When generating vendor/etc/linker.config.pb, STUB_LIBRARIES was used.
Because it lists *all* libraries with stubs regardless of its
install location (vendor or system), using it for vendor
linker.config.pb doesn't make sense. Instead, use vendor-specific
STUB_LIBRARIES, which is named SOONG_STUB_VENDOR_LIBRARIES.
Bug: 313806237
Test: m (aosp_cf_x86_64_phone)
Test: $OUT/vendor/etc/linker.config.pb not listing libz.so
Change-Id: Icd0aaf92d9630c07f58c4739a9f0ac713516db43
Afdo creates more variants to build, and optimizing LTO is costly to
perform. Turn off these two optimizations for eng builds for a faster
build speed.
This reduces total C/C++ invocation time for aosp_arm64-eng by 6.28%.
Test: presubmit
Bug: 307753064
Change-Id: Ibac4269c66a64e896dba2074b607d71a2da37546
This is to handle libz special case. libz has stubs but not an LLNDK.
So, libz.vendor should be treated as non-stub-providing libraries and
Vendor APEX should bundle it if it's used by its contents.
libz will set no_stubs for vendor/product variants.
Bug: 313806237
Test: go tests ./apex/...
Change-Id: I10759d7073838909126f8bfe87654f11aa02fd32
The next CL will need a TestContext parameter in
ContentFromFileRuleForTests in order to retrieve the file rule contents
from the Config. Add it and update all the tests that use it in order
to simply review of the next CL.
Bug: 306029038
Test: go test ./...
Change-Id: Ia4b4c9854017ea3472fa2f8ba42cf7f72720496e
By default clang assumes that on riscv64 jumps are really expensive.
That's probably not true for the kind of SoCs we're dealing with, but
more importantly (see the bug) it causes clang to do loads from the
right hand side of a `&&` before the left hand side has been evaluated.
This found one latent bug in libcore, and it doesn't seem like the best
use of anyone's time to have to chase similar issues if they're going
to be similarly latent for riscv64 when we get clang's default fixed.
Bug: https://github.com/google/android-riscv64/issues/124
Test: treehugger
Change-Id: I640f1b43ea3d2452366ab86e97a9189fa9f5326c
To generate module-info.json in Soong for b/309006256 Soong needs to
know the test data paths. Moving test data installation into Soong will
also help later for test suite packaging.
Add ModuleContext.InstallTestData that installs the files listed in a
[]DataPath alongside the test. The files will also be passed to Make
to allow it to continue packaging them into the test suites for now.
Update the module types that are producing LOCAL_TEST_DATA entries
in their Android.mk files to go through InstallTestData instead.
Bug: 311428265
Test: atest --host toybox-gtests --test-timeout=120000
Change-Id: Ia8b964f86e584ea464667fd86a48d754d118bead
cc modules can use __ANDROID_VENDOR_API__ to read BOARD_API_LEVEL
that is the API level of the vendor surface.
Bug: 313822931
Test: check ninja commands to have -D__ANDROID_VENDOR_API__=<version>
Change-Id: Iceb5918cbfce0f24929d02d7e3caf1e9942b88e8
This can be used to add apex-specific shared_libs to a CC module. It
would be nice if we could have apex.shared_libs. But it would make code
much harder to follow because we need to record the shared_libs list as
"exclude list" for non-apex anyway.
Bug: 312510312
Test: m libhardware libhardware.vendor libhardware.vendor.com.google.cf.ir
# check DTNEEDED for libapexsupport
# only apex-variant should have the dependency
Change-Id: I56a3dc280127d8ba44337707444ea226a49ccf0f
For LTO compilation, we do not need MLGO flags in cflags, since codegen
happens during link only.
Fixes compiler warnings for unused command line argument.
Test: presubmit
Change-Id: I361e4292b10a3582fd5c69fa7b5678c654b44a0f
This Cl sets testModule field to True in cc.NewTest(). This will cover "cc_test", "cc_test_host" and "art_cc_test".
Change-Id: I4a8db86835b195db34fd9f86560e7bf9321fbd7d
Provider added for the following test modules in this change: art_cc_test, cc_benchmark, cc_fuzz, cc_test, cc_test_host, rust_test,and rust_test_host.
Bug: 296873595
Test: Manual test
Change-Id: I815680529bcbecacb3a2bdb8f3746053afdee48c
Fix various issues that prevent building or running tests with the
Go tools:
Remove testing/test_spec_proto/go.mod. There's already a go.mod at
the top level, and adding testing/test_spec_proto/go.mod breaks
compiling with the Go tools in IntelliJ or from the command line.
Fix go vet issues when trying to print []DataPath.
Fix aconfig_conversion_test.go and add it to Android.bp so it runs
in presubmit.
Test: go test ./...
Change-Id: I18af2f05631446fe507744a2951e4b02c7ab59be
We should update gnu++2a to gnu++2b too, but not in this CL.
Bug: 311052584
Test: treehugger
Test: m checkbuild
Change-Id: I3637444d74868b3a29b425fd749ecbbe1a45fbf7
This reverts commit febf8b9bdf.
We are still seeing flaky performance alerts from microbenchmarks even
with this change, and this has caused binary size / memory usage
increase. Let's add this flag back and observe whether performance
regression reports become more frequent.
Change-Id: I0fec1296b1f49b21b59d81add677f6a926ee88dc
To simulate 16k page size in x86-64 cuttefish, it is necessary
to compile the shared libraries and binaries with max-page-size greater
than 4096.
Bug: 309816695
Test: source build/envsetup.sh
lunch aosp_cf_x86_64_phone_pgagnostic-trunk-userdebug
m
Change-Id: I32670ef2c25fdcefec11bd07ba41cd0ea96c92bd
To simulate 16k page size in x86-64 cuttefish, it is necessary
to compile the shared libraries and binaries with max-page-size greater
than 4096.
Bug: 309816695
Test: source build/envsetup.sh
lunch aosp_cf_x86_64_phone_pgagnostic-trunk-userdebug
m
readelf -l out/target/product/vsoc_x86_64_only/system/bin/toybox
Change-Id: I38e4449a7826a2eca0c1a1cccc916af21f0b3a93
The MLGO model we embed in the Android Clang is trained on ARM64
ThinLTO artefacts. Applying the model across architectures does not
necessarily translate to performance improvement. Therefore only enable
the optimisation on ARM64 ThinLTO for now.
Test: presubmit
Bug: 293827654
Change-Id: Ie64a65c11191cf700a463637a0746c8470d3aa3c
In a trunk stable branch, the ABI-monitored libraries under different
release configurations share the source code. The exported symbols are
filtered by version scripts and API level. header-abi-linker needs to
take the API level as a parameter.
Test: make
Bug: 308192597
Change-Id: I4b55b7ec023451b3a4cd1da308a1de10842d51d2
Also add -Wno-everything as suggested by Pirama
Bug: b/308179696
Follow up: https://b.corp.google.com/issues/308179380
Change-Id: Iaa73f2b52363562b2391346d8b2e5dc73bab4f2d
Product modules install required vndk modules in /product/lib[64].
But still the product modules can link to the llndk from system
partition using the list in /system/etc/llndk.libraries.txt.
To install /system/etc/llndk.libraries.txt file, use
llndk_libraries_txt singleton without versioned name.
Until VNDK is fully removed, the vndk apex must install the versioned
llndk.libraries.<ver>.txt file to the vndk apex.
Use llndk_libraries_txt_for_apex singleton for this purpose.
Test: lunch cf_x86_64_phone-next-userdebug; m
Test: atest GtsVndkDependencyTestCases
Bug: 299867815
Bug: 302255959
Change-Id: I058616081fe31597a9b76b772b69927807436896
This warning is triggered even when initializers don't have any
side effects, so it very rarely surfaces actual bugs/mistakes.
99% of the time it complains about a purely cosmetic issue.
Disable it by default in commonGlobalCFlags, which still allows
individual modules to opt into it.
Bug: 145210666
Test: presubmit
Change-Id: I671b5c457ee3d92b9cc37f020ba54e01e1a04f59
More might be added later when we have an allowlist.
Bug: 305998969
Test: b build with the change in ag/25094394
Change-Id: I92df843ac81f4e33ce7506db61c989a380015975
* changes:
Revert "Decompress debug sections when creating debug info"
Revert "Compress debug info in clang builds with zstd"
Revert "Use zstd to compress debug info in cc libraries"
Revert "Use zstd to compress debug info in rust libraries"
* changes:
Use zstd to compress debug info in rust libraries
Use zstd to compress debug info in cc libraries
Compress debug info in clang builds with zstd
Decompress debug sections when creating debug info
We can save a significant amount of disk space by compressing
debug info in `lld` cc builds. This change configures a build
flag to instruct lld to compress debug info when building cc
libraries. Rather than adding a global config we add this to
each target individually to avoid linker failures on windows
cross build targets.
Test: m
Bug: 305277519
Change-Id: I4ab3d47fb0de7e31a39fb671cccde1acee3a2018
We can save a significant amount of disk space by compressing
debug info in `clang` builds. This change configures a build
flag to instruct clang to compress debug info.
Test: m
Bug: 305277519
Change-Id: Ieeaa655452d847e83ebc6bec0864e47c46f4c6fb
This reverts commit 4e28d0eda2.
Reason for revert: Reverting now that we should have sufficient data from Crystalball https://b.corp.google.com/issues/306023954
Change-Id: I9f92c203eda309956ec192432ba0063eb11f59e2
This reverts commit 07c362bf7f.
Reason for revert: DroidMonitor: Potential culprit for Bug 304839693 - verifying through ABTD before revert submission. This is part of the standard investigation process, and does not mean your CL will be reverted.”.
Change-Id: Iafb894510c48af2dd8b015d7264a031ff9348884
This enables sharing of platform constraints between platform and kernel builds.
Bug: 304594734
Test: treehugger
Change-Id: I48091eeaf2e529c3f180ba056da3935911868b14
FSAFDO (Flow Sensitive AutoFDO) is an improvement over AutoFDO. It
attempts to capture and represent the flow-sensitive profile in LLVM
optimization passes with the mechanism to expose and iteratively
reload/match the profile with better precision.
Enable this for all AFDO projects so that we can capture data from the
field, and then evaluate the performance gain.
Test: presubmit
Bug: 304282732
Change-Id: I0de6b4071e4f0a75500f3ffd5171a7be5117ec86
This reverts commit 80b54d2502.
Reason for revert: caused hello_world_test to not be converted, but it's a test that builds/passes
Bug: 300117121
Change-Id: I805cfb336b7f58a5a1e295cade16be4c471e2ed5
(This CL is a cleanup, and should be a no-op)
Currently we support three variations of cc api stubs.
1. publicapi stubs, i.e. ndk stubs (empty additional args to ndkstubgen)
2. module-libapi stubs that are also an ndk library (--systemapi --apex)
3. module-libapi stubs that are not an ndk library (--systemapi --apex
--no-ndk)
ndk_libs.bzl was used to differentiate between (2) and (3). This creates
an additional layer of indirection - users will need to modify this
external .bzl file if they would like to add a library to an ndk.
Replace this with an explicit atttibute in cc_stub_suite macro for better UX.
Test: go test ./bp2build
Test: b test //build/bazel/rules/cc:cc_stub_library_tests (added in
sibling CL)
Bug: 299501496
Change-Id: Idd3579e8013bae7a1740534f90d2767df5bac1a5
GLOBAL_THINLTO is now default, since it improves build speed, produces
more performant code and smaller binary. Remove the option to turn it
off since we no longer maintain that variation, and that simplifies the
logic (esp. for moving CFI builds to ThinLTO).
Also fixed a bug where ThinLTO is not propagated to its static deps on
non-default targets (32-bit or host).
Test: presubmit
Bug: 169004486
Change-Id: I31f41ba27c2b94a384d2ba5027049c307d6f4334
TARGET_VNDK_USE_CORE_VARIANT enables vendor to use some of the VNDK
libraries with core variant installed in /system/lib. However, this does
not make sense when VNDK is deprecated. This change is to ignore
TARGET_VNDK_USE_CORE_VARIANT when the VNDK is deprecated.
Bug: 303754049
Test: aosp_cf_x86_go_phone boot succeeded
Change-Id: Ie9fa75e0fa452e48924d51d64201690ffb271f33
The data attributes for cc library rules is dropped by their starlark
macros, so serves no real purpose. Furthermore, this unused dependency
proves problematic for allowlist v2, as there are many cases at HEAD
where the corner-case "requires" dependency would otherwise mark a module
as unconvertible.
Fixes: 303307456
Test: New unit test
Test: Manual verification to unblock allowlist v2 `bp2build.sh` runs in
AOSP
Change-Id: I6ca6104b958d2b428fc2ca5b3fa794106571acca
abort() seems friendlier, but in practice it puts a couple of irrelevant
libc.so frames on the top of the stack trace and clobbers all the
registers that might have helped remotely debug a sanitizer failure.
Let's switch to trapping directly in the sanitized code instead.
Bug: http://b/298741930
Test: treehugger
Change-Id: I6c971920c8506f97133749aaddd73b8cea6a12bc
This is now the default, and we will be removing non-GlobalThinLTO configuration.
Also changed handling for lto: never, it doesn't need to create an
lto-none variant.
Change-Id: I971baa920df867fb762923d925eed22215a89c27
Test: presubmit
Just knowing it failed is pretty useless. Tell people how to update
the ABI dumps.
Bug: http://b/156513478#comment43
Test: m ndk # after breaking the ABI
Change-Id: I7ed6a0acf2d1fbd942410678fc6aba2afbc5a2b6
This is the bulk of the "allowlist v2" feature. It will disable bp2build
generation for modules which have transitive dependencies without a
bazel build definition.
This CL includes this mutator, but doesn't register it as a bp2build
mutator (outside of a few unit tests). This allows us to easily iterate
on completion of this feature and ensure there are no launch blockers
before we finalize the change in AOSP.
Bug: 285631638
Test: Unit tests
Change-Id: Ifb0a079c409ca19b02cafa3fab2efa0d3deebc50
This module type is used by a single soong module - `common_libc` in
bionic/libc
Implementation details
- Convert this module type to ndk_headers rule. Bazel's ndk_headers rule
will have a boolean attribute `run_versioner` to determine if
verioner should be run on the headers
- Add this module type to the alwaysConvert bp2build list
- Add the converted target for `common_libc` to the deps of
`ndk_sysroot`. This ensures that unbundled apps link against the
versioned NDK headers of libc
Test: go test ./bp2build
Test: b build //bionic/libc:common_libc --config=android
Test: for f in $(find bazel-bin/bionic/libc/common_libc.versioned -type f); do cmp $f ${f/bazel-bin\/bionic\/libc\/common_libc.versioned/out\/soong\/ndk\/sysroot\/usr\/include}; done # no diff
Bug: 301169067
Change-Id: I55be202f0589db9bdc743c8be41c9c5afd74c352
PRODUCT_PRODUCT_VNDK_VERSION is set to 'current' by default. Now, we
can generate product variants without checking the
PRODUCT_PRODUCT_VNDK_VERSION build variable. Remove reading the
PRODUCT_PRODUCT_VNDK_VERSION variable from soong and generate product
variants by default.
Bug: 302255959
Test: m
Change-Id: I9a9b2076f4367c5ce9a393bbb206f8dee3884bd8
Test: go test soong tests
Test: verify no NeuralNetworksTest_random_graph sdk variants in
build.ninja when aosp_x86 is lunched
Bug: 302093869
Change-Id: I6429aa75518cca6974abcf7330236a6a07135ea1
Error message in case of not valid "nextAbiDump" included wrong
"prebuiltAbiDump.InvalidReason()". Refactor error message to include
"nextAbiDump.InvalidReason()" instead.
Change-Id: I2581734dd16649af2b282b3459f461658abd3d87
Signed-off-by: Aleksei Vetrov <vvvvvv@google.com>
This target will have a dependency edge to every bp2build equivalent of
Soong's ndk_headers. In b builds, sdk variants will compile against this
aggregated CcInfo providing target
A non monolithic alternative was discarded after conversations in
b/300504837#comment1-5
Contents of bp2build generated target: https://paste.googleplex.com/6643820291686400
Implementation details
- Since there is no equivalent Soong module for ndk_sysroot, hardcode
bp2build/build_conversion.go to collect all ndk_headers soong modules.
Add them to `deps` of a ndk_sysroot target
- Create `ndk_sysroot` in build/bazel/rules/cc/BUILD.bazel. This is
expected to be a temporary location. This will use the
cc_library_headers macro
- Update SetStubsForDynamicDeps so that sdk variant of rdeps depends on
//build/bazel/rules/cc:ndk_sysroot. This will provide a CcInfo during
compilation. Since ndk_sysroot is of type cc_library_headers, it will
not get packaged into the apk.
- Refactor `goBazelTarget` to a generic `bTarget` so that it is
representative of the expanded usage by ndk_sysroot
Test: b build //build/bazel/examples/android_app/java/com/app:app_with_sdk_variant_of_jni_deps --config=android (with aosp/2755284)
Bug: 300504837
Change-Id: Ifa427dd78115703ab251b0e1a0b71d3f19e91008
cfi and hwasan cannot be enabled at the same time. Skip building
vendor snasphots for the snapshot variants with both cfi and hwasan.
Bug: 277174974
Bug: 290779821
Test: HWASAN_INCLUDE_PATHS=frameworks/native/libs/gui m
Change-Id: Iaae913367a70e150d77a614de5f942aea98d1086
This should be no-op, as the underlying mutator has not changed yet.
Some other refactoring is required and done in this CL:
- Delete some old, dead ApiBp2build code
- Fix casting to TopDownMutator when it's not necessary
This change is required to prepare for allowlist v2 work, as only
BottomUp mutators can AddDependency.
Bug: 285631638
Test: m nothing
Test: presubmits
Change-Id: I5212a5f5634cc13056195783e6df37ff8eb000da
This is a followup fix for aosp/2753352 that added the name of the
`cc_stub_suite` as deps of `unbundled_app` config setting. However,
cc_stub_suite is really a macro that expands to <name>-<ver>.
Add the version explicitly to rdeps
Bug: 298085502
Test: b build //build/bazel/examples/android_app/java/com/app:app_with_sdk_variant_of_jni_deps --config=android (with aosp/2755284)
Change-Id: I8cf2804141bb14589e11b03a138b715538a3f1ee
This no-op refactoring facilitates some upcoming functional changes for
"bp2build allowlist v2". The work requires that the bp2build conversion
mutator be changed from a TopDown mutator to a BottomUp mutator.
Refactoring all bp2build-related methods so that they use Bp2buildMutatorContext
makes it easier to make this functional change without touching tens of
files and multiple projects.
Bug: 285631638
Test: m bp2build
Change-Id: I3d1ef3064146e959c6f0dc315350fc9764bf2bd2
Instead, we return an error. This allows us to access some product
variable information earlier when it will not be used as an attribute
without panicing
Test: m nothing
Change-Id: Id094b2b9e1364a8d174d99b3824fa149fb235b3e
Test: go tests
Test: CI
Test: patch in CL using prebuilts for mainline modules and build
Bug: 300640274
Change-Id: I73f70dac679e341670e1484ce57bc984f413d409
NDK has two ABI checks:
1. Check that prebuilt ABI exactly matches compiled binaries.
2. Check that ABI of level X compatible with level X+1.
The second check used compiled binaries for level X and prebuilts for
level X+1. But this approach may confuse people, because ABI change will
be shown "reversed". For example, adding new field to ABI monitored
binary of level X would be shown as "removed" when compared to level X+1
prebuilt.
This change uses prebuilts for both level X and level X+1. Given, that
prebuilts are checked to match compiled binaries, this should be enough.
Change-Id: If942e4319744bc5a2674cdd65f6a6f664fdfaa7e
Signed-off-by: Aleksei Vetrov <vvvvvv@google.com>
This reverts commit 49dcf79550.
The previous init crash was caused by the kernel defaulting to Sv57 and
the CFI shadow implementation in bionic assuming a 48-bit virtual
address space. With 5-level paging disabled in cuttlefish, we can again
enable CFI.
Bug: 293884796
Link: https://github.com/google/android-riscv64/issues/45
Test: cuttlefish boots
Change-Id: Ibcc02cf7ca6e396fd58674c648abf9f84daa260e
If a dependency has ndk stubs, then the sdk variant of the library
should link against the ndk stub variant and not the impl/apex_stubs variant
Unlike module-libapi, the depdendency does not go through an
@api_surfaces external repo indirection. This indirection was created to
support Multi-tree, and will likely be removed in the future
Test: Added a unit test
Bug: 298085502
Change-Id: Ie081e153fa586b6c22db0b8e42f91149fd8e5d9b
ndk_libary will be converted to a cc_stub_suite target. Its api_surface
attribute will be publicapi
The headers corresponding to this stub target will be added in a followup bug
(tracked in b/300504837)
Bug: 298085502
Test: Added a unit test
Change-Id: If9745083b18e0bcf5ecb89229a0f709b949d401c
This commit adds support for compiling rust rules inside the sbox
sandbox. To compile a rust module with sandboxing enabled, the entry
point to the crate must be specified via the `crate_root` property, and
all input sources and compile-time data must be specified via the `srcs`
and `compile_data` properties.
Bug: 286077158
Change-Id: I8c9dc5cf7578037a583b4be2e2f73cf20ffd4408
Ignore-AOSP-First: ag/24746588, in the same topic, is in an internal repo. This CL will be cherry-picked to AOSP afterward.
Test: go test
Bug: 277091218
Change-Id: I389d9535ea176991a1faa9beb46352b93363acd2
Merged-In: I389d9535ea176991a1faa9beb46352b93363acd2
cc_stub_suite now has an additional attribute correponding to the api
surface. This attribute will be used to pass additional args to the stub
generator.
Update bp2build to set the api_surface of the apex stubs. This ensures
that its stubs are generated with "--systemapi --apex"
Test: unit tests
Bug: 298085502
Change-Id: If4c479f85e6e485b5c795a565a0c559f1f013bf5
This module type will be converted to a cc_prebuilt_* bazel target. It
will provide
1. a prebuilt (.a/.so) file
2. headers (as -isystem)
Test: added a bp2build unit test
Bug: 298088835
Change-Id: Ib58cc7f6fde8f4ca34516f6f18a4c048a02a049a
Currently, we have ~4 of these modules which are all defined in
prebuilts/ndk. However, using ctx.ModuleDir() instead has the following
advantages
- makes bp2build simpler since we do not need to relativize this path
- prevents soong modules from reaching into another directory (as the
test setups were doing).
Test: m nothing
Change-Id: I780e2564cb37ebf4b800f0cd184789f3fc6f2fc8
Collect all NDK exported headers paths into a file that is used to
detect public types that should be ABI monitored.
Assume that we have the following code in exported header:
typedef struct Context Context;
typedef struct Output {
...
} Output;
void DoSomething(Context* ctx, Output* output);
If none of public headers exported to end-users contain definition of
"struct Context", then "struct Context" layout and members shouldn't be
monitored. However we use DWARF information from a real library, which
may have access to the definition of "string Context" from
implementation headers, and it will leak to ABI.
STG tool doesn't access source and header files, only DWARF information
from compiled library. And the DWARF contains file name where a type is
defined. So we need a rule to build a list of paths to public headers,
so STG can distinguish private types from public and do not monitor
private types that are not accessible to library users.
Bug: 156513478
Test: development/tools/ndk/update_ndk_abi.sh with enabled canDumpAbi
Change-Id: I9fa41e73450a41379638debb3dc56f421e0fb870
Signed-off-by: Aleksei Vetrov <vvvvvv@google.com>
For a dependency edge A --> B (stublib), Soong will link A's platform
variant against impl of B's platform variant if either of these are true
1. A and B have same apex_available
2. B has bootstrap: true
3. B is only available to platform
(3) was missing from bp2build. This CL adds that.
To implement this, we check the `apex_available` property of B.
Test: updated bp2build tests
Test: Built the internal module b/299191635 that was failing due to this
Bug: 299191635
Change-Id: Iafb173a3ab20d69b89f7949ce40c6f4096396f24
Soong currently adds -isystem
prebuilts/ndk/current/sources/cxx-stl/system/include to modules that
have ndk_system STL. This does not translate well to Bazel because of
its stricter sandboxing constraints.
In preparation for building sdk variants with Bazel, create an
indirection where sdk variants that use this STL depend on a
`ndk_system` header library module.
This should be a no-op in Soong, but the resultant ninja files are not
identical. (e.g diff https://diff.googleplex.com/#key=OOLtc1GFmDDF). The
-isystem now appears _before_ the local cflags. However, this should be
fine because we have a check in `CheckBadCompilerFlags` that bans use of
-I/-isystem in user-provided cflags.
Test: diff'd ninja file
Test: TH
Bug: 298258442
Change-Id: I8c4e2b66bb9ac25c44ceedd52298ba474a554a04
This enables Machine Learning Guided Optimizations (MLGO) for register
allocation. App launch performance on oriole-userdebug is improved by
0.53%, and binary size is very slightly reduced (e.g. 64 bytes for
libart.so).
https://dashboards.corp.google.com/_0e1b06dd_a79e_4433_b625_b62e5cac004b?p=BootImageProfileId:36265
Test: presubmit
Bug: 229056536
Change-Id: I0e5d40bb90a1555a1343f114ebffffca281f20b2
-Wmultichar is enabled by default, so we don't need to explicitly
enable it.
Bug: 298703966
Test: presubmit
Change-Id: Iae820145e6412365024a97ab3a34d760563785fa
Sandboxing the rust rules requires having explicit inputs for the rust
toolchain. This module type makes the rustc prebuilt and its
dependencies explicit as a module that can be depended on by all rust
rules.
Bug: 286077158
Test: rust sandboxing topic + go test
Change-Id: If2b80b32e329e6a6ee11ba824de868cf04714553
This warning usually indicates an incorrect size being computed,
which can cause severe bugs.
Bug: 296422292
Test: presubmit
Change-Id: I381dea70a6a408883ec772ccc006746ad66b835f
This adds support for correctly handling Rust sanitized snapshots,
ensuring they only have one variation.
The presence of multiple variations were causing build failures
when a rust_fuzz module for host was defined and a snapshot
build was requested.
This also sets -Z link-native-libraries=no on host modules
(in addition to device modules) to avoid emitting extra linkage
flags due to link attributes.
Bug: 282897366
Test: SOONG_SDK_SNAPSHOT_USE_SRCJAR=true m
Change-Id: Idf980c29145f11c530ad635a4eb5b01a1730ac24
Now built with MLGO. This reduces the total Clang invocation time for
building AOSP by 0.7%.
TOOLCHAIN_RUSAGE_OUTPUT=/tmp/rusage.txt m -j32
PGO + LTO + BOLT 83839.78s
MLGO + PGO + LTO + BOLT 82949.30s
Test: presubmit
Bug: 293616148
Change-Id: I5364c341520172e8e1acafb4aa3330922805ac0e
Sanitized glibc binaries get their runtime added by clang, not by
Soong, but unsanitized glibc binaries with sanitized dependencies
need their runtime added manually by Soong. Fix adding a static
runtime dependency to static glibc binaries.
Fixes: 297250415
Test: TestUbsan
Change-Id: I4913326604f9efebd8ecd4aad2f109fad8b6a80c
Adds the ability to centrally disable HWASan for multiple modules when
building with SANITIZE_TARGET=hwaddress. Soong version of the patchset.
HWASan takes precedence over CFI and several other sanitizers that it's
incompatible with[1], which can be problematic for modules that require
those sanitizers (e.g. those that depend on vendor prebuilts where only
sanitized variants are provided). This patch adds an easy way to disable
HWASan for such modules while still keeping it globally enabled.
Test: build with HWASAN_EXCLUDE_PATHS set and verify with readelf that
relevant modules have no references to __hwasan symbols
[1] bb31ca1168/core/config_sanitizers.mk (236)
Change-Id: I5824f71f2a400c64cde29e2c7afdd167d851d337
For the trunk stable next build, we have duplicated modules for the
vndk libraries txt files: one from the generated list, the other from
the prebuilt vndk snapshot files.
If the current vndk version provided by the vndk snapshots, use the
txt file from the prebuilt vndk snapshot.
Bug: 296777146
Test: lunch cf_x86_64_phone-next-userdebug && m nothing
Change-Id: I18bd7b7c77bd37c26c5e0b15cb020a59d50f4f2f
I want to make this warning an error in my project and the
current setup is preventing me from doing so.
This has to be -Wno-error=reorder-init-list and not
-Wno-reorder-init-list because the later is overridden by some
modules explicitly enabling warning sets. -Wno-* flags are
overridden by -W* flags coming later in the command line,
but -Wno-error=* flags are only overridden by later -Werror=
flags.
Bug: 145210666
Test: presubmit
Change-Id: Ia08f8a20da37cdc57167324b0cd19413d8786990
Source tree may include prebuilt vndk snapshot libs that are newer
than or equal to the PLATFORM_VNDK_VERSION.
Ignore those prebuilt vndk snapshot libs.
Bug: 296488609
Test: lunch cf_x86_64_phone-next-userdebug; m nothing
Change-Id: I3adaf3b7636f53884f08540959d2ec2fddfb6921
There are three types of tests, deviceless tests, device-driven tests
and host-driven device tests. But currently we don't have information
to get the type of a test and can't generate test targets on desired
types.
Test: b test //platform_testing/tests/example/native:hello_world_test
Test: b test //packages/modules/adb:adbd_test
Bug: 296312548
Change-Id: I3f022ef769636d508e055477623a4d1a6a1d9044
Without an explicit deduping in this helper function, a dependency may
appear multiple times in the generated BUILD files. I noticed this while
converting some internal google tests.
This currently does the deduping for the non-configurable root value. I
ran into some issues deduping the configurable values, and I will
address it in a followup CL.
Test: go test ./bp2build
Change-Id: Ie7b588a9c922d72e4fabfb1414a7cc3dc402ff86
Soong's proto.include_dirs becomes the -I path for aprotoc cpp code
generation. This does not translate well to Bazel because it runs this
action in a sandbox. Even if we construct an -I path, the .proto files
will not be there. This CL attempts to handle this kind of dependency
automatically via bp2build.
For this hypothetical example
```
foo.proto # contains `import bar.proto"
Android.bp # cc_library {srcs:"foo.proto", p.include_dirs:["subdir"]},
subdir/bar.proto
```
Implementation details for CcProtoGen of foo
- Glob the labels of .proto files for each includeDir, and create a
proto_library that encapsulates them.
- Since Bazel poses a contraint that proto_library target needs to be
in the same pacakge as the .proto file, the proto_library might be created
in a subdirectory with an import_prefix
- Add bar's proto_library as transitive deps of foo's cc_proto_library.
This will be added to -I during aprotoc. We cannot add them to `deps`,
otherwise bar's symbols will be statically linked into foo's .a
file.
Implementation details for clang compile of foo
At the end of CcProtoGen, we have converted foo.proto
to .cpp/.h files. To compile them to .a files, we need the .h files
generated from bar.proto. Soong specifies this at the
top-level cc_library soong module, so add those deps as the
implementation deps for clang compile.
(Will add support for java in a follow-up CL)
Test: go test ./bp2build
Test: built some internal modules that were previously blocked by this
Bug: 285140726
Change-Id: I7e1f9f0d1b1ba916a7ba8278f6cfb342b381d695
llndk.libraries.txt file is currently located within the VNDK APEX.
However, this file is still required even if VNDK APEX is deprecated.
This change removes llndk.libraries.txt from VNDK APEX, so it can be
installed within the system image.
Bug: 290160925
Test: aosp_cf build succeeded with llndk.libraries.txt in the system
image
Change-Id: I09a0a43babaa58ff16fc04ea71ab41ab68b54b70
Because enabling ThinLTO makes some constructor functions not called
in arm64 executables.
Bug: 295944813
Test: build
Change-Id: Ic0cab5fd80aa86d0d9ca1131564b747a445c6e46
The orderfile use flags are unused as cflags causing warnings which becomes errors in some order file builds
In addition, I fixed the test cases so it only looks at ldflags not cflags for orderfile use flags.
Test: mma build/soong
Output: #### build completed successfully (05:14 (mm:ss)) ####
Change-Id: I2a2d846d6688fd5256cf753267c000ff054d56f1
- ARM 32-bit targets will be built with 4096 ELF alignment.
- ARM 32-bit targets will only support 4K Page sizes.
- ARM 32-bit targets won't be page agnostic.
Test: source build/envsetup.sh
lunch aosp_cf_x86_64_phone
m
Bug: 295550555
Change-Id: I61cc45fa5f2ea18c494dfc9a0ca44cf9b694dfa3
If `srcs` contains a gensrcs/genrule module, the current bp2build module
will put it in the catch-all `srcs` attribute. This is reserved for .cpp
sources, so if the genrule produces a .proto/.aidl/... file, this will
fail.
This handles genrules that produce .proto files. To implement this, this
creates an additional partition that detects if the other module is a
genrule/gensrc that produces .proto files. If true, it will append it to
the proto partition.
This CL does not handle
- genrule that produce .c/.aidl/.yacc/.... files. They will continue to
be partitioned into the catch-all partition
- java modules
Test: unit tests
Test: TH
Bug: 293205700
Change-Id: Ib720fdf3a053ff5dd256e6faa632e3fa7776966d
Revert submission 2681060-dumpstate-isstrictrun
Reason for revert: This change has been successfully cherry-picked to udc-d1-dev, and can now be reverted in AOSP so that libdumpstateutil doesn't need a vendor variant.
Reverted changes: /q/submissionid:2681060-dumpstate-isstrictrun
Change-Id: Ia337b557c71c5a64d350f45199929440474fa278
Reduce the import instr limit from the LLVM default (100) to 40. This
helps reduce the binary size as well as improving performance slightly.
Size:
Default(100) 40 30
libhwui.so 11059040 11058912 11060872
libart.so 10697576 10697160 10696568
Performance:
40 vs. default: 0.37% improvement
http://go/art-benchmark?p=BootImageProfileId:36054
30 vs. default: 0.36% improvement
http://go/art-benchmark?p=BootImageProfileId:36058
Test: presubmit
Change-Id: Id800ff7818cde908daab784bac0a312c6a71272d
Now migration to STG is complete, hence remove the legacy extraction
mechanisms.
Bug: 156513478
Change-Id: I55b7626e2bfa78c9bb5157ade73d71a97597da72
Signed-off-by: Matthias Maennich <maennich@google.com>
Replace the abidw->stg pipeline by just stg for extraction directly from
ELF/DWARF. This is to migrate to STG entirely.
As a migration tooling, keep the existing method active via
`legacy_use_abidw` as a flag on ndk_library.
Bug: 156513478
Change-Id: Ideaa9908b31591f49f9a167cfa3f3d5c95d8b198
Signed-off-by: Matthias Maennich <maennich@google.com>
STG's native .stg format has been designed with version control in mind
as is the best format to represent STG's internal data representation.
Hence, migrate the ABI representation dump to .stg. That is done by
converting `abidw` extracted xml to stg at build time. STG has builtin
`abitidy` functionality, so `stg` can replace the prior `abitidy` step
to generate the final format.
Bug: 156513478
Change-Id: Ifcac8998ab899637a5d8c54f9e72d79d3e28ec52
Signed-off-by: Matthias Maennich <maennich@google.com>
In addition, I added a test file to check if flags are added and propagated correctly.
Test: mma build/soong
Output: #### build completed successfully (07:24 (mm:ss)) ####
For testing with an actual binary or shared library, steps are in this
README:
https://android.googlesource.com/toolchain/pgo-profiles/+/refs/heads/main/orderfiles/README.md
Change-Id: Idcf169156ef691bcacb8adc92828ef09450085f8
This change adds libdumpstateutil to VndkMustUseVendorVariantList. This
is done to avoid a VNDK re-spin on aosp/2679475 and aosp/2681060, which
add the dumpstate util function IsStrictRun and make it framework-only.
Test: Presubmit; verify that builds still work on internal (refer to
aosp/2681060 test).
Bug: 283326935
Change-Id: I3c60dd919a65182937d6e0b233635ff6ca9414d3
This reverts commit 73fe13c9b1.
Reason for revert: b/293884796, crash on boot of riscv64 aosp phone.
Change-Id: Ie90e0e99ddaa7bb5d496e2b1d0ab0302702cb2bc
If the target SDK of the module is a preview, the current logic passes
it through to llvm-rs-cc, which expects an int. Convert codenames to
their API level first, then extract the final or preview int.
Also simplify the logic to look for 'current' more generically via the
common ApiLevel library functions.
Test: Verified local build of riscv64, confirmed RS attempts to build
with API level 9000+ instead of 'VanillaIceCream'.
Change-Id: I1fa7577181bfd16de0ea4e77f7ab8fbd8fdb55e5
