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 prevents bp2buld conversion of modules which have transitive deps
that are not converted.
This does not change most allowlist semantics -- that change is still to
come. As a result, this effectively removes conversion of a few modules
which were previously converted under old semantics, however, these
modules are not currently used in any meaningful bazel builds, and will
be fixed at a later time.
Test: bp2build.sh
Test: m nothing
Test: manually spotchecked allowlisted modules in metrics to ensure the
diffs were minor
Test: manually verified bp2build performance regresses by about 0.4s
Change-Id: Id5c44fa5394917b28a3e707a81555b9e467d6621
This change makes sure that prebuilt module has correct module name
prefix("prebuilt_") inside method shouldConvertWithBp2build, to avoid
the bp2build conflict issue when a normal module and a prebuilt module
with the same name are both allowlisted.
Test: CI, added unit tests and the steps to reproduce in b/303376793
Bug: 303376793
Bug: 303868211
Change-Id: I4f1126182ec1f52ae6a08046a77204939230edef
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
In make, all product variables are assigned to an empty string before
including the product config makefiles. In starlark, we don't do that
just so the memory usage is smaller. This means that in make, using
?= to assign a product config variable has no effect. Replicate this
behavior in starlark by not emitting any code for ?= assignments of
product variables.
Fixes: 304324003
Test: go test
Change-Id: Id31531506ac9e372a1bea92ce9ae8e17ae0ca70c
(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
System_modules property provides the jars passed as bootclasspath when
compiling the stubs in the java_api_library where its creating
java_sdk_library's sdk_version is none, as the jars will not be provided
from the full_surface_stub_libs but compiled by itself in the child
change.
The jar provided by the system_modules will also be passed to metalava
to resolve hierarchy coming from outer dependencies.
Test: m --build-from-text-stub
Bug: 288624417
Change-Id: I8f3b89efa24bceb070d7a37fae3c7334dd7f0868
rust_bindgen modules which defined cpp_std with a .h file were not
correctly emitting the `-x c++` flag. This CL addresses that, and
ensures that if either cpp_std or c_std is set then the appropriate
behavior occurs no matter the file extension.
Bug: 304269101
Test: Soong tests
Change-Id: I71a8ae30ac0ed502d9d3fbf2f3039b0c56529d39
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
Make generates manifests for modules that are missing AndroidManifest.xml
files in order to paper over differences between aapt and aapt2. These
manifests don't have an <application> tag. which was tripping up an
unexercised error path in manifest_check.py.
Change manifest_check,py's extract_uses_libs_xml to consider manifests
with no application tag as having empty uses_library and
optional_uses_library lists. Also fix up a few nearby pylint warnings.
Fixes: 303554426
Test: m ConnectivityMonitorRobotests
Change-Id: I3815a1c86e24d7b3dea7f4ea26c34d3af182ded1
These were added with the android 14 release.
Bug: 290816499
Test: run genrule_sandbox_test.py with a local change to check all genrules in the tree
Change-Id: Icf3627c245638ab3f73b83e24ef04c916d7ab58b
This dir contains the following jni libs
- libnetworkstatsfactorytestjni
- libandroid_net_frameworktests_util_jni
Bug: 297405814
Test: b build $(bmod libnetworkstatsfactorytestjni) --config=android
Test: b build $(bmod libandroid_net_frameworktests_util_jni) --config=android
Change-Id: I3905396cab8df5ebbbf12eafcd9f96cde8305697
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
Bug: 290816499
Test: run genrule_sandbox_test.py with a local change to check all genrules in the tree
Change-Id: Id18b801c2306dd59b5b593b004b513b578ce3705
Revert submission 2761821-suppress-hiddenapi-check
Reason for revert: have some typo - break next build
Reverted changes: /q/submissionid:2761821-suppress-hiddenapi-check
Change-Id: I9fce1e1a9389d58928f1eec50c0eaf016f5f63ac
