New stream type: AUDIO_STREAM_ASSISTANT
This is intended to be used by a virtual assistant like
Google Assistant, Bixby, etc.
The audio stream has own volume alias and the volume
does not change by volume changes of other streams.
Bug: 123745215
Test: make
Change-Id: I17ec0480cad4c3507a97454ccb832255adda9897
Merged-In: I17ec0480cad4c3507a97454ccb832255adda9897
Signed-off-by: Baekgyeong Kim <baek.kim@samsung.com>
This was causing a null deref in cts/security/gce-all.
Bug: 142652210
Test: Run cts/security/gce-all, observe no more crash.
Change-Id: If2003b650f69e88825219113ce1b7d70edc4731d
AUDIO_SOURCE_DEFAULT is translated to AUDIO_SOURCE_MIC in
AudioPolicyManager. That means AUDIO_SOURCE_DEFAULT should behave
as same as AUDIO_SOURCE_MIC.
In this commit, argument for addInputEffects() is translated for
avoid different behavior between AUDIO_SOURCE_DEFAULT and
AUDIO_SOURCE_MIC.
This means same preprocessing effects are applied for both Sources.
Bug: 137898521
Test: Add pre-processing effect for AUDIO_SOURCE_MIC then check if
it works for AUDIO_SOURCE_DEFAULT as well
Change-Id: I3df1bcc7309dc535a721113cc4465b20733b4058
Merged-In: I3df1bcc7309dc535a721113cc4465b20733b4058
DIRECT and OFFLOAD Audio threads use the mixer when using SW effects.
This requires them to use the default sample rate.
Test: vendor specific, see bug
Bug: 147209612
Merged-In: I553fe43695d64160dd64808e18299699a4047225
Change-Id: I553fe43695d64160dd64808e18299699a4047225
If ENFORCED_AUDIBLE is set for a stream, it should output
to Speaker and not select output device(s) for VOICE_CALL.
Bug: 147209610
Change-Id: I9553efd677979351ed6f7191079da84dac2a7ed1
`realloc` doesn't deallocate `param` if it fails. Since the loop ends in
a `free`, we just need to avoid overwriting `param` until after the
check.
Caught by the static analyzer:
frameworks/av/services/audioflinger/Effects.cpp:1868:9: warning:
Potential leak of memory pointed to by 'param'
[clang-analyzer-unix.Malloc]
Bug: None
Test: TreeHugger
Change-Id: I58fe4319927b3ea99989e1fb1dc2dabe89c72ef8
For the formats compatible with IEC61937 encapsulation, assume
that the record thread input from MSD is IEC61937 framed (for
proportional buffer sizing). Add the
AUDIO_OUTPUT_FLAG_IEC958_NONAUDIO flag so downstream HAL can
distinguish between raw and IEC61937 framed streams.
Test: manual test on BCM TV device
Change-Id: I8f09f936ccce63985df4ad00433299b063769419
Use stable C interface instead of C++ interface.
Bug: 123543273
Test: compilation, boot, dumpsys
Change-Id: I04935343917ce400d0df531e06b4d23ab7269fa6
Merged-In: I04935343917ce400d0df531e06b4d23ab7269fa6
When converting audio attributes to stream type, do not log an
error if no product strategy support USAGE_VIRTUAL_SOURCE.
Bug: 146558301
Test: check logs after using AudioPolicy for injection
Change-Id: I8313b9d27583934e7a64edbfe49d52bad587f372
* clang-tidy segmentation fault when checking
util/C2InterfaceHelper.cpp in multiple threads.
Test: make with WITH_TIDY=1 and tidy checks for nightly builds
Bug: 146435095
Change-Id: I6487a69636d4cdc1d3449ceb193d347d402beea6
Root Cause:
Google C2 H264/H265 decoder request 64-aligned stride from graphic block; but ARM GPU would return
with 128-aligned stride when input height ONLY satisfies 2-aligned (e.g. 130).
Solution:
Revise stride alignment from 64 to 128 of C2 H264/H265 decoder
Bug: 142924202
Test: Build C2 Codec
Test:
run cts -m CtsMediaTestCases -t android.media.cts.ImageReaderDecoderTest#testGoogH264Image
run cts -m CtsMediaTestCases -t android.media.cts.ImageReaderDecoderTest#testGoogH264ImageReader
run cts -m CtsMediaTestCases -t android.media.cts.ImageReaderDecoderTest#testGoogH265Image
run cts -m CtsMediaTestCases -t android.media.cts.ImageReaderDecoderTest#testGoogH265ImageReader
Change-Id: I6eaff1b858e031b64744bc67d8aee5cc51cfd92d
As we described "libavservices_minijail" in Andorid.bp, we don't need
to keep "libavservices_minijail_vendor" as a workaround, but may make
"libavservices_minijail" vendor_available and remove
"libavservices_minijail_vendor".
We will keep "libavservices_minijail_vendor" until we replace
"libavservices_minijail_vendor" to "libavservices_minijail" from all
vendor modules.
"libavservices_minijail_vendor" will be removed once the vendor
modules are cleaned up.
Bug: 146313710
Bug: 145580164
Test: build and boot
Change-Id: Iffb16c10ddc93e645d5839ac8aa320353e46d224
This patch avoids pseudo LCM buffer calculation oversized buffer
causing posix_memalign error in cases of large play/record frame
counts (such as 24596) for PassthruPatchRecord patches.
Test: MSD to Primary audio patch with play/record frame counts
of 24596.
Change-Id: I3a2f3795330e99984388b01033d8e7c1685d01f2