Currently vendor APEX can link to VNDK when 'use_vndk_as_stable'
property is set as true. However, when VNDK is deprecated, all former
VNDK libraries should be included in the vendor APEX despite of property
value. This change ignores use_vndk_as_stable when VNDK is set as
deprecated.
Bug: 290318998
Test: aosp_cf_x86_64_phone with WITH_VNDK=false build checked that
Vendor APEX contains all required VNDK libraries within the APEX
Change-Id: I648277d734274e7852b3effc24e7780e55089f75
Adds dylib support for vendor and recovery images.
This changes the default linkage for vendor and recovery images to
dylib, which matches the platform default linkage. This also means that
by default, dylib-std variants are used for rlib dependencies.
Bug: 204303985
Test: Soong tests.
Test: m dist vendor-snapshot
Test: RECOVERY_SNAPSHOT_VERSION=current m dist recovery-snapshot
Change-Id: If84074b8615a70c45e7e162abeb853dc8c34d49a
In the past, dexpreopt for boot jars was very inflexible, and it was
incredibly hard to make a change that is as simple as adding a jar to a
boot image. Boot image generation was handled by
"platform_bootclasspath" and "bootclasspath_fragment" separately. This
caused not only code duplication but also the inflexiblity as such a
design did not fit today's use cases, where a boot image may take jars
from multiple mainline modules and the platform, and a mainline module
can contribute to multiple boot images. The design casued a huge
maintenance burden as any change to the boot image cost multi-week
efforts.
In recent years, efforts have been made to improve this a bit by a bit.
This change is another step towards making dexpreopt reasonable.
After this change, all boot images are generated by "dex_bootjars",
which is in build/soong and is therefore available on both the full
source tree and the thin manifest (master-art). The change decouples
profile generation/extraction from boot image generation. Profiles for
mainline modules are still handled by "bootclasspath_fragment"
because they need to be packed into APEXes when building mainline
modules and extracted from APEXes whem building the system image from
prebuilt modules. Boot images are not handled by
"bootclasspath_fragment" anymore.
Bug: 290583827
Test: m (all existing tests are still passing)
Test: Manually checked that the boot images are exactly the same as
before.
Change-Id: Ib5a5f401bee334ffcab5c26618e0c8888b84575a
This is to decouple profile generation from image configs. A boot image
profile is either for a mainline module or for the platform, which is
orthogonal to boot images.
Bug: 290583827
Test: m nothing
Change-Id: I0918f2fa945a2af1839f1a08ecede331c64d0317
The Modules() method was for the ART boot image in the ART APEX. It's no
longer needed because we don't install the ART boot image in the ART
APEX anymore.
Bug: 290583827
Test: m nothing
Change-Id: I674a5fa1ed18908413a64129df5947c7d26e638e
This CL is to prepare for the changes in http://r.android.com/2652081.
See the description of that CL for the reasons.
Some tests were set up in a fragile way and were easy to be broken when
the implementation changes. This CL is a pure test change that fixes
them.
This CL also removes TestNoUpdatableJarsInBootImage. That test checks
what should go to the ART boot image and what should go to the
platform one, which does not apply today because boot images are not
associated with mainline modules anymore: a boot image may take jars
from multiple mainline modules and the platform, and a mainline module
can contribute to multiple boot images. In practice, we have ART jars in
the platform boot image, and we are going to add core-icu4j and
consrypt to the ART boot image, which is now for testing only.
Bug: 290583827
Test: m nothing
Change-Id: I22c45cbf6f853b030b68edb51197854e9c53a02e
We will deprecate flattened apexes. In this change, GSI-specific make
variable (PRODUCT_INSTALL_EXTRA_FLATTENED_APEXES) is removed. The
variable was used to install both image/flattened apexes in the GSI, so
that it works on ro.apex.updatable devices and not-updatable devices.
Now, GSI will have only image APEXes in it.
Bug: 278826656
Test: lunch gsi_arm64-userdebug && m # no flattened apexes
Change-Id: Ie5fe505b9f2a6a1687252513788fd3e16bc147c5
deapexer doesn't need it.
Bug: 279858383
Test: presubmit
Test: m MODULE_BUILD_FROM_SOURCE=false droid
Change-Id: Ie8311c74682f366cf6aeebb669500d01675cd0bf
Bug: 285138555
Test: b build //build/bazel/examples/... --config=android
Test: b test //build/bazel/tests/... --config=android
Change-Id: I7350930d48ccd13e96bdb4f994f47071a0dde70f
Bug: 285138555
Test: b build //build/bazel/examples/... --config=android
Test: b test //build/bazel/tests/... --config=android
Change-Id: I11372cbc01b89d530a35f90375ca6fa87834edd2
This property allows modifying the variant version that is put into the
APEX manifest.
Bug: 285138555
Change-Id: I42b8c0ddad8822a0a329e99ec4ec62f8c3546ee9
This property allows developers to create multiple modules that provide
the same APEX implemenation with some differences. For example, an APEX
that targets the current SDK can be developed alongside an APEX that
targets a pre-S SDK. Common dependencies of these APEX modules can both
use the value of the apex_available_name property in their
apex_available lists and be packaged in either APEX module.
Bug: 285138555
Change-Id: Ia2cb027e2f19bd6d3d1c02e3d3742a0122a43c97
Since vendor APEXes are installed in vendor partition, there shouldn't
be system_file according to Treble. Instead, we can force-label / and
/apex_manifest.pb as vendor_apex_metadata_file so that apexd (and other
system components) can still read them.
Bug: 285075529
Test: m nothing (soong test)
Change-Id: Idb36b8c4c68b29e2235dbda38ee323d4b781e1d5
AndroidMk assumes that the app is the last file installed, and it uses
this assumption to populate the LOCAL_SOONG_INSTALLED_MODULE entry. This
CL moves the privapp_allowlist installation to before the app
installation to respect this assumption.
Bug: 242509786
Test: go test
Test: OUT_DIR=out.ref m nothing &&
cp aosp/2562351 && OUT_DIR=out.change m nothing &&
GOWORK=$PWD/build/bazel/mkcompare/go.work \
go run android/bazel/mkcompare/cmd -json \
<(sed -e "s/out\.ref/out/g" out.ref/soong/Android-aosp_cheetah.mk) \
<(sed -e "s/out\.change/out/g" out.change/soong/Android-aosp_cheetah.mk)
&& verify manually that the only diffs are related to the removal of
the prebuilt_etc module.
Change-Id: I95ec27070f575e79fb976de68493a219717ed89a
For a non-override android_app, we can assume that the privapp_allowlist
already contains the correct package_name, and so we don't need to
overwrite it in this case.
Bug: 242509786
Test: go test
Merged-In: I0f137e34cae3478dc8b9178d138121ff1d936f07
Change-Id: I0f137e34cae3478dc8b9178d138121ff1d936f07
Verify logic for transitive shared vs static deps to
ensure in Bazel that these are handled similarly.
Test: go test soong tests
Change-Id: I015935fa4d7eb1ac40666d47a8de8e48150a043e
* changes:
Dexpreopt ART jars and framework jars together.
Unify installDirOnHost and installDirOnDevice.
Remove the ability to install boot images by bootclasspath_fragments.
Prepare tests for dexpreopt changes.
We don't need this anymore because we are going to compile ART jars and
framework jars together.
Bug: 280776428
Test: m
Change-Id: I070157530449a1bb5779e25984c367df3dde7b36
After this change, there is a clear separation between tests that are
related to dexpreopt and tests that are not. The former uses
PrepareForTestWithDexpreopt, while the latter uses
PrepareForTestWithJavaDefaultModules. The benefit is that the latter
will no longer affected by any dexpreopt changes.
Bug: 280776428
Test: m nothing
Change-Id: Ib957765b9287d51c082e0a33cee17a6bb56daeef
* changes:
Special case platform variant of bootstrap libs
Select stub/impl per apex variant
For test apexes, base_apex_name is the api domain
Broaden the granularity of config_setting from apex_name to api_domain
Print default val if all vals in axis match default val
apex_test do not "override" the source apexes that they test. However,
similar to override_apexes, test apexes have the same path
(/apex/<apex_name>) on device. Instead of creating another property,
reuse the existing base_apex_name property. The use case for this will
be for creating selection statements for stub/impl selection.
Test: go test ./bp2build
Change-Id: I4b7548e0e0fc920e407e1c5e5c00e87efc6e70c9
so that Vendor APEXes don't embed the LLNDK prebuilts (which are just
stubs of LLNDK libraries).
Bug: 280697209
Test: m nothing (soong test)
Change-Id: I9e6c123e73fa0ab56d8494a01652ee32a9b6a6cd
Since these apps explicitly set com.android.cellbroadcast in their
Android.bp files, they do not need to be present in the baseline
apex_available map.
Test: m nothing
Bug: 281077552
Change-Id: Ibea86bfbc7e460e27c5e780b611674e66e9cf727
These libraries now set the correct apex_available in their Android.bp
files. Therefore, they do not a separate entry in the baseline
apexAvailable map.
Test: m nothing #passes
Bug: 281077552
Change-Id: Icf6b97c3bcc4bda96eefa14ba6d4bf10614bc378
These are not needed in this map since the Android.bp files for these
modules are created by pom2bp script
which explicitly set apex_available as ['//apex_available:apex',
'//apex_available:platform'].
Bug: 281077552
Test: m nothing
Change-Id: I53bb9a104b1e7a2535fe568e27431b1a48a885e3
This change allows override_android_app to use the same
privapp_allowlist as the non-override module so that they will always
remain in sync.
Test: go test ./java -v -run TestPrivappAllowlist
Test: go test ./apex -v -run TestApexWithApps
Test: m com.android.permission com.google.android.permission and verify
manually that apex_payload.img contains correct privapp_allowlist
Test: m com.android.permission before and after change &&
`diffoscope apex_payload_reference.img apex_payload_with_change.img`
&& verify that there are no semantic changes
Bug: 242509786
Change-Id: Ifdcb28af40763aed7a4aac9a7f681153554bc256
This effectively undoes both If6c3ee82d588e2742c85cef7244c090c93f38b8e
and I682e4f1f477f3024f7719dfaa67006ef335e0640. SDK variants are now
installed again, which will fix unbundled builds of cc_test modules.
The platform variants used by com.android.virt are now packagable
even though they are not installable.
Fix the original problem in b/194403710 by adding a flag to platform
variants of modules in apexes that are not platform available, and
using that to prevent install and packaging dependencies. That
allows the HideFromMake flag to go back to being used for preventing
install dependencies but not packaging dependencies.
Test: TestPackagingWithSkipInstallDeps
Test: TestFileSystemShouldInstallCoreVariantIfTargetBuildAppsIsSet
Test: TestFileSystemShouldSkipApexLibraries
Bug: 194403710
Bug: 268582372
Fixes: 274443025
Change-Id: If5418df3ddbb940bd631caebdf38daa81e71f40e
Prevoiusly, we were partially correcting for override modules in
bp2build/mixed builds in some but not all places. Now we always check
for override modules and ensure that Bazel_module properties are
propagated properly for override modules.
Bug: 279609939
Test: go test soong tests
Change-Id: I5445aa71f4c8013315415a2ca9ab9c6b3be6bce0
As-is, dexpreopt config and bootjar is in the dir including
DeviceName(). It causes unnecessary dexpreopt invocation when target is
changed repeatly. To avoid it, rename dir to common name.
Bug: 278833696
Test: m
Change-Id: I2818d9ae09e6c22ca6989cc8efdb70b470ec502c
If a library contributes to an API surface, it will have only a single
copy on device. Therefore, we should disallow installation to muliple
apexes/platform.
There are some exceptions to this rule today, and they have been relaxed
using allowlists.
Bug: 277651159
Test: go test ./apex
Change-Id: Ice3023ecd28412a2610d8b98628cb727b58c5c3b
The core problem I am trying to solve is making sure that stub
libraries in Bazel have a single apex available. In Soong, this will be
enforced using some graph walk which is not easy to port to Bazel.
However, we might need to revisit this when we build the enforcement
mechanism in Bazel. We likely need a `test_for` on the top level
apex_test Bazel rule so that the test apex is _allowed_ to link against impl
of the library it is trying to test.
(This CL retricts this to cc_library, I can expand this to other modules
if this is the right approach.)
Bug: 277651159
Change-Id: Iaeec22c5626df79a33785c766ed29102b1da403e
For now ext4 is the only supported fs_type.
Bug: 279689445
Test: m apex.apexd_test_erosf
Test: m apex.apexd_test_f2fs
Change-Id: Idd1b2e61e25110da6d3805b082195e6ec031f573
This reverts commit 9551384f65.
5bedfee448 is a better solution for the
same problem.
Bug: 279247159
Test: m checkbuild (presubmit)
Change-Id: Ifd3117a4450b5275c09258d1ce896adb41a75158
These should be handled by Makefile dependency (LOCAL_REQUIRED_MODULES).
Bug: 278832320
Test: build and see symbol files
Change-Id: I80ad9d8e7e8384e2ae3537cc681bc1d1b327d04c
If any of apexes in apex_available is an apex_test, then that name will
be propagated down from that apex to each apex variant. This metadata
will be used to enforce that stub libraries cannot have more than one
apex_available.
This logic is necessary so that bp2build can select the correct
stub/impl.
(To avoid replicating this complexity in Bazel, we should consider
dropping the test apexes in Bazel BUILD files, next CL)
Bug: 277651159
Test: go build ./apex
Change-Id: I63617c1dc2a2d5c9cd7758c416fec7b4db1f10a7
Apexer rule with symbol files causes "no build rule" error when the apex
is replaced with prebuilt.
Bug: 251299786
Test: m checkbuild (presubmit)
Change-Id: I7390f72c321beaeb49fcee4e02a219c07f184664
Teach rustc to generate object files instead of a linked executable
for binary and shared library crates. This lets us skip the lengthy
Rust compilation phase when only the C++ dependencies of these crates
have changed.
This works using a Python script that replaces the linker invoked by
rustc and converts the linker command line into a rspfile that can be
used during the link phase. The script makes certain assumptions about
how rustc invokes the linker, but since we control the version of Rust
we use, we can update the script if those assumptions ever break. I've
asked the Rust developers about adding an official way to do this so
hopefully we won't need to rely on this hack in the future.
The rustc rule for Kythe has been altered to pass linker=true, somewhat
consistent with the main rustc rule, which now doesn't invoke the actual
linker either. `m xref_rust` still succeeds, but hopefully we can find
someone who knows more about the rest of the Kythe pipeline who can let
us know whether it should still work. Otherwise, let's use this CL as
an opportunity to find out whether the rest of the pipeline is happy
with linker=true, since otherwise we will find it hard to maintain the
Kythe rule in the future.
Change-Id: Ifb9c8689c1bd78c0afcf6fdfd111140d20377536
Updatable apexes should set a min_sdk_version that is not current. This
enforcement does not exist today, but will soon be added to the build
system. In preparation for that, cleanup the min_sdk_version property
used in this unit test.
Test: go test ./apex
Bug: 221087384
Change-Id: I8a163595f85cb0f8dabccc3321ef210215675714
Updatable apexes should set a min_sdk_version that is not current. This
enforcement does not exist today, but will soon be added to the build
system. In preparation for that, cleanup the min_sdk_version property
used in this unit test.
Test: go test ./apex
Bug: 221087384
Change-Id: Idabc4c6f20e134b38ed0ae3ce68f6b7d54d7a13c
.. as a validation for apex.
apex_sepolicy_tests checks the contents of the apex against
precompilec_sepolicy. This is to avoid common mistakes in file_contexts.
For example, etc/vintf directory should be read-able by servicemanager.
Bug: 267269895
Test: m com.google.cf.wifi (with wrong file_contexts entry)
Change-Id: Ic363345df43e5b2108b07de787f5f6817bbe6af4
libc_hwasan is a new library in the runtime apex that lives in
bionic/hwasan/libc.so and is symlinked to /system/lib64/hwasan/libc.so.
This is chosen by the linker if an app or binary requires HWASan
support.
Bug: 276930343
Change-Id: If331744ad84241ad99a41805ea3110d37cf9b0af
Otherwise when other modules use the apex as data (for example, in
tests) they get the full path from the root of the workspace to the
apex.
Fixes: 276416485
Test: atest timezone_data_e2e_tests:com.android.tests.apex.TimezoneDataHostTest#testStageActivateUninstallApexPackage -- --abi x86_64 then tree out/host/linux-x86/testcases/timezone_data_e2e_tests
Change-Id: Id740856794770df2edc55dfff5f6f938fc31d0c5
There are no tests to verify the generated action for the canned fs
config entries, so add some.
Also update the prop's docstring to reflect the actual logic.
Bug: 275209284
Fixes: 275280970
Test: soong tests
Change-Id: I37f2a8640bf4c307068a77db7a635c9bbeb9f38f
This includes apex_test and apex_vndk. Rely on mixed builds module
allowlist to filter further.
Test: presubmits
Test: m test_com.android.adbd; showcommands test_com.android.adbd | tail -n1 | grep bazel-out
Change-Id: Id73db0917ad4a6dd9c0646d5186753ed940e735f
This relands aosp/2457063. The original change broke T and U since those
branches still contain soong modules of type (kind+level). Those soong
modules have been cleaned up now
Test: Used go/abtd to test T and U branches with this change
Bug: 208456999
Change-Id: I0ef7933c055f88cb512a02108f1173e51156ef1c
For vendor apexes, use_vndk_as_stable: true means the apex is tied with
the current VNDK version while min_sdk_version: N means the apex should
work with N+ VNDK versions. Setting both properties together doesn't
make sense.
Bug: 274396342
Test: m nothing (soong tests)
Change-Id: I00dcac43f95ce6a90b3f3273ae4611a300ea5028
Add skip-sdk-check to skip checking the SDK version when extracting an
APK/APEX from an App Set Bundle. This can be used when the platform SDK
version is not defined and the APEXs/APKs use SHA based SDK versions.
This check should not be set to true for non Beta dessert releases
Bug: 274518686
Test: # Add SHA targeting modules to platform
m SOONG_SKIP_APPSET_SDK_CHECK=true #Build Passes
m SOONG_SKIP_APPSET_SDK_CHECK=false #Build Fails
m #No config supplied, build fails
Change-Id: I1919437d3410f09c991e1de39031bd88e1f8246a
* changes:
Revert "Create two sentinel api levels"
Revert "Create EffectiveVersion* functions for ApiLevel"
Revert "Prework for migrating min_sdk_version from (kind+level) ..."
Revert "Always include host variants in the sdk snapshot"
Revert "Update min_sdk_version from SdkSpec to ApiLevel"
min_sdk_version signifies device version and does not need an sdkKind to
describe it fully. Update the type and cleanup existing usages. As a
side benefit, we also get better error handling since users can no
longer enter something like `public_30` as a valid min_sdk_version in bp
files
Will do a similar cleanup for targetSdkVersion and maxSdkVersion in a
followup CL
Test: m nothing
Test: no change in ninja files (this should be a no-op)
Bug: 208456999
Change-Id: Ie6ae7e267d093c5e4787e82685daaca1021d202e
Vendor APEXes can be updatable unless it sets "use_vndk_as_stable:true".
When "updatable:true", it should meet the "updatable" requirements like
"should set min_sdk_version".
Bug: 274396342
Bug: 274041915
Test: m nothing (soong tests)
Change-Id: I746d8941c21e883c1eba3e1f1e5a414a67db071a
This CL adds a few things:
1) Populate the filesInfo struct with cquery'd information from an
apex's ApexMkInfo provider. This filesInfo is then used in
apex/androidmk.go to generate Make modules (soong_cc_rust_prebuilt.mk),
which are then used in packaging to generate zip files of symbols in $PRODUCT_OUT.
2) Make a list of dicts of primitives JSON-encodable.
3) Tests.
Bug: 271423316
Bug: 271423062
Test: presubmits
Change-Id: Iaa34f51044de310510e580d9cf1fe60bbef801c1
When evaluating whether to build the VNDK for the multi-target
configurations, we construct the API level based on the VNDK
version. When the VNDK version is 'current', we replace the
version name with the platform SDK version e.g. 'UpsideDownCake'.
For architectures that currently have a MinApiArch of FutureApiLevel
(i.e. 'current'), comparing 'UpsideDownCake' with 'current' will
assume 'current' is at a higher API level and skip building the
VNDK.
Before perofrming the API level comparison, we check that the VNDK
version does not match the platform VNDK version. If true, it is
necessary to disable the VNDK when it does not meet the minimum API
level for the architecture. If false, we skip this check and allow
the VNDK to be enabled, as this implies we are targeting the 'current'
VNDK version and should include the target platform VNDK.
Test: m (soong test) & boot
Change-Id: Iacc33d2858fef670ddfdff5bbde9d32b084979c9
override_apex's bp2build converter had a bug where it was looking at
the product variables for the override_apex module itself instead of
for the base module it is overriding.
Fixes: 271424349
Test: go test
Change-Id: If1e2653d3751fa908faf0ab97dfa2e943ebe98ec
Provide a list of the installable apexes in the tree, so that the
multi-tree orchestrator can generate rules for packaging.
Bug: b/266729952
Test: unit tests pass
Change-Id: If2b2585a7d14aed0618ddbacd116bc5728109d87
strict_updatability_linting currently only makes it so
that NewApi can't be added to the lint baseline.
However, since we're updating NewApi to work on a lot
more apis than before, we need to baseline many issues
across the android tree. Temporarily disable
strict_updatability_linting so that we can add these
baselines.
Bug: 193460475
Test: Presubmits
Change-Id: I8d92df95a46e9b903f0cc0e3be56f17722c50430
When set true, use file_contexts file as is and do not force-label / and
/apex_manifest.pb
Bug: 268439003
Test: m (soong tests)
Change-Id: I252eef9569deeac87bb71a3425a8eb07bab3c634
The type of min_sdk_version is being migrated from
android.SdkSpec(kind+level) to android.ApiLevel(level). This affects
`ShouldSupportSdkVersion` for java modules. This function skips the
check for modules compiling against `core`, and that requires access to
SdkVersion and not MinSdkVersion after the migration.
Skip the check explicitly using SdkVersion.
Test: go test ./java
Test: No change in ninja file
Bug: 208456999
Change-Id: I14eca4f8e8c5d7477ded00c4fe54097323fab4a2
Implement APEX stub of API surface so any stub can be replaced with API
surface when APEX stub interface is required.
Unlike other stub interface, APEX stub can be decided if it should be
used after APEX postdeps mutator analyzes which modules should be
included in which APEX. To cover this, APEX stub is being added to the
dependency if the dependency should not be covered with LLNDK or NDK
stub, and APEX stub exists. From depsToPaths, if dependency to both
original module and API library exists, then it choose one of the
dependency and ignore the other.
To cover this logic, a new property is added to the api_surface :
apex_libs. This is introduced as it is difficult to
gather all api library with apex stub before DepsMutator.
Bug: 264963986
Test: cf_x86_64_phone_vendor build succeeded
Change-Id: I9f0b1f70968e32eba94d3e0d7bb1f4bb29ff2438
There are no current users of overriding rros, so it's dead code.
Additionally, there is some evidence of accidental misuse, so remove it
for now as we don't want arbitrary properties of AOSP modules to be
overridable in the general case.
Bug: 270542561
Bug: 269568074
Test: presubmit
Change-Id: I030abd644358baefc0d1543649c8a2a81a7491d0
This reverts commit be90172f0a.
Reason for revert: not used, and not desired
Bug: 270542561
Test: presubmit
Change-Id: I415440dbae965f3c07c6f3ccf6ab4bc9ec57fbf6
This reverts commit dd2c76e772.
Reason for revert: This was submitted without build system team +2 with no rationale for why it's needed.
Change-Id: I86b9849122ec2b7a382d4f39bbbc0ea1ff70c4bc
This property allows an APEX bundle to ignore the apex_available rules
defined in its dependencies, effectively allowing it to use any APEX
as its dependency.
Bug: 269660351
Test: m
Change-Id: Ib2797e04438ad908ac10b256dce07a7819e836b3
Previously, the symlink optimization for APEXes assumed that the target
of the symlinks are in the system partition. The assumption however
doesn't hold always because the file that was added to the APEX might be
with system_ext_specific: true or vendor: true.
Bug: 265598720
Test: m nothing
Change-Id: Ieb9a6769320c0ec697a88c0cae977e7d65288362
* changes:
Use the profiles in the APEX to dexpreopt system server jars.
Add tests for the ignored "profile_guided: true" property.
Output dex_preopt to SDK snapshot.
After this change, if "profile_guided: true" is set, profile-guided
compilation will be enabled for the jar and the ".prof" file next to
the jar in the APEX ("javalib/<name>.jar.prof") will be used as the
profile when dexpreopting for the prebuilt APEX.
Bug: 241823638
Test: m nothing
Test: -
1. (on internal master) Patch aosp/2426453.
2. Build the APEX bundle and the module SDK of com.android.wifi
3. (on tm-dev) Patch this CL and aosp/2141972.
4. Copy the APEX bundle and the module SDK built on step 2 to the
source tree
5. Disable hiddenapi check
6. lunch cf_x86_64_phone-userdebug && m MODULE_BUILD_FROM_SOURCE=false com.google.android.wifi
7. cat out/soong/.intermediates/prebuilts/module_sdk/Wifi/current/prebuilt_service-wifi/android_common_com.android.wifi/dexpreopt/oat/x86_64/javalib.invocation
8. See the profile being used.
Change-Id: I55a5a295e9c5d6f0564afb139c5fb7da91ab8cae
The tests in bootclasspath_fragment_test setup the environment to have
two APEXes: a source one and a prebuilt one. Before this change, the
tests for prebuilt incorrectly check the contents in the source one.
This change introduces a new function that properly checks files
deapexed from the prebuilt one.
Bug: 241823638
Test: m nothing
Change-Id: I1865c20b198d50e7ebc8ebfb9f7c71394a225ab7
filesInfo should not be used for an apex that's converted to Bazel,
since all information should come from the Bazel starlark provider.
filesInfo is populated when there's a full Android.bp module graph of
the apex deps.
Test: presubmits
Change-Id: I4b9c596ca7841f4ea76e21f929039282113b6752
So that it can be used in the bazel implementation.
Bug: 268006095
Test: m nothing, check that it's in out/soong/soong_injection
Change-Id: I1520dd874076dee7fa083648d0790b060d658e5a
This reverts commit d1209a89c7.
Reason for revert: apex_name was still in use in the internal/intermediate branch. Fixed(removed apex_name) in the internal branch.
Change-Id: I10b8e0cfdb01dcdb6b623bddf9c04f4dac71fc90
After this change, if profile-guided compilation is enabled for an APEX
system server jar, the profile will be installed next to the jar with
the ".prof" suffix, ("javalib/<name>.jar.prof"). This file will later be
used by odrefresh and dexpreopt from prebuilt APEX.
Bug: 241823638
Test: m nothing
Test: -
1. Patch ag/20581649 PS2.
2. banchan com.android.btservices x86_64 && m
3. Check that "javalib/service-bluetooth.jar.prof" exists in the APEX.
Change-Id: Ibcfc6257dade92bd40a6d4b7368148717d0863b9
apex_name was used to install symbol files under /apex/<apex name>,
which caused duplicate rules when there's two apexes with the same
apex_name. (e.g. override_apex or multi-install apexes)
Since symbol files can be looked up via ELF Build-Id, we don't have to
install symbol files with the actual "apex name", but instead we can
install them with soong module name.
Bug: 267581665
Test: m (soong test)
Change-Id: I1cbe5c86d287f180c7deef93ba85ea5552680011
Previously in mixed builds, only the tidy files for the boundary module
were built, whereas all of its transitive dependencies' tidy files were
not being built. Instead we should export the list of transitive tidy
files for a module so that we can run clang-tidy for the boundary module
as well as its dependencies.
Bug: 195029134
Test: WITH_TIDY=1 DISABLE_ARTIFACT_PATH_REQUIREMENTS=true mss tidy-packages-modules-NeuralNetworks --bazel-mode-dev
Change-Id: I463646d2ae1fc4aa075a54c264e1c34571c3fd5c
This has been a workaround to avoid duplicate build rules for
multi-install apexes. The reason for duplicate errors was that apex_name
was used to install symbol files.
We don't need to rely on apex_name for symbol files, hence no need for
the workaround.
Note that symbol files for apexes are still available via the paths
based on "soong module name".
Bug: 267581665
Test: cuttlefish (which uses multi-apex) build & run
Change-Id: Icaca667e139ef2aca32339a75b54b12a00449c2d
VNDK APEX has been using "apex_name" property to set "name" field in
apex_manifest. But setting apex_name had a side-effect of emitting
duplicate rules for symbol files for VNDK APEX and its override_apex.
(Please see the removed comments for more details about error)
Not using "apex_name" fixes the duplicate errors.
There're a few more vendor apexes with the same issue, which I will
handle in follow-ups.
Bug: 267581665
Test: m (soong test) && boot
Change-Id: I00f617cef8af4d21880d4e8b9c0b9ab9322ba15c
Since fe4ff29b4d47fa8ae242fb7e084215d3afada4b5 (in system/apex), apexer
doesn't check keyname (--key vs --manifest.name). So, passing
--do_not_check_keyname doesn't make any sense.
This is a preparation to remove (or deprecate) "apex_name" property from
"apex" module.
Bug: 267581665
Test: m
Change-Id: I3737d1e36d305375a523a23a6bbeb2d23586e578
It makes sense to make min_sdk overridable because the override_apex is
a variant of the original apex, and this variant might be
created/introduced in the newer Android release compared with the base
apex. Despite the fact that the source of the variant apex is the same
as the original one, the variant may not support the older version, like
Go modules, trimmed modules. If we consider the variant as a "new"
module, then it makes sense to have a new min_sdk to make it only
available for the newer Android released version only.
Bug: 266945903
Test: Build locally, and check if the override works
Change-Id: I7f4c7169e15b7e57f44a4abafe012151dbc226a0
Disable VNDK apexes when their version is below the minimum
supported API level for the primary architecture.
Test: TestVndkApexCurrent
Change-Id: Ie72a5beb9da432660d3fec5c520402224326f961
This piggybacks onto the ApexInfo cquery handler, so we're issuing a
single bazel query call that reads two providers in the starlark expr.
Also rename requiredDeps to makeModulesToInstall to differentiate it from
APEX's required/provided libs in the apex manifest.
Test: unit test
Test: mkdiff
Fixes: 263123189
Change-Id: Ib7e43f1586f29864eee8627dba3631bfaff27afa
This reverts commit eb8cd3122a.
Reason for revert: mainline module jacoco jar do not contain JAVA_LIBRARIES needed for coverage, see: b/266679012
Change-Id: If756f7aaf40c6bcfed170c59c494bcc426620039
Apex builds are setting LOCAL_REQUIRED_MODULES to include external
dependencies into the system image. This make sense for device builds,
but just causes extra unnecessary builds for unbundled apex builds
that are not building device images. Skip the dependency during
unbundled builds.
Test: aosp-master-art mainline_modules_x86_64-userdebug builds
Change-Id: I363995b7309cc3c90b0a584d051050f105b5b915
variable.go:
add a new product variable TrimmedApex which is controlled by
environment variable PRODUCT_TRIMMED_APEX
config.go:
all config to check if trimmed build is turned on or not
apex.go:
add overridable apex module property:
use_DCLA: <DCLA apex module>
"use_DCLA" allows user to just use the lib list in DCLA to determine
which native shared libraries to trim. This property is no-op if
TrimApexEnabled is set to false.
builder.go
1, if it is to produce a trimmed variant, fix the default version code
by making the last variant digit to 2. this is only useful in local
build. in production build, there is apex manifest injection by
coastguard that determines the actual version code.
2, create a new apex rule called: "TrimmedApexRule", this rule invokes
an apexer wrapper called apexer_with_trim_preprocessing which will
properly trim the target libs.
BUG: b/259381334
TEST: adbd banchan build for both static and trimmed variant
Change-Id: I38a91cdef86a3c9185d00610ab3dd7acd07fd90a
The goal of this cl is to simplify the python rules,
mostly by removing the "decorator" pattern that they
currently use, and instead making separate module
types for libraries, binaries, and tests that inherit
from each other.
Bug: 259718110
Test: Verified ninja files are unchanged (they only change in the list of soong sources because I added/deleted files)
Change-Id: I1e836e2cc4782c7818f91db7df7895de3b8db7ca
Currently, Bazel sets __ANDROID_APEX_MIN_SDK_VERSION__ based on the cc library's min_sdk_version while Soong does based on apex's min_sdk_version. This causes a diff in the clang commands.
Digging further, we realized __ANDROID_APEX_MIN_SDK_VERSION__ isn't used anywhere so we might just delete it altogether instead of fixing it.
Bug: 265134933
Test: presubmit
Change-Id: I08e89dc8f6ef86669248ea5c47de6603e9d2ffbb
The image apex does not need all the files that make up a flattened
apex.
Bug: 254205429
Test: go tests
Change-Id: I7c574c5b30b55ab6d1b3981ee26d168c2f2427c4
Previously min_sdk_version did not handle soong config vars
Test: m bp2build and verify com.android.adbd
Change-Id: I48426a8e6e03b61234b77ce7d7ec07b1cab36b7b
The .capex is the compressed apex. Depending on various conditions, the
.capex should be installed, and thus the
apex.go#ProcessBazelQueryResponse should set the .capex as the main apex
output file.
Fixes: 264873511
Test: new unit test with install rules
Test: m com.android.adbd --bazel-mode-staging gives:
$ ls -lah
out/target/product/generic_arm64/system/apex/com.android.adbd.capex
-rw-r----- 1 jingwen primarygroup 3.6M Jan 10 14:44 out/target/product/generic_arm64/system/apex/com.android.adbd.capex
Change-Id: Ic3143cc6999f111428c6adb76760be8ddc4b6344
These instances could use WriteFileRule instead of
making their own shell code to write a file.
Test: Presubmits
Change-Id: I9c809b2164a68b4ce1c22fbbd0d7497240110b39
This was revealed by the new check added in
Ib3c6efd16bdf13b59f79be66d4f9dba49100f6cc.
This was okay since the old implementation relied on the fact that impl
variant comes before stub variants.
To make the new check happy and keep the intention explicit, stub
variants are not added to VNDK apex.
Bug: 191770320
Test: m (apex_test)
Change-Id: I455f2c9b6bc471c579379286c0198f36d325adf6
A single file generated by different modules is almost never correct;
make it an error.
The check is skipped for some error cases, which will be handled in follow-ups.
Bug: 191770320
Test: m (soong test)
Change-Id: Ib3c6efd16bdf13b59f79be66d4f9dba49100f6cc
Add AndroidMkEmitAssignList to emit a line to assign the items from
the given list of string arrays.
Test: treehugger
Change-Id: Id5acbef38ea4e91349bd2461f226db352d4b8123
Generally, when building system images from prebuilt APEXes the APEXes
should be released versions and the build ensures that is the case when
extracting .apex files from prebuilt .apks files. However, it is useful
to test building against prebuilt APEXes before finalization to flush
out any issues early on. Previously, the build would fail in that case.
This change adds support for using the SOONG_ALLOW_PRERELEASE_APEXES
environment variable to allow prerelease APEXes to be suppported.
Bug: 259958590
Test: m MODULE_BUILD_FROM_SOURCE=false droid
# Above fails because it could not find non-prelease APEXes in the
# prebuilt .apks files.
m SOONG_ALLOW_PRERELEASE_APEXES=true MODULE_BUILD_FROM_SOURCE=false droid
# Above works.
Change-Id: I20b8e5db0b9fd2a21434be25652195fae929bd3f
The apex and prebuilt_apex module types make the output APEX available
using ":<apex>" syntax, e.g. for use as a data file in a test. This
change adds compatible behavior to apex_set to ensure consistent
behavior when building against prebuilts.
Bug: 259958590
Test: m MODULE_BUILD_FROM_SOURCE=false nothing
Change-Id: Ic6f105976cfdd6af4920cc1e11ab0b4ff72e6a6f
A previous change to the deapexer script to check that blkid_path and
fsckerofs_path were set broke the scripts/unpack-prebuilt-apex.sh
scripts use of deapexer. This change fixes it by passing the paths
through to the script which then passes it on to deapexer.
Bug: 255963179
Bug: 240288941
Bug: 259958590
Test: m MODULE_BUILD_FROM_SOURCE=false droid
Change-Id: Ief7f36219b7fe1cf7104c08163e10d8e39f181d0
Vendor apex with "use_vndk_as_stable:true" sets the dependency in
"requireNativeLibs: ":vndk"". But this isn't enough because the APEX
won't work if VNDK version mismatches.
Now, when a vendor apex uses VNDK libs, "vndkVersion" is set as well in
apex_manifest so that apexd can abort the installation/staging when
device's VNDK version mimatches.
Bug: 222620439
Test: VendorApexHostTestCases
Test: built vendor apex how has "vndkVersion" field in its
apex_manifest
Change-Id: If03340c230efe854ab932cdf472c276f7646ad0c
Add an option to override defaultManifestVersion using environment
variable. The environment variable will be used to override the apex
version locally when developing a desert release feature in
mainline-prod branch. Overriding the apex version in mainline-prod
branch allows us to install apex built in mainline-prod branch to device
running code from git_master which has a higher apex version than
mainline-prod branch.
Bug: b/233608815
Bug: b/196860838
Test: presubmit
Test: go test -v ./build/soong/apex/
Change-Id: I8ac2aae2b6b29da7f09db6edb893e0a0f0691e51
The first line of each section will be
```
include $(CLEAR_VARS) # <module type>
...
```
Bug: 257037252
Test: treehugger
Change-Id: Iba586155b682fe4e1e5817d8397eda8f9d9c8789
The tags will then be read by an aspect applied from apex attributes to
validate that the deps are explicitly opt-in to the appropriate apex.
The semantics of this is the same as the Soong apex_available property.
This also adds the "base_apex_name" attr to override_apex conversion so
that apex_available checks continue to work on the non-Google and non-Go
apex names.
Test: presubmits
Fixes: 218841706
Change-Id: I19b3637da9aa47573b252d74f62b601fbdc3784d
This leads to an arm64 apex being used on a device that is mixed
x86_64 with nativebridge=arm64. A device like that doesn't appear
to work with arm64 binaries. For example, the boringssl-self-check
binary crashes on boot.
Bug: 260115309
Test: unit test
Test: boot emulator with this combination
Change-Id: Ic4a91974290a05b1799f755fcf52ef226d68f4c2
When defining a stubs library, allow specifying that the
implementation library does not need to be installed. This allows for
cases where the implementation is deployed in some non-standard way -
e.g. inside a Microdroid virtual machine.
Without this, we get build errors like: "TARGET module
com.android.compos requires non-existent TARGET module: libvm_payload".
Default behavior is unchanged. The change is protected by an allowlist
to limit usage to the immediate use case.
Bug: 243512108
Test: builds; soong tests pass
Test: Remove allowlist, see build failure
Change-Id: Iaae75f2e93b842f5944a7518cc95069d62c5a638
It's not read outside of the function / action generation logic, so keep
it inside the function.
Test: presubmits
Fixes: 190817312
Change-Id: I0626e6c86adc584abb2a251cbc7ffb1bec2c7f59
This file contains the java APIs parsed by the 'dexdeps' tool within the gen_java_usedby_apex.sh.
Bug: 239084755
Fixes: 239084755
Test: presubmits
Change-Id: Ia271783a6be3ea3a343481306cde1aaba2166e88
The image apex does not need all the files that make up a flattened
apex.
Bug: 254205429
Test: go tests
Change-Id: Iff05f2d01f11397bc9e9bf31868112575268969b
The libcronet prebuilt JNI library doesn't exist for riscv64 yet.
Add exclude_* properties to apex arch-specific properties so that
the libcronet dependency can be excluded from the apex that contains
when the arch is riscv64 until it exists for riscv64.
Test: TestApexWithArch
Change-Id: Ic395a077824f0f60b90178530fbfae8a96b3782f
There are no riscv64 apex prebuilts, which breaks the build when
provenance_metadata singelton tries to run a tool on all the input
apexes that don't exist. Hack it to provide the arm64 apex for now
if no riscv64 apex is specified, which will allow the build to pass.
Test: lunch aosp_riscv64-userdebug && m droid
Change-Id: Ic39936539803615ea4b7f817daf7b59ab7c40e7e
Soong supports string properties, but they are overloaded, and can mean
one of three things:
* path reference
* module reference
* string literal
Bazel has different types: label and string attributes. Thus there needs
to be a way to categorize them correctly in bp2build.
This CL introduces a new function to be used on properties like
apex_key.private_key / apex_key.public_key, as well as
android_app.certificate / apex.certificate.
It is important to disambiguate the prop betenn a string literal
attribute or file/rule target label attribute, so this functions does
just that. The new attributes are then further handled by their
respective macros (apex_key, android_binary, apex).
Bug: 253557437
Fixes: 253557437
Test: presubmits, new tests
Change-Id: Id8111cdd60d3aabcae7d17fe9da84d0ee3966023
This fixes a bug that only the primary output of an appSet is installed
when the appSet is part of a flattened APEX.
Bug: 247072627
Test: m GoogleExtServices on a device using flattened APEX.
system/apex/com.android.extservices.gms/priv-app/GoogleExtServices@TM
has these apks:
GoogleExtServices.apk GoogleExtServices-arm64_v8a.apk
GoogleExtServices-hdpi.apk GoogleExtServices-ldpi.apk
GoogleExtServices-mdpi.apk GoogleExtServices-tvdpi.apk
GoogleExtServices-xhdpi.apk GoogleExtServices-xxhdpi.apk
GoogleExtServices-xxxhdpi.apk
(before the fix, there only was GoogleExtServices.apk)
Change-Id: Icbc4dd002f856a3f751badec781ad132c423ac9b
In Soong, decodeMultilib, used to get multilib to determine the dep variations, return "first" if defaultMultilib is set to "common". apex sets defaultMultilib to "common" which means equivalent compileMultilib in bp2build for apex should be "first" (See new Soong unit tests for more context).
This CL fixes bp2build for apex to be more correct.
Bug: 251559512
Test: go tests
Change-Id: Id1cb4407980fc1fab91822c81326f37fb4adfa0a
This add support for src field in bp file.
Signed-off-by: Chen Guoyin <chenguoyin.cgy@linux.alibaba.com>
Signed-off-by: Mao Han <han_mao@linux.alibaba.com>
Change-Id: I9393696951db375d53436f9d5a9841b6a072cea3
