurora.org/platform/frameworks/av into cm-13.0
|
| |\ \ |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Extend shutdown sequence used for pcm offload to compress
offload as well. The current tear down sequence for compress
offload calls clear on the NuPlayerDecoderPassthrough object directly.
This is not guaranteed to prevent an encoded buffer from getting
queued into the Renderer. Going through a proper flush+shutdown
sequence helps prevent this (rare) scenario.
CRs-Fixed: 914308
Change-Id: Ie2040cbbd42b8c67bb53a7ef8a628faf1ab7acda
|
| |\ \ \
| | |/ /
| |/| | |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
cherry-pick of AOSP commit b75de9d4ba6725104342b5376a1047ff1c9b2765
Bug: 24712426
Change-Id: I58cca0bde708a5c3b073ed77e18b62fe81d6c017
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This reverts commit a5c3b990807732d1c4a7c1290cba06a605675973.
Change-Id: I1f267cf7ab90b1f5c38e0659df02c8ef518382ed
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
* In which our hero unites the legions of distant worlds together
as one at last to fight anew until the segfaults boil the lands
and the New Version cometh.
Change-Id: I869b2ab254968ed7a133827ea9e5ba6056f68a5c
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
* Standardize parameter names, clean up stub calls
* Attempting to make QC's blob work. We've entered
our last days on earth. All is lost, all is lost.
Change-Id: Idef9aad100f414e64a737e0f0a8c9c112ad4cb46
|
|\ \ \ \
| |/ / /
| | | |
| | | |
| | | |
| | | | |
git://codeaurora.org/platform/frameworks/av into cm-13.0
Change-Id: Ia41df61c3ccfeb809572b63a4b1a8ca6bd85dfb2
|
| |\ \ \
| | |/ /
| |/| | |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Drop frames in NuplayerDecoder if the corresponding
OMX_BUFFERFLAG_DATACORRUPT is set from OMX.
Add a util function to easily switch off this feature.
Do not declare MediaSource explicitly.
Change-Id: I97bd238d52d660e1ab61eac222cdff28334cc5cc
|
| |\ \ \
| | |/ / |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Use a generic function to get flags while creating
extractor, so that there is no restriction to a
particular usecase.
CRs-Fixed: 911338
Change-Id: I8520d2ea9b25ab564b7eee6610744fcaa364fe87
|
| |\ \ \ |
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
mDataSource can be deleted by client thread while GenericSource
looper is accessing dataSource in sniff functions, which can
lead to race condition. Add local sp<> to hold an additional
reference to dataSource.
Change-Id: I9e51f75f3c2bd703e381230dd93cb4d3f6621dab
CRs-Fixed: 923729
|
| |\ \ \ \
| | |/ / / |
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
DataSource and HTTPSource can be accessed/modified by
GenericSource's looper and Client's thread which can
lead to race conditions when copying the sp<>.
Add a Mutex lock to synchronize such accesses and avoid
race conditions.
CRs-Fixed: 906899
Change-Id: I2fb4b4a7079e638e151f4fe67a780007a4011652
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
int overflow when converting current position in ms to Us
Change-Id: I12882e52ed1ac92644d601efae5c10ec25ae9e63
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
AwesomePlayer never did, and neither should NuPlayer.
Bug: 18656047
Change-Id: Ia9abe42cfe6ea643633cec859e9da770092569bb
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Bug: 24712426
Change-Id: I58cca0bde708a5c3b073ed77e18b62fe81d6c017
|
|\ \ \ \ \
| | |_|_|/
| |/| | |
| | | | |
| | | | |
| | | | | |
Android 6.0.0 release 26
Change-Id: I8a57007bf6efcd8b95c3cebf5e0444345bdd4cda
|
| |\ \ \ \ |
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Bug: 23748678
Change-Id: I8b65786f25f4524e5e2e9476ecc2a4f4ab3aea9e
|