afdo.addDep would never add a depencency on a profile for a host
module, but afdoDepsMutator would still propagate AfdoRDeps to
dependencies, afdoMutator would still create afdo variants, and
afdo.flags would still add the -funique-internal-linkage-names
flag. Exit early from all of these functions, since the afdo
variations will never be useful.
Test: afdo_test.go
Change-Id: I255fcbb86c99871916e571fbc74075d918b24745
Add more coverage to the afdo tests for the arm32 variant of a module
that only has an arm64 profile, for the -funique-internal-linkage-names
flag, and for the interaction between LTO linker flags and AFDO.
Test: afdo_test.go
Change-Id: I3ed4ce033c2431ea3e2fee536744b5e5b4cad296
This CL is the java_sdk_library_import equivalent of aosp/2928483.
With trunk stable, we will have multiple apex prebuilts in the tree.
Each apex prebuilt will have its own module sdk. This means that it is
possible to have mutliple versions of `framework-foo` in the tree. This
CL introduces a mechanism to select a specific versioned
java_sdk_library prebuilt.
Implementation details
- Add a `source_module_name` property to java_sdk_library_import. This
will identify the source equivalent of the jsl in packages/modules.
This used to be implicit - i.e. the name without the prebuilt_ prefix.
With multiple prebuilts, this has to become explicit.
- Set appropriate `source_module_name`(s) in the dynamically created
child modules. e.g. the source_module_name on sdklib.v1.stubs and
sdklib.v2.stubs will both be sdklib.stubs, assuming
`source_module_name` on the top-level jsl_import is sdklib
- Add a private Created_by_java_sdk_library_name property to java_import
and prebuilt_stubs_sources modules. This will be used to idenfity the
top level java_sdk_library_import that was used to create these child
modules. This is necessary because java_sdk_library_imoprt is a macro
that creates 1:many modules. However, to avoid toil, only the
top-level java_sdk_library_import will be listed in
`apex_contributions`. This new property will be used for
source/prebuilt selection in android/prebuuilt.go
- Rename BaseModuleName in commonSdkLibraryAndImportModule to
RootLibraryName. This is necesssary because the former is now reserved
to identify the source equivalent of a prebuilt module (maybe we
should rename it?)
Bug: 322175508
Test: Added a unit test
Change-Id: If6aa6b0e939a07f8263874941413908383c81a78
I want to be able to use certain characters to separate a namespace
from a variable, but currently it's possible for soong config namespaces
and variables to have any character in them.
Restrict them to just identifiers, which should cover all of our
existing usages.
Bug: 323382414
Test: Presubmits
Change-Id: I6d55dd3378ac4d257efde450189b81a4aa3932c3
this.teams_for_mods uses a dictionary and `range` yields a
non-deterministic ordering. This would cause all_teams.pb to be
non-determinisitic. Since this file is created during Soong analysis, it
would cause the .ninja file to also be non-deterministic.
Use SortedKeys to do the iteration instead.
Test: go build ./android
Test: m nothing a couple of times and checked that the .ninja files are
identical
Change-Id: Ife2d2520d118ef25639f86390912d98b5f057655
The aconfig files are collected by visiting the direct deps of the
module. VisitDirectDeps does not guarantee ordered traversal. To prevent
non-determinism in the generated ninja file, sort the inputs.
Test: go build ./android
Change-Id: Ic67fc0859bf18de62b6297ed502d1d495cc3a780
Similar with aosp/2897612, rename LOCAL_USE_VNDK into
LOCAL_IN_VENDOR or LOCAL_IN_PRODUCT to make variable useful from
VNDK deprecation.
Bug: 316829758
Test: AOSP CF build succeeded
Change-Id: If7bead24add014ed9e0e9e1d265b0199b911be89
In order to use soongdbg, you must run analysis with GENERATE_SOONG_DEBUG=true set
in the environment.
Test: GENERATE_SOONG_DEBUG=true m nothing ; soongdbg ...
Change-Id: If43676fe2784f05cd87c0ecb4a46ab676b91023f
android/prebuilt.go#isSelected has a special-case inside it to ignore
apex_contributions contents for the top-level java_sdk_library hook.
This was necessary because even though we might want source stubs in
next builds, we still needed the top-level prebuilt hook to be active to
emit the dexpreopt rules to .mk.
This worked fine for rdeps that create a dependency edge on the child
stub modules of java_sdk_library. Notable examples include the full
android api stubs created by f/b/api.go. In postdeps mutator, these
expanded deps get rewritten to source/prebuilt if necesssry.
The exception to this are workflows which depend on the top-level hook
directly via `libs`. We resolve these rdeps to an appropriate provider
during GenerateAndroidBuildActions stage. This meant that rdeps were
getting prebuilt stubs of these even in next builds.
Bug: 323454855
Test: Added a unit test
Test: lunch cf_x86_64_only_phone-next-userdebug
Test: aninja -t query
out/soong/.intermediates/packages/modules/Permission/SafetyCenter/Config/safety-center-config/android_common/javac/safety-center-config.jar
| grep prebilts/module_sdk # empty now
Change-Id: Id91333d88055519f3c58ab40466f9628085f5180
The order function passed to ExtendMatchingProperites no longer takes
a property, dstValue or srcValue parameter.
Test: builds
Change-Id: I19fc554f705fbf42100f0a38df87e2d1fb2454ec
Mutator contexts are created for every module, and in the case of
transition mutator contexts for every dependency of every module.
Unlike the Blueprint mutator contexts that they wrap, the Soong
mutator contexts can be relatively large. Add global pools for all
of them that can avoid the repeated allocations.
Test: SOONG_PROFILE_MEM=/tmp/mem.pprof m nothing
Change-Id: I64a5f3c91292cff6352300f99c11ac50c713f96d
Every InstallPath will have String() called on it eventually, often
more than once if it is in a slice that is sorted. Precompute the
full path so it can be returned from InstallPath.String() without
recomputing every time.
Test: paths_test.go
Change-Id: I1ed0a3801806854356865c0a5fc35d5cf6d349fe
LOCAL_DROIDDOC_STUBS_SRCJAR is used to copy entries to OUT_DOCS in
build/make/core/soong_droiddoc_prebuilt.mk, which defines the copy rules
to generate the docs. When generating the documentation, the exportable
stubs should be set as dependencies, so that disabled flagged apis are
not exposed. Thus, this change modifies the entries passed to
LOCAL_DROIDDOC_STUBS_SRCJAR from the everything stubs to the exportable
stubs.
Likewise, all other LOCAL_DROIDDOC_* entries are copied to dist during
sdk builds. Therefore, this change replaces the everything artifacts to
the exportable artifacts.
This change also config-guards the exportable stubs generation, so that
the exportable stubs are not generated when the droidstubs module does
not generate the stubs.
Test: m docs dist && inspect ninja file to verify that everything stubs are not exported
Bug: 323261972
Change-Id: Ic04fb47ad311aefafc2f1a7b8bb15525050ecd2e
How it works:
1. build/make/core/Makefile generates a txt file with the kernel
version, which is taken from an explicit BOARD_KERNEL_VERSION value,
or extracted from the kernel image on the source tree, or extracted
from the kernel image extracted from the prebuilt boot.img.
The file is saved at
$ANDROID_PRODUCT_OUT/obj/PACKAGING/check_vintf_all_intermediates/kernel_version.txt.
2. If PRODUCT_ENABLE_UFFD_GC is "default", meaning the GC type needs to
be determined by the kernel version, build/make/core/Makefile copies
kernel_version.txt to
out/soong/dexpreopt/kernel_version_for_uffd_gc.txt.
3. build/soong/dexpreopt/config.go writes the the UFFD GC flag to
out/soong/dexpreopt/uffd_gc_flag.txt. The flag is determined by an
explicit PRODUCT_ENABLE_UFFD_GC value or by contruct_uffd_gc_flag.py,
which reads kernel_version_for_uffd_gc.txt and determines the flag
accordingly.
4. dex2oat takes the UFFD GC flag from uffd_gc_flag.txt.
5. post_process_props.py mangles ro.dalvik.vm.enable_uffd_gc based on
the same logic.
Bug: 321751629
Bug: 319554951
Bug: 318763448
Bug: 319648491
Test: m --no-skip-soong-tests nothing
Test: atest uffd_gc_utils_test
Test: Build with `OVERRIDE_ENABLE_UFFD_GC=default m` for device with no
UFFD support -
1. Check the existence of `-Xgc:CMC` in
out/soong/dexpreopt_arm64/dex_bootjars/android/system/framework/arm64/boot.invocation
(dex2oat invocation for a boot image)
2. Check the existence of `-Xgc:CMC` in
out/soong/.intermediates/packages/apps/Settings/Settings/android_common/dexpreopt/Settings/oat/arm64/package.invocation
(dex2oat invocation for an app defined in .bp)
3. Check the existence of `-Xgc:CMC` in
$ANDROID_PRODUCT_OUT/obj/APPS/Dialer_intermediates/oat/arm64/package.invocation
(dex2oat invocation for an app defined in .mk)
4. Check the value of ro.dalvik.vm.enable_uffd_gc in
$ANDROID_PRODUCT_OUT/product/etc/build.prop
Test: Build with `OVERRIDE_ENABLE_UFFD_GC=default m` for device with
UFFD support, and do the steps above.
Test: Build with `OVERRIDE_ENABLE_UFFD_GC=true m`, and do the steps
above.
Test: Build with `OVERRIDE_ENABLE_UFFD_GC=false m`, and do the steps
above.
Change-Id: I035ad32233e49e2a30ce11f6c7c318a648511ef8
Api levels annotations flags should respect the stubs type, but the
everything api versions xml file was being passed to the metalava
invocation for both everything and exportable metalava invocation.
This change fixes such issue by setting api versions xml file (passed to
metalava as --apply-api-levels argument) to respect the stubs type.
Test: m nothing --no-skip-soong-tests
Bug: 319162970
Change-Id: If7155b3701b32fb4cd81dc918648b449de090201
Timestamps in pyc files and zip archives were causing cache misses in RBE.
Bug: b/322788229
Test: Ran a few local builds and verified improved cache hit rate for metalava actions.
Change-Id: I9375694d26a107f5c0c43bbd685bd8e86349d187
