Revert "Revert "Revert "Make libbinder not available to media.sw..."
Revert submission 11485868-reland-libbinder
Testing potential culprit for b/156778943 ("pthread_mutex_destroy called on a destroyed mutex" inside libstagefright).
Reason for revert: broken above test
Exempt-From-Owner-Approval: revert due to test breakage.
Reverted Changes:
I045836e10:Use NO_IMEMORY version of libstagefright_foundatio...
I23f41fece:Revert "Revert "Make libbinder not available to me...
bug: 156778943
Change-Id: I2834070f492db0318e215bf16058e0c43e00b03d
Currently swcodec apex build uses vendor variant which no longer
depends on libbinder. However, in some builds VNDK is not used
and libbinder dependency still exists. Although the prebuilt
apex no longer contains libbinder, some build still fails when
libbinder is disallowed in apex.
Create a separate version of libstagefright_foundataion_no_imemory,
and use that only if there is no vendor build.
Also do not use __ANDROID_VNDK__ to disable the binder related
code in libstagefright_foundation, it's never defined if there is
no VNDK. Use NO_IMEMORY instead.
bug: 139201422
bug: 139016109
test: various builds in presubmit
Change-Id: I045836e10b93f71c7c31b30bb555eea8d6a2ef3a
Do not set SMPTE2086 if max/min luminance fields are 0,
these may indicate that max/min luminance is not present
in the bitstream. Let GPU composition use default if needed.
Bug: 153847809
Test: Play Netflix HDR content in GPU composition.
Change-Id: Ibce5319f7832095314fe07b0ce13600658ce4627
- Handle various corner cases with regard to REQUEST_ERROR, RESULT_ERROR, and BUFFER_ERROR.
- Drain the codec outputs in case the input buffer isn't dropped.
- Allow APP_SEGMENT to drop while still producing valid output image.
- Add a status tracker to manage active/idle state.
- Use frame number as key for pending input frames since with ZSL, 2
capture result could have the same timestamp.
- Also removed some deprecated variable/methods.
Test: CTS, vendor testing
Bug: 145579077
Change-Id: I9c3e929469b8fb75b32b016f9006036c954f663f
This reverts commit 47cdcf399f.
Reason for revert: Each ATRACE_CALL operation has around 10us of perf cost. For cases
like 240fps recording, each removeInFlightMapEntry operation, on an
average has 20-30 CameraMetadata constructions, costing us .2 - .3ms in
latency. This is around 6% of the ~4ms budget that we have.
According to
https://source.android.com/setup/develop/new-device#build-variants
'Any debugging features that are enabled by default in userdebug should
be clearly defined and shared with all developers working on the
project. You should enable debugging features only on a limited-time
basis until the issue you're trying to debug is resolved'.
Reverting this change now that http://b/144028609 has been resolved.
Bug: 155910869
Test: builds
Signed-off-by: Jayant Chowdhary <jchowdhary@google.com>
Change-Id: I56e471a83097527a1a211c61c3bd73136a82b6b6
Existing definition was inconsistent, so update it to be consistent
and match what implementations have actually done.
Test: Builds
Bug: 150331548
Change-Id: I037edfbf462ae9822194761d68bcfa34d27f79c6
refactor some mutex for how nuplayer sets up mediametrics data.
expanded locking to eliminate a couple race conditions.
Bug: 151644303
Bug: 151643722
Test: poc attached to bugs
Change-Id: I75f29a6254c5eab5d4f524ee7a7ef59f93a0b405
The unparceling code for face metadata in API1 dynamically allocates
a camera_metadata structure, and a recent fix forgot to delete that
structure in an error path.
Fix this by making the metadata structure statically allocated; it's
tiny anyway.
Test: atest CtsCameraTestCases
Bug: 155512103
Change-Id: I9b81cdb3ab47c7b6c94c6ed65213ddafef529dc9
Bug: 155096545
Test: adb shell dumpsys media.audio_flinger | more
Test: then search for "Device"
Test: There should be no garbage characters on the next line.
Change-Id: I798d914cede3f45151f97178de3ceee09b4e24c0
When a new device is attached to hardware module, query the HAL to get
the supported encapsulation modes and metadata types.
Test: atest EnumDevicesTest, manually
Test: atest audiofoundation_parcelable_test
Bug: 155799506
Change-Id: I903e46e35a44416c9f2d206c128face8a6c61dcb
Spawn a thread to handle the disconnects for onRoutingChanged()
and onTearDown(). This is to prevent deadlocks when calling back into
AudioFlinger.
Bug: 155819038
Test: See bug for repro steps with OboeTester
Test: Also run TEST DISCONNECT in OboeTester. It was failing.
Test: atest CtsNativeMediaAAudioTestCases
Change-Id: I83212fea140d3b6beca9aa737606a77a7c97b007
Chromebook sometimes runs 32 and 64-bit code together.
That caused messages between the code to be rejected because the
struct sizes did not match.
So we used alignas(8) to force members to 8 byte boundaries.
Bug: 155901581
Test: Launch OboeTester
Test: Tap TEST OUTPUT
Test: Tap OPEN and START
Test: You should get "MMAP YES" if MMAP is supported by the device.
Change-Id: I0cd6a59a974888ef17b06b38fa78f032d602cd57
Do not set CTA861_3 bit if content light level
fields are 0, these may indicate that CLL is
not present in the bitstream. Let display or
compositor use default if needed.
bug: 152581483
Change-Id: I6764623130f839df83fd6f5db1b97ab768a8642d
Merged-In: I6764623130f839df83fd6f5db1b97ab768a8642d
(cherry picked from commit cbbbfab2126050a84ab9067824c3cdb62d581e83)
- Trigger audio power usage logging when
1. endAudioIntervalGroup event for track/record
2. setMode to NORMAL for telephony
- Aggregate audio usage by type and device
- Send audiousage data every 24hrs by default, using property
persist.media.audio_power_metrics.interval_hr to config the interval
- Add property persist.media.audio_power_metrics.disable to disable
audio power metrics
Bug: 121318157
Bug: 140909217
Test: atest mediametrics_tests
Test: adb shell dumpsys media.metrics --all
Change-Id: I94ff00a5609a49c96a14827d3260d481c1310abe
Signed-off-by: Robert Lee <lerobert@google.com>