In practice, the AoC can crash while the stream is in
any state, thus this should be allowed for correct
handling of error conditions in the framework.
The list of states when `IStreamCallback.onError` can
be received is not enforced in VTS tests, thus no
test regressions are expected.
Bug: 339400414
Test: repro steps for simulating AoC crash in b/339400414
Change-Id: I9a52779f773307882109e159d715072de5af2ece
In aosp/2908743, the intended behavior of the "halReservedExit"
command (stream exit) was inadventedly changed. Instead of
exiting from the thread's loop only when the cookie has
the expected value, it was actually exiting when the cookie
value is any but zero. The intended behavior is as follows:
- the cookie has expected value: do not send reply, exit loop;
- the cookie has unexpected but non-zero value: ignore,
do not send reply (that's the point of using the cookie);
- the cookie has a zero value: ignore, send a reply (this is
for VTS testing).
Bug: 300181540
Test: atest VtsHalAudioCoreTargetTest
Change-Id: I9945eb0ba9042993adac8599b18f241c4f69ca7a
also update VTS data path test skipping for offload effects
Bug: 335547630
Test: atest --test-mapping hardware/interfaces/audio/aidl/vts:presubmit
Change-Id: Ie92f04091658e210dccb7a2d60ebbab14c49fb58
Merged-In: Ie92f04091658e210dccb7a2d60ebbab14c49fb58
Issue a fatal error when the data FMQ detects internal pointers
corruption. This condition indicates some low level problems
and should be diagnosed as fast as possible.
Bug: 338974476
Test: atest VtsHalAudioCoreTargetTest
Change-Id: I962eba620149c7169f0f0a2dad4c4064232966d7
relaxing several parameter checking to align with HIDL
- no need to have stage in use to set bands/channels
- band settings no need to be sorted by frequency
Bug: 328012516
Test: atest VtsHalDynamicsProcessingTargetTest
Change-Id: If12d978ee69ee7f087a7e8758513a9c6bacf817f
Merged-In: If12d978ee69ee7f087a7e8758513a9c6bacf817f
Some audio outputs use A/V sync and requre mode bursts
in order to start reporting the presentation position.
Bug: 300735639
Bug: 328010709
Test: atest VtsHalAudioCoreTargetTest
Change-Id: Icad0942f2ba1dcd6f030a7dc4f37e22fdbd6dd71
According to the definition of the 'PAUSED' state
in StreamDescriptor.aidl, s/w (the client) stops
writing once the buffer is filled up. That means, it is
allowed for an output stream not to consume data
from the MQ while in the paused state, so allow that
in the test.
Also, update the state transition sequence in the test
to flush any data after making a burst while in the 'PAUSED'
state.
Bug: 328010709
Test: atest VtsHalAudioCoreTargetTest
Change-Id: Icb5fd02ca4ede63d7ae33613ab66cb96f3e6df29
The test was using '0' as an "invalid" patch ID value, however
this value is valid in the context of 'IModule.setAudioPatch'
method and means "create a new patch and allocate and ID for it".
Bug: 328010709
Test: atest VtsHalAudioCoreTargetTest
Change-Id: Icd33f3cbd1602ec5aa162fa72fc3ddd59ccffbef
This is aligned with the HIDL implementation VTS. The echo reference
device can't provide any input until certain preconditions are met,
and modeling these preconditions in the test is not trivial.
Also, add the information into the mix port into the trace scope
for easier identification on test failure.
Bug: 328010709
Test: atest VtsHalAudioCoreTargetTest
Change-Id: I737479d8ef1961791ac3bd82aeb779453d2e49f4
Since "active" means "used by the stream for acquiring data,"
it was unreasonable to expect the list of active microphones
to be non-empty prior to actually starting data acquisition.
This change adds running of 'burst' commands before calling
'getActiveMicrophones'.
To reuse existing code some refactorings have been made.
Added 'AudioInputFlags::HOTWORD_TAP' to the list of port
config flags for which I/O testing is not performed.
Bug: 328010709
Bug: 328362233
Test: atest VtsHalAudioCoreTargetTest
Change-Id: I876c0b6d7365e104ec9ed8cf5033a83f822006b6
When staring data transfer, if resuming of BT session fails,
the legacy implementation does not indicate an error, thus
causing an automatic retry on next transfer. The AIDL
implementation was switching the stream into an error state
instead. This could cause audio mute when the BT stack is
slow on resuming. This CL aligns AIDL with legacy behavior.
Bug: 333307789
Test: repro steps in the bug
(cherry picked from https://googleplex-android-review.googlesource.com/q/commit:3f4c7fca29db4b82d597bb56478f40a75cef618c)
Merged-In: I5080a7275671287f188930f041c5ff2215ed2e1e
Change-Id: I5080a7275671287f188930f041c5ff2215ed2e1e
Skip the data path tests for offloading Virtualizer, as offloading
effects don't publish fmqs and don't actually process the data.
Bug: 334851716
Test: run vts -m VtsHalVirtualizerTargetTest
Change-Id: I0d96fe4b036bdbd6c7b65af906edaaaac39f7a10
Added tests to verify increase in energy with decrease in room and hall size.
Remove preset reverb default implementation from the config file.
Bug: 305866207
Test: atest VtsHalPresetReverbTargetTest
Change-Id: I565612d90b7c7e52cb23bd5dbfaf02235f65cc67
Limit the connection types to test to the following:
- HDMI*
- IP_V4
- USB
Only these connection types can be easily checked by the HAL
for presence of an external device.
Bug: 326888643
Test: atest VtsHalAudioCoreTargetTest
Change-Id: I659e14a150b3043ead8d844cd89a2c4700d57efd
As of VNDK deprecation, any libraries that defines vndk is no longer
valid anymore. This change removes all VNDK definition(s) from modules
which was VNDK.
Bug: 328994089
Test: AOSP CF build succeeded
Change-Id: Ifdade9c052db29c989ce7d804ba725361c4c0f8e
Merged-In: Ic4f19113c1a25dc1fb60cdfbf9c0f2d8d7d09726
The HIDL implementation blocked for the audio buffer duration
when the output side of the remote submix pipe does not have
the sink connected. This behavior was accidentally removed
when fixing b/327220024.
Also, limit the amount of the debug messages displayed when
the sink is shutdown to avoid spamming the syslog.
Bug: 328347445
Test: repro steps from the bug
Test: atest VtsHalAudioCoreTargetTest
Test: atest --test-filter=".*AudioPlaybackCaptureTest.*" CtsMediaAudioTestCases
(cherry picked from https://googleplex-android-review.googlesource.com/q/commit:6507376fb940198e8c39ad6d5345d132b1dd7219)
Merged-In: Ia66cb7b4567d42a41bf4715b7d725d36510ac50c
Change-Id: Ia66cb7b4567d42a41bf4715b7d725d36510ac50c
When calculating buffer size, use powers of two
for "large" buffers only on SRs >= 44100 Hz.
With lower SRs, sizes rounded up to a multiple
of 2 give too high latency.
Bug: 328432688
Test: repro steps from b/328432688#comment3
Test: atest VtsHalAudioCoreTargetTest
Change-Id: I605ef8507fef030d349e85c33ed0229c9566ddb5
The events about the status of A2DP and LE are sent by
the framework asynchronously with attempts to connect
an external device. As a result, the connection attempt
may be rejected by the bluetooth HAL module.
Bug: 327711955
Test: atest VtsHalAudioCoreTargetTest
(cherry picked from https://googleplex-android-review.googlesource.com/q/commit:a46d52771a3e4a7816ea722088153ee6b65f59c3)
Merged-In: Iecef0ee6a66cd290237fdcccd6a7016e9622c4b0
Change-Id: Iecef0ee6a66cd290237fdcccd6a7016e9622c4b0
Module base class can be used by multiple module,
to identify the module in use from logs, use mType in the logs as well.
Change log type for getVendorParameter and setVendorParameters,
as all parameters might not be supported by each module and unsupported parameter does not mean an error.
Change-Id: I0f46f8a9be733edd11186140bc9c3f51ea8d0969
