These interfaces are marked as vintf, so it should be used from both
system and vendor. So system/hardware/interfaces is proper place over
frameworks/base
Test: m
Bug: 249177028
Change-Id: I75d6cedd2b5d621d9e3831b3e59faf1d2387de74
Add Boolean, Byte, Double, Float, Long parcelables that
wrap the corresponding primitive type. This is similar
to the already existing Int type. Boxed types can be
made @nullable, passed as 'inout' parameters, etc.
Bug: 205884982
Test: m
Change-Id: I0328f0b3a37835bf1f320fc35156c078a42ca067
There are still issues with importing implicit versions. Switching to explicit versions while b/188871598 is worked out.
Bug: 242678729
Test: m; AIDL_TRANSITIVE_FREEZE=true m aidl-freeze-api && AIDL_FROZEN_REL=true m
Merged-In: If5a96c05b6448b588aa6326225dc6962ca7da6c7
Change-Id: If5a96c05b6448b588aa6326225dc6962ca7da6c7
(cherry picked from commit 56ec5d9ed5a02e7302b150c6f608ffdcb013c7bd)
This requires a version bump to android.media.audio.common.types.
cc_defaults and java_defaults are used in order to avoid explicitly updating version references in several Android.bp files when we do future version bumps.
Test: m
Merged-In: I481d4b4e7b4f1302d508ab6eb2c7ed4336a8f801
Change-Id: I481d4b4e7b4f1302d508ab6eb2c7ed4336a8f801
(cherry picked from commit 6088f4b7bb3d3f569328b7c332dd7bce0dcedbed)
It's still available from vendor.
Bug: 234181591
Test: m
Merged-In: I8e177d32852b548645a8bd1d7b7bb8a128c7b0bc
Change-Id: I8e177d32852b548645a8bd1d7b7bb8a128c7b0bc
(cherry picked from commit 7746c0e18967d728a673ed71562ddcf451d36d48)
As part of the AIDL API freeze for TM, the V1 of
android.media.autio.common.types was frozen. As a result,
unversioned reference to the interface now refers to the V2
of the interface, yet to be frozen. This causes ODR violation
where V1 and V2 of the interfaces are both used in the
platform.
To fix that, replaced the unversioned reference explicitly
with the reference to V1.
Bug: 225941299
Test: m
Change-Id: I1d4faa3cd541acdecacec5c711050d0e925ca36d
V1 of android.media.audio.common.types is frozen.
V1 of android.se.omapi is frozen as well.
Bug: 225941299
Test: m
Change-Id: Iebf09810195f09b8656aed56e63ee2050c257ab6
sountrigger.types now has V1 (frozen) and V2 (unstable).
`sountrigger.types` without the version suffix references the unstable
one and causes an ODR violation where there's a transitive dependency to
V1 (via other stable AIDL interface).
Bug: 225941299
Test: m
Change-Id: Ia7a3f002a768a4a879bb8495467c75cb11844459
Since APM XML configuration is being replaced with data
provided by the HAL, add data fields used to be in the APM XSD
schema to HAL structures. Summary of changes:
- Audio*Flags enum types moved to HAL types because flags
are specified for mixPorts in the APM config;
- AudioGainSys.useForVolume -> AudioGain;
- added AudioPort.flags because they are specified in the
APM config for mixPorts; since AudioIoFlags is a union,
it will be used for determining the mix port role;
- added AudioPortMixExt.max{Open|Active}Count;
- added AudioPortMixExt.recommendedMuteDurationMs;
- AudioPortConfigSys.flags -> AudioPortConfig
this is for symmetry with AudioPort;
- added AudioPortDeviceExt which encapsulates device
and encodedFormats information for device ports.
Bug: 198812639
Test: atest audiofoundation_parcelable_test
Change-Id: Id5fd537b11e43e85b3caa93b2cb3bc7f482a3199
a.m.a.c adds a version of AudioPort and AudioPortConfig structs
which are similar to Audio HIDL HAL V7, and the types they
depend on.
Bug: 198812639
Test: m
Change-Id: I40f15fd8e9adbcea7268f2645d2dfc2c4d704e9b
a.m.a.c adds AudioGain* types which are similar to
the ones from Audio HIDL HAL V7.
Bug: 198812639
Test: m
Change-Id: I6471d3a85fea81e45357f5d7e4363313bd36f7c7
a.m.a.c adds a version of AudioProfile struct which
is similar to Audio HIDL HAL V7.
Bug: 198812639
Test: m
Change-Id: Ibb4beb21f6c7975eb0400956bca10111f186280c
Move types that are known to be used by audio HALs
(those listed in h/i/audio/common/7.0/types.hal):
- AudioContentType
- AudioEncapsulationMetadataType
- AudioEncapsulationType
- AudioMode
- AudioSource (renamed from AudioSourceType to
match the name of MediaRecorder.AudioSource)
- AudioUuid
Bug: 198812639
Test: m
Change-Id: I3331d7e3d76db50f5b3c4db57891a20e4f935ad8
These types are intended to be shared with HAL interfaces and are
moved to a more generic directory/namespace and a stable build target.
There is no semantic change here.
Test: m
Change-Id: I4cedbab7e87d4895994ce8d26bd7e89f059a0103
... by switching from packageName/uid/pid to the Identity class.
This allows up to track which parts of the app trigger audio-recordings.
Test: atest CtsAppOpsTestCases
CtsNativeMediaAAudioTestCases
Fixes: 160150145
Change-Id: I0913a336a862e4a1cb38b9d967cfbdc490513ab0
Those are needed by new dependecies introduced from libaudioclient as
part of transitioning it to AIDL.
Bug: 160253486
Test: See main topic CL.
Change-Id: If4796aa91680290bcc40b496ed23911e498b4f9e
This moves the android.media.audio.common, android.media.permission and
android.media.soundtrigger_middleware AIDL files to a their own
separate aidl_interface module for Java.
Previously, because of how the framework JAR is configured, those
files were automatically compiled into it and we had to disable Java
for this target, creating a discrepancy between C++ and Java, which
would require us to similarly diverge any dependent targets if we want
cross-language support.
As part of this change, we also had to switch from using the
FileDescriptor type to ParcelFileDescriptor, as this is the only type
supported outside of the framework JAR and is also the recommended
practice.
Test: Manual verification of basic soundtrigger functionality.
Change-Id: Ib65c19dd7c7f5433cfa527fa74bb61ac677f5cf5
This change formalizes the permission enforcement patterns in the
sound trigger middleware layer. Every sound trigger session is
associated with an originator identity, established during attachment.
This identity is the used to authorize any operations performed on
this session, including data deliver via callbacks.
Temporarily, for b/w compatibility the existing SoundTrigger.java API
is preserved. Follow up changes will use the newer API, which requires
an identity to be provided.
Logging / dumpsys aspects have been modified to include the originator
identity associated with every call / session.
Change-Id: If83b151bd182af5a0dd98ff23dce252018de936b
Bug: 163865561
With b/152655547, all aidl_interface modules are considered as stable
unless it is explicitly with "unstable: true". This change marks the
aidl_interface that are not used across updatable module bounraries
as unstable, so that the build system does not run the API
dumping/checking on them.
Bug: 152655547
Test: m
Change-Id: I1257c66de6dd42b2d32d47ed74cb2878f79d14fb
This is not a stable API surface, so modules shouldn't be building
against it. Compiling against core_platform also disables link-checking,
which can hide other problems.
Bug: 137191822
Test: m
Change-Id: I7ce7dad8c02b094c565daf8c178bc6f89bb610ef
* changes:
Migrate SoundTrigger implementation to new service
Implement the soundtrigger_middlewware service
Add a permission for preempting sound trigger sessions
Sound trigger middleware service definition
Add audio.common types AIDL definition
This service is intended to replace:
frameworks/av/include/soundtrigger/ISoundTriggerHwService.h
This change only adds the replacement service, follow up
changes migrate the clients to use the new service and remove
the old one. The new service is feature-equivalent to the new
one, but offers the following advantages:
- AIDL interface (as opposed to hand-written parceling code).
- Pure Java implementation all the way to the HAL.
- Better documentation.
- Rigorous error handling.
- Unit tests.
- Reduced code complexity (less layers, better separation of
concerns).
- Permission-based security model (as opposed to some baked-in
assumptions about process affinity).
Change-Id: I79f4eff105d3e6245990be068b933d4d48c35a0d
Bug: 142070343
These are the AIDL files that define the sound trigger middleware
interface.
This service is intended to replace the existing
frameworks/av/include/soundtrigger/ISoundTriggerHwService.h
using AIDL in order to avoid a large amount of hand-written
parceling code and other forms of boilerplate, and provide
cross-language support.
Change-Id: Ia783ba4f1ea7335a984396e7024cac0411699403
Bug: 142070343
We have decided to reuse the existing annotation @SystemApi for all Java
APIs regardless of whether they are for apps or platform internal
modules. This was because introducing new annotation types every time
when we have to create new API surfaces will only increase the confusion
without giving much benefit.
Instead, to differenciate the different API surfaces of @SystemApi, the
annotation type is parameterized. Specifically, it has to axises.
client: specifies the intended client of the API.
process: specifies the process(es) that the API is available.
The default for client and process are priv-apps and all, respectively,
which corresponds to the today's @SystemApi for privileged apps like
GMS.
Bug: 140202860
Test: m
Change-Id: I3305b71e22970e80db95f3daf3d7713603c7d68d
This is a subset of the types defined in
hardware/interfaces/audio/common/2.0/types.hal.
Specifically, the subset that sound trigger depends on.
These files were auto-generated using the hidl2aidl tool, and
later manually edited to workaround some missing features in
the conversion tool or AIDL (mostly around enum support).
Change-Id: I1c0ffe8a2ebaf09eb4516beaaf4b6fc80985bfc9
Bug: 142070343
droiddoc modules for the SDK API documentation and stubs library
generations have depended on the 'framework' (which was recently changed
to framework-minus-apex' module to get the list of Java source files to
be processed.
This however caused a circular dependency when we tried to modularize
some classes in the framework library as a separate library. The
separate java library depended on the stubs library (because it should
only use SDK APIs) and the stubs library depended on the framework
library. The framework library itself depended on the separated library
(or its stub) to use APIs from the separated library, thus forming a
circular dependency.
This change fixes the problem by directly giving the framework source
files via a filegroup `framework-sources-to-document` where all Java
and AIDL files that are to be documented are included in.
This change also put the generated R.java and Manifest.java files from
framework-res into the filegroup for framework sources.
Bug: 70046217
Bug: 135922046
Test: m
Change-Id: I09ad88da47540d31ad089aad5e1151a4b6877ec2
When there is no *.aidl file in a filegroup, include path for the aidl
compiler is not set to the path property of the filegroup. This blocks
the plan to cut the dependency from API stubs libs to the framework,
because then we have to rely on the include paths deducted from the
input filegroups, rather than AIDL include paths exported from the
framework.
To make that happen, this change first removes the *-without-aidls
filegroups and replace the reference to them with the corresponding file
groups having AIDLs. Adding the AIDL files for media apex to the API
stubs is fine because the AIDL files are @hidden ones.
Bug: 70046217
Bug: 135922046
Test: m
Change-Id: I794f6d0a7bcbef311d345d6e18ca311800af8ef4
The path property specifies the base directory for Java files. This can
be used by the build system to filter (or filter-out) Java files by
their package names.
Bug: 135922046
Test: m
Change-Id: I6bece92e702b560e53bc087b5960bff577dc3175
(cherry picked from commit 7954524e3cc618830f0bf7b23687bf222ecd285b)
Bug: 134566750
Test: m
Change-Id: I0260942108a13cc5e9bed1d6005abfb8daffe0ac
Exempt-From-Owner-Approval: minor build change
