Add toolchains to support cross compiling to aarch64-linux-musl and
arm-linux-musleabihf.
Bug: 236052820
Test: build arm and arm64 musl sysroots
Change-Id: I47a9322929baff2492c6e8db989ece01fcbeb133
* Reject -warnings-as-errors in the tidy_flags.
* Fix one grammar error in comment.
Bug: 229801437
Test: WITH_TIDY=1 make; make tidy-soong_subset
Change-Id: I4fae229add65131b64cd5a1d06507d27198010ed
* A follow up CL will set const NoWarningsAsErrorsInTidyFlags
and make it an error to use -warnings-as-errors in tidy_flags.
* Append TidyGlobalNoChecks after local tidy_checks to disable some checks.
* Append TidyGlobalNoErrorChecks after local tidy_checks_as_errors
(-warnings-as-errors) to allow some warnings globally.
* Move bugprone-signed-char-misuse and bugprone-branch-clone to
globalNoErrorCheckList so we can find and fix some of those warnings.
Bug: 229801437
Test: WITH_TIDY=1 make; make tidy-soong_subset
Change-Id: I0128b859b7be6eb9bbda1916b89a6a471b052150
* Save repeated strings in TidyFlags of build.ninja rules.
* Some of these checks were disabled to upgrade clang-tidy.
They could later be moved to TidyDefaultGlobalChecks and
TidyExternalVendorChecks if not breaking the build.
Some projects can then enable those checks locally.
Test: WITH_TIDY=1 make; make tidy-soong_subset
Change-Id: I70e4218c929e3c88f766f2c68c56c51356110e72
Vendor variants of AIDL-gen CC modules are by default different from
core variants. So instead of adding them manually to the list, AIDL
builder can mark them as MustUseVendorVariant.
Bug: 233616647
Test: m
Merged-In: Ifca0eb64d60ecb49d72cd0347336b69fb332c58c
Change-Id: Ifca0eb64d60ecb49d72cd0347336b69fb332c58c
Now only keep the ones that do not have thousands of warnings.
We can add more such checks locally or when more warnings
are fixed or suppressed in source files.
Bug: 112478838
Bug: 218759304
Test: make tidy-soong_subset
Change-Id: Iac93310441cdc64ed80e8c30159a08f95e1e9ebc
Previously we ran mutators in bp2build mode to add dependencies, now we
look up modules by name directly. Remove workarounds to allow bp2build
mode to not fail when adding/handling dependencies.
Test: m bp2build
Change-Id: Ibf6fd905150cac306e5c395902ef28f609f4df2a
* This tidy check is too noisy,
with more than 5000 cases over many files.
Bug: 218759304
Test: make tidy_soong_subset
Change-Id: I8a4b2589ecafdfcd9c19e8a4d59ce1784d1be694
This reverts commit fbc5effdfd.
Bug: http://b/197965342
Bug: http://b/230930120
Test: m and also test internal targets
Change-Id: I49b2d80ef4191ebb08bb6e80507042ee628c7f09
Merged-In: I49b2d80ef4191ebb08bb6e80507042ee628c7f09
Too many warnings from this library.
Treat it the same as tidyExternalVendor projects.
Bug: 231626164
Test: make tidy-vendor-google_arc_subset
Change-Id: Ieaf748517390cb6dfac7e85500666dc9f49ffd52
This respin contains performance improvements to clang-tidy and lld, and
fp16 cost fixes.
Test: presubmit
Bug: 219872481
Change-Id: I322a680cdc6ebc0f1fe3735ed087477e7f2508fe
Most of the variable export code for cc modules can be re-used for
exporting variables for java modules. Refactor this code into a more
composable structure for reuse.
Test: build/bazel/bp2build.sh
Test: manual comparison of
out/soong/soong_injection/cc_toolchain/constants.bzl
with previous output
Change-Id: Ie5a6fee08cc888b7dc69c3e324e5c3f8aa269a8f
The libgcc dependency is only used for _Unwind* symbols, remove
it now that libc_musl.so embeds libunwind.a.
Bug: 215802826
Test: m USE_HOST_MUSL=true host-native
Change-Id: I1ad4c1435fc0090c6df2d50d542d2203dfb94d7b
This reverts commit c75994d623.
Reason for revert: BTI shipped in S, and we're going to work on app compat outreach rather than regress this security feature.
Bug: https://issuetracker.google.com/216395195
Change-Id: Ic7143d756907ceb574a686258751840ce62c920f
The clang prebuilts now provide a single module with per-architecture
variants instead of a module per architecture.
Bug: 220019988
Test: m checkbuild
Change-Id: I39e2cf8ae14edf8510276dab38011afaef85822c
Merged-In: I39e2cf8ae14edf8510276dab38011afaef85822c
* Only prebuilts/gcc/linux-x86/host/x86_64-w64-mingw32-4.8/x86_64-w64-mingw32/bin
exists for x86_windows_host.
* Remove config.ToolPath;
add required -B flags into ToolchainCflags and ToolchainLdflags.
Bug: 218883919
Test: make droid tidy-soong_subset
Change-Id: I9a18bf8cc0cf84e091c7463b3bda316eaab53aa3
Bug: http://b/214080353
Bug: http://b/218805949
Reland switch to clang-r445002. The math test failures are fixed by
disabling FMA (-ffp-contract=off).
Test: - atest -a libcore.java.math.RunCSVTestsStrict
- select bionic tests on go/abtd
- atest -a \
gatm_algo_test:gatm_algo_test.TestAllGatmTestData/GatmAlgoTest#VerifyOutputValues/05temporaltracing_5frames_1lux_h_drc_free
- presubmit
Change-Id: I8383edb116caa63884507932406a5f49a223d006
Prevent linking against host glibc libraries by passing
--sysroot /dev/null to compile and link rules targeting musl.
Bug: 190084016
Test: m USE_HOST_MUSL=true host-native
Change-Id: Ief2c9507af9ae7cf5376cb874fcef0f2cbaf9994
Previously this would result in a Ninja variable lookup, instead, let's
error sooner in Soong.
Test: CI
Change-Id: Ib3e15dce84364e6f7a83b82cb5f0e8694ffc9d2a
Use x86_64-linux-musl and i686-linux-musl as the clang triple when
targeting musl.
Bug: 190084016
Test: m USE_HOST_MUSL=true host-native
Change-Id: Ibd19d9a5fbf4b67950745480d1e0ed8e02eeeba1
This will suppress warning like:
system/chre/chpp/app.c:637:3: warning: Call to function 'memset' is insecure
as it does not provide security checks introduced in the C11 standard.
Replace with analogous functions that support length arguments or provides
boundary checks such as 'memset_s' in case of C11
[clang-analyzer-security.insecureAPI.DeprecatedOrUnsafeBufferHandling]
Test: WITH_TIDY=1 CLANG_ANALYZER_CHECKS=1 make tidy-system-chre-chpp_subset
Change-Id: I87a7bba31aa10260fc408317a8cd54bf80c03152
This flag has never done anything for ld.lld, since it has this behavior
by default (and actually can't turn it off either). It's now being
removed as a valid flag from ld.lld, so it's time for Android to drop
the unnecessary flags. See https://reviews.llvm.org/D118942.
Test: builds
Change-Id: I22c344b7a55e275c1fbd5fef25f85e97a39de5c4
noOverrideExternalGlobalCflags is a new set of flags intended to be
added at the end of the command line for 3rd party projects like
external/, hardware/, vendor/, and more. Our previous flags for external
projects occur way too early on the command line, leading to issues with
use for suppressing diagnostics. Note that support for this variable in
build/make is currently unimplemented, as there are no projects that
would depend on it. It could be added in the future, if it turns out to
be useful.
This change initially applies `-Wno-unused-but-set-variable` and
`-Wno-unused-but-set-parameter` to external projects because these
upstream projects aren't easily fixed, and the risk of actual bugs is
low.
Bug: http://b/197240255
Test: Build and check warnings
Change-Id: I26f56c5c52725dddb70dd8130ad61270eac7a9aa
This is configured from Make by setting up Darwin+Arm64 as a HOST_CROSS
target (which is largely true, as binaries can't be executed on X86_64
machines). On the Soong side, it's a bit blurier, as we don't current
have any other users that are the same OS but not natively executable
(Linux/musl is executable, Windows/x86 is a different OS).
Instead of requiring cc_binary/etc to become multi-architecture-aware
and using something like common_first/MultiTarget, this defaults all
non-multi-architecture-aware modules to be built with both
architectures. It then adds a dependency with the
DarwinUniversalVariantTag so that supporting modules can get the outputs
of the other variant.
Cc uses that dependency tag to run lipo on shared libraries and binaries
so that the output of the x86_64 variant is actually a fat binary
including both architectures.
Bug: 203607969
Test: build sdk-repo targets on a Mac
Change-Id: Icbddb0a177c0ba19d3e0d11f8cf568e0d1ea3245
We do not profile host binaries. Moving to device binaries only helps
saving some disk space as well as shortening build time.
Test: presubmit
Change-Id: I09afb106d0e2ee250b3bc75d9dfd9cb8b6552c97
As they can't built arm64 binaries. All of our master-based builds are
now running on macOS 11 buildbots.
Bug: 203607969
Change-Id: I24c34a8365a399fbe43629ab5a22a1d53e3429b3
This just sets up the toolchain and allows Darwin+Arm64 to be specified
as a HostCross target. These variants will not be exported to Make, or
be installed on a Soong-only build. A future CL will add support for
universal binaries using these variants.
This config is a bit stranger than the regular 64/32 multilib, as it's
two primary 64-bit configs. And on a Darwin/X86 machine, the Arm64
versions are HostCross (doesn't work on the current machines), while a
Darwin/Arm64 machine, either version works (if Rosetta is installed).
Bug: 203607969
Change-Id: Iacaed77d267773672da027cd74917e33fb1c1e94
We're now packaging llvm-windres (aosp/1870074) which can be used as a
substitute for GNU binutils' windres, but we've also removed the last
consumer of .rc files in aosp/1867048 (external/mdnsresponder/).
In the worst case, we can bring back support for .rc files but instead
process them with llvm-windres.
Bug: 203794880
Test: lunch aosp_cf_x86_64_phone-userdebug && m
Change-Id: I30a873537e004b46a87191a932d076a89463c873
Modern Xcode versions only have a single macOS SDK installed anymore, so
we don't need to iterate through our supported versions, only check
whether the current version is supported.
This also makes it so that we can specify "11" (or soon, "12") to
specify all 11.x SDKs, instead of having to update our supported list
multiple times a year. This matches previous years that had multiple
patch releases on 10.14.y, 10.15.y, etc.
Bug: 203607969
Test: build on Mac, changing the supported SDK list
Change-Id: I42715f7e8dfec811de440a435927c14c742dede9
Test: USE_BAZEL_ANALYSIS=1 m adbd
Test: Manually verified config.bzl contains various x86_host flags after
bp2build.
Test: Unit tests
Change-Id: Ie9201ea2be4cd1c6659bea088a797cedbae37403
The Android SDK has been built with Make up until now, monolithically,
then split up into several sdk-repo zip files for different purposes.
The Mac and Windows SDKs really only need to contain the platform-tools
and build-tools pieces, but due to this monolithic sdk zip, we had to
build the whole SDK first.
This adds an `android_sdk_repo_host` module that can build these
platform-tools and build-tools zips.
Bug: 187222815
Change-Id: I55809e1d7447dd65e22461f921b2b8abb6d5f822
We turned this off in local tidy flags, but didn't disable it on global
tidy builds. Let's remove this for now, since the upkeep cost is large
on it, and it also is currently crashing in some cases.
Bug: http://b/199534745
Bug: http://b/194865868
Test: WITH_TIDY=1 m
Change-Id: Ife39ab4949c5baf825efff66326b276c6a2a59f4
