Original aosp/2374663 was reverted, try to submit it again with some fixes in this.
Test: m sbom
Test: m dist
Test: on aosp, lunch aosp_bluejay-userdebug && m dist
Bug: 266726655
Change-Id: Icf305770473f0c448a1ad721cbe7addf737115e4
riscv64 can't build the device version of bcc and ld.mc due to a
dependency on an old version of LLVM, but they are listed in
base_system.mk which can't add them conditionally based on the target
architecture. Add a hack to base_system.mk to silence the warning
that the host-only modules are listed in PRODUCT_PACKAGES. This
can be removed once renderscript is removed from the platform.
Test: lunch aosp_riscv64-userdebug && m
Change-Id: I8efd769a34b110dfb838f7016e029369cb9a0f01
Currently we only have vbmeta_system and vbmeta_vendor. Add
BOARD_AVB_VBMETA_CUSTOM_PARTITIONS , which allows OEMs to define
their own vbmeta_*.img partitions. example: aosp/2381823
Bug: 263568805
Change-Id: Id671e2c3aee9ada90256381cce432927df03169b
This is determined by:
- a product config flag
- the vendor API level
It is then passed to the device as a system property
"ro.dalvik.vm.enable_uffd_gc".
This change is a no-op change. It doesn't enable userfaultfd GC by
default. OVERRIDE_ENABLE_UFFD_GC=default can be passed to the build
system to enable userfaultfd GC for testing purposes.
Bug: 242553398
Test: -
1. lunch aosp_redfin-userdebug
2. OVERRIDE_ENABLE_UFFD_GC=default build/soong/soong_ui.bash --dumpvars-mode --vars=ENABLE_UFFD_GC
3. See "false" in the output
Test: -
1. lunch aosp_oriole-userdebug
2. OVERRIDE_ENABLE_UFFD_GC=default build/soong/soong_ui.bash --dumpvars-mode --vars=ENABLE_UFFD_GC
3. See "true" in the output
Test: -
1. lunch aosp_redfin-userdebug
2. OVERRIDE_ENABLE_UFFD_GC=true build/soong/soong_ui.bash --dumpvars-mode --vars=ENABLE_UFFD_GC
3. See "true" in the output
Test: -
1. lunch aosp_oriole-userdebug
2. OVERRIDE_ENABLE_UFFD_GC=false build/soong/soong_ui.bash --dumpvars-mode --vars=ENABLE_UFFD_GC
3. See "false" in the output
Change-Id: Ifd6e6cddb502315912ff949619a5b526ae0d73ff
This variable is always false or empty now, so clean up remaining
references of the variable.
Remove "buildvariant=" kernel commandline as it was used by
system-as-root kernels to determine the system build variant. It is
unsupported now.
Bug: 241346584
Test: Presubmit
Change-Id: I0f97b6e0bf8b03429f65366119ff78f9c7983516
This should only be set to '1' by flashing debug boot image,
so default to '0' in system/build.prop.
Bug: 223517900
Test: Boot with and without debug boot image and check value of property
Change-Id: I0e0df7d86a31fd771248433513483e8ad600b5ed
The SDK no longer includes system images, so no longer needs to depend
on targets that are only relevant for the system images.
Remove all the explicit dependencies that end up depending on
"everything" but are not included in the sdk zip.
Two positive side-effects:
- fewer unused things get built (now ~32k targets, down from ~108k)
- apexes no longer get built, eliminating the risk of mismatch between
module sdk prebuilts and source-built AOSP apexes.
There are some potential risks with this CL:
- because we no longer forcefully rebuild everything, if there are bugs
in dependencies of exising files in the SDK, incremental builds may
not rebuild them.
- some files that get built into the SDK may be being built as a
side-effect of other files included in the SDK, so removing one file
may cause another to not get built. The solution to that is to add
the file not getting built to ALL_SDK_FILES.
Bug: 205008975
Test: m sdk sdk_repo dist && diff sdk.zip before/after (identical)
Change-Id: I97284864f66aa88556fbe16864f45b04be97634e
cc_test that can run on hostside are default to be part of
host-unit-tests suite. The suite is set up similar to device-tests and
general-tests, and it's not a compatibility suite. So following variable
will always be empty:
$(COMPATIBILITY_TESTCASES_OUT_host-unit-tests.
It should use HOST_OUT_TESTCASES as testcases dir.
The same logic should be applied to art-host-tests.
Bug: None
Test: build
Change-Id: Ieddacd5157da73c6df473c103be109d9c3a55281
The long-form variables (PRODUCTS.<makefile>.<variable>)
are used to get information about multiple products.
However, they've never really worked correctly, and so
importing multiple products is deprecated behavior.
Remove as many usages of the long-form variables and
multi-product imports as possible.
Bug: 228518445
Test: Manually
Change-Id: I0b67f16360ff8bdcdb39638de739440472bccf76
This reverts commit 59ecb9e016.
Reason for revert: symbols_map is fixed to ignore non-elf files
Bug: 218888599
Change-Id: I88654b2dba3dfb9ab2809b0555544277279f345d
Artifact path requirements requires setting some new variables,
and also dumping the state of the product config variables
after evaluating partial products that set artifact path requirements.
Bug: 188079133
Test: Removed the code that disabled artifacts path requirements from the rbc dashboard script,
and verified that it was still green for aosp_arm64
Change-Id: I228e35285d788f4c83aa695c0f28b7c7db02544c
Run the symbols_map tool on each unstripped binary and proguard
dictionary while copying it to produce a textproto that maps the
hash to the symbol/dictionary file location. Combine the textprotos
into a mapping file when producing the zipped build artifacts.
Bug: 218888599
Test: m dist
Change-Id: I78997e2f3b631ed74213903cc8e72415b2d7eee0
If0e4b958b3dfaa02771a5da70f970379635f904e made `droidcore` stop
depending on `files` so that installed files from unbuilt images
would not be built. That also disabled the dependency on installed
files that were not in any image, but the vendor notice file had
overly broad dependencies that caused them to still be built.
Icdb11d3c72b180cd02231e8dc98ae500d6566a98 made the vendor notice
file dependencies less broad, exposing the missing `droidcore`
dependencies.
Track the list of installed device files that are not included in any
image, built or unbuilt, and add them as a dependency for `droidcore`.
Bug: 225187583
Test: m checkbuild
Test: check dependency path to encryptionkey.img
Change-Id: I084158ffb52b353e9a3eca9ad06a68c80d3bd371
Add vendor_kernel_boot image for vendors whose bootloader support
extra first stage booting kernel modules ramdisks. This benefit
kernel repo to build kernel-artifacts only image without Andorid
artifacts dependency.
Bug: 214409109
Signed-off-by: Lucas Wei <lucaswei@google.com>
Change-Id: If07218b86a7751b3d452a172610af960f5f9ec74
The goal of this change is to cause ro.product.first_api_level to be
set in /system/build.prop in targets with PRODUCT_FULL_TREBLE=false
(which may be the case for targets without the usual set of
hardware interfaces, such as the mini targets being added in
https://r.android.com/1966051).
Setting ro.product.first_api_level is necessary for devices with newer
kernels to boot because it controls which metadata encryption format
is used, and an old format that is incompatible with new kernels is
used by default (see commit 4073c0b5241b635cb59e9e8a75ddcc783b5a99ce in
system/vold). There also appear to be e.g. bionic tests that check this
property. So it seems reasonable to set it even if PRODUCT_FULL_TREBLE
is false.
To begin with, stop adding it to ADDITIONAL_SYSTEM_PROPERTIES. There's
basically no point in adding it there because it will be filtered
out by the property blocklist for the system partition, whether
PRODUCT_FULL_TREBLE is true or false. This lets us remove the
blocklist entry, which causes ro.product.first_api_level to
appear in /system/build.prop via ADDITIONAL_VENDOR_PROPERTIES on
PRODUCT_FULL_TREBLE=false targets.
Bug: 217455793
Change-Id: I8c07baf9d308ea734ffed5d9b1776b60cdbf3435
Some product config makefiles are being refactored for
the conversion to Starlark. In order to test that the
refactor doesn't cause any meaningful change, we test
that the ninja files are the same before/after the refactor.
Currently, they have differences in the order of AB_OTA_PARTIONS,
which shouldn't matter. In order to increase our confidence
and make the ninja files more stable, sort AB_OTA_PARTITIONS
before it's used.
Bug: 220169197
Test: Presubmits
Change-Id: I6a9f9f2990406d810577cf8930416925f2987311
There are 2 choices to build system_dlkm.img for
the system_dlkm partition for Android T launch
devices and must choose one.
1. Use kernel prebuilt system_dlkm.img
- BOARD_PREBUILT_SYSTEM_DLKM_IMAGE to point image
2. Build from kernel prebuilt system_dlkm_staging
- PRODUCT_BUILD_SYSTEM_DLKM_IMAGE
Both requires: BOARD_SYSTEM_DLKM_PARTITION_SIZE and
must be 64MB or higher in size (enforced via vts).
Bug: 200082547
Test: TH
Test: atest vts_system_dlkm_partition_test
Signed-off-by: Ramji Jiyani <ramjiyani@google.com>
Change-Id: I83435123bd8aa3d04ab8a8b650a95fbab0bc49f2
This was assigning the INSTALLED_INIT_BOOT_IMAGE_TARGET to bootimage
instead of initbootimage.
Test: m
Bug: 203698939
Change-Id: I5fc6e2296dac8cc35c4d46601fea6ab0ee18fd18
Keep the property for now to keep the ability to toggle on/off the
library we want. If we want to remove this functionality,
we need to cleanup all references to persist.sys.dalvik.vm.lib.2.
Test: build and boot with debug apex
Bug: 214049018
Change-Id: I66c7eb510f5a8b7974391ceafd309fba59a524a1
Right now if PRODUCT_EXPORT_BOOT_IMAGE_TO_DIST is given,
$ m dist bootimage
would copy boot*.img to the dist dir.
After this change, both
$ m dist # droidcore
$ m dist bootimage
would copy boot*.img to the dist dir, the difference is that the latter
builds & copies *only* the boot images.
Bug: 212486689
Test: m dist bootimage
Test: m dist => boot*.img are copied to dist dir
Change-Id: I1c714ce5d75eba93e32078b1ac291a181b05405a
This ramdisk used to be in boot.img, and is now placed into this new
init_boot.img instead.
This new image is used for a new init parition to seperate Android
platform artifacts from the kernel artifacts in boot.img.
Test: boot Cuttlefish
Bug: 203698939
Change-Id: Iaaf82486259979ab728730ce72a4e847ae005c18
The main issue with board configuration up till this
cl was that it didn't have access to the product configuration
variables. Pass those in by dumping the make variables to a
temporary file, which is then converted to RBC, loaded,
and passed to the starlark board config..
Bug: 201700692
Test: build/bazel/ci/rbc_product_config.sh -pb sdk_phone_x86_64-userdebug
Change-Id: I9a4946b970ca43c5b5f53a6c507ad2c1a2eca61e
Support moving some install rules into Soong by allowing Soong to
provide a Makefile containing install rules, and by skipping parts
of base_rules.mk when Soong already created an install rule.
The install rules are exported as a Makefile instead of directly
in Soong's build.ninja file so that base_rules.mk can add dependencies
from the `required` property that are hard to resolve in Soong.
Bug: 204136549
Test: m checkbuild
Change-Id: Ife2964d76469b385e90c0c5f6adf21bd2b5e8bf6
Device builds are no longer supported on Mac, but we do support building
various host tools, including the SDK build-tools and platform-tools
packages. All of these are defined in Soong now, so we stop reading all
the Android.mk files, and defining and depending on the device output
targets.
Bug: 187222815
Test: `m`, `m dist`, etc on Mac
Change-Id: Ibb19756d44f2cbc6948381043115718ffc3f4b7d
Since those were the only components of the Windows SDK, this removes a
lot of special casing around that as well.
Bug: 187222815
Change-Id: I2690aadcf78e9ccad43314c8475893e81a2e2735
This is to support a bootimage only target in AB.
Bug: 198363484
Test: lunch aosp_arm64-userdebug; make bootimage dist
Change-Id: I9a32c365f635ec4693675a1969d7a1e684c8f55a
The Mac SDK build is being reduced to just the build-tools and
platform-tools packages. I'm still working on refactoring those builds
not to require the monolithic SDK to be built first, but we can at least
stop distributing it (and other related files) now.
Most of these checks can likely go away once the build refactor is
complete, since we'll be able to move away from the `sdk` target to
something more specific. Right now there's a lot of logic around sdk
being in MAKECMDGOALS that I don't want to unravel with the current
build logic.
Bug: 187222815
Change-Id: I1f711ce063170e41078ad4a125bcd77340a1746c
soong_docs will be run directly from soong_ui so it won't have a
corresponding Make target.
Test: Presubmits.
Change-Id: Id7abccdf2ea3830edf6043c10b617bddf4f9c86d
Copy generated API list during building cc_library to cc_apis/
Test: build/soong/soong_ui.bash --make-mode SKIP_ABI_CHECKS=true TARGET_PRODUCT=aosp_cf_x86_phone TARGET_BUILD_VARIANT=userdebug droid dist DIST_DIR=/usr/local/google/home/sophiez/my_dist_dir EMMA_INSTRUMENT=true EMMA_INSTRUMENT_FRAMEWORK=true CLANG_COVERAGE=true SKIP_BOOT_JARS_CHECK=true
Change-Id: I05e6b7b9c9bbbab72718fd926ddda468bde05565
If RBC_PRODUCT_CONFIG variable is set, obtain product configuration
variables by converting product configuration makefiles to Starlark
files and then executing them.
Also, introduce RBC_NO_PRODUCT_GRAPH variable to suppress product graph
generation. We cannot generate product graph with Starlark, so this
option allows to verify that the rest of the contents of the generated
Ninja files remains the same when Starlark-based converter is used.
This allows to perform the regression testing, i.e. running
`RBC_NO_PRODUCT_GRAPH=t DISABLE_ARTIFACT_PATH_REQUIREMENTS=t m nothing`
and
`RBC_PRODUCT_CONFIG=t m nothing`
should generate identical *.ninja files.
Bug: 181797530
Test: Manual
Change-Id: Ic6173a9640f32766b71c02a2b1833ce7a278e4cc
boot-test-harness.img is used to allow adb root on user build
images. It also sets properties: ro.audio.silent=1 & ro.test_harness=1.
GKI 2.0 devices will use BOARD_PREBUILT_BOOTIMAGE so
boot-test-harness.img will not be generated.
Therefore, we have to introduce the vendor_boot-test-harness.img
as an alternative for boot-test-harness.img.
In the future, we'll simplify the flow as:
+ If a device has a /vendor_boot partition, builds
vendor_boot-(test-harness|debug).img.
+ Otherwise, builds boot-(test-harness|debug).img.
boot-(test-harness|debug).img needs to be kept for some clients
to gracefully transit to using vendor_boot-(test-harness|debug).img.
Bug: 194654549
Test: make then `unpack_bootimg --boot_img $OUT/vendor_boot-test-harness.img`
Test: Check the ramdisk content in ./out/vendor_ramdisk
Change-Id: If3a1393b4ff3e69bb9b62f3b843b7858437d47bf
When rule action contains something like
cp $(FOO) ...
and FOO is set with
FOO := a \
b \
the generated Ninja file constains extra spaces, making it
difficult to compare it to the same file generated by the
Starlark-based product configuration.
Bug: 181797530
Test: manual
Change-Id: I278bd8edf0f017a31c5b5115b2a38f4f663c55fc
Bug: 188179858
Test: compare dexpreopt_config.zip files from
1. TARGET_BUILD_UNBUNDLED_IMAGE=true m dexpreopt_config_zip
2. m dexpreopt_config_zip
(note that m clean should run between steps)
Change-Id: I92623b59e4fcac397c7fed3d7ab1070fe5281027
But only for a TARGET_BUILD_UNBUNDLED_IMAGE build. Since
the unbundled image build attempts to remove as many
framework sources as possible, we'd like to remove the
sources and dependencies of some blueprint_tools as
well.
Test: m nothing dist
Test: Verify graph is unchanged
Test: m nothing dist TARGET_BUILD_UNBUNDLED_IMAGE=true
Test: Verify that graph no longer includes blueprint_tools
Bug: 192629969
Change-Id: Ibfb52b841379e20855d0a935a13ad047af2fe58d
The droidcore-unbundled target replaces the droidcore
target when TARGET_BUILD_UNBUNDLED_IMAGE=true. This
target is similar to droidcore in terms of dependencies,
but it has a smaller set of dist files. The intention is
to use TARGET_BUILD_UNBUNDLED_IMAGE=true with a
VSDK-using vendor build to supply Java dependencies from
prebuilts to avoid building (or needing) Java framework
sources.
Test: build with and without TARGET_BUILD_UNBUNDLED_IMAGE=true
Test: m nothing dist (before patch)
Test: m nothing dist (after patch)
Test: m nothing dist TARGET_BUILD_UNBUNDLED_IMAGE=true (after patch)
Test: examine ninja build graph for previous 3 tests
Bug: 188176942
Bug: 187064252
Change-Id: Ia7f8200c1bb1aaaa57c177d1698a114efe26bec7
(cherry picked from commit 0a30cc41bc)
The droidcore-unbundled target replaces the droidcore
target when TARGET_BUILD_UNBUNDLED_IMAGE=true. This
target is similar to droidcore in terms of dependencies,
but it has a smaller set of dist files. The intention is
to use TARGET_BUILD_UNBUNDLED_IMAGE=true with a
VSDK-using vendor build to supply Java dependencies from
prebuilts to avoid building (or needing) Java framework
sources.
Test: build with and without TARGET_BUILD_UNBUNDLED_IMAGE=true
Test: m nothing dist (before patch)
Test: m nothing dist (after patch)
Test: m nothing dist TARGET_BUILD_UNBUNDLED_IMAGE=true (after patch)
Test: examine ninja build graph for previous 3 tests
Bug: 188176942
Bug: 187064252
Merged-In: Ia7f8200c1bb1aaaa57c177d1698a114efe26bec7
Change-Id: Ia7f8200c1bb1aaaa57c177d1698a114efe26bec7
This prop is owned by OEM, OEM can set this if they want to disable
VABC.
Test: m dist, make sure generated OTA has VABC disabled
Bug: 185400304
Change-Id: Iceb2fb1f399d38a51722352a86ddf68af05fa24e
Merged-In: Iceb2fb1f399d38a51722352a86ddf68af05fa24e
This prop is owned by OEM, OEM can set this if they want to disable
VABC.
Test: m dist, make sure generated OTA has VABC disabled
Bug: 185400304
Change-Id: Iceb2fb1f399d38a51722352a86ddf68af05fa24e
VTS and some other tests would replace the system images with GSI. To
put the correct fingerprint in the test report, we need to know if the
new fingerprint format is in use. So, add a vendor build prop.
OEMs are reponsible for setting this build prop, or using other ways
to put the correct fingerprint in the test report.
Bug: 188824341
Test: boot the device, check build prop
Change-Id: I6bc7f01903865fc2c256d209debdab68cd9d1bb3
To dexpreopt in post processing, store config files for dexpreopt
Bug: 158843648
Test: m dist and check dexpreopt_config.zip
Change-Id: I5c63a5ffc10023994b14e3a63f50defb9194739f
Rename ro.product.enforce_debugfs_restrictions appropriately to
ro.product.debugfs_restrictions.enabled as per the sysprop naming
scheme.
Test: build, boot
Bug: 184381659
Change-Id: I65265098caf6382583aa53811f1b3330e58141c0
