The 'fuzzer' sanitizer enables 'hwasan', which is not supported for Rust
static binaries. Make sure we skip applying this sanitizer to those
binaries.
Bug: 204776996
Test: SANITIZE_TARGET=fuzzer m <static_rust_executable>
Change-Id: I619cfab32b46c0811590973344eb5cdbe3f1a119
On HWASAN global sanitized builds, fuzz targets had hwasanFlags applied
twice. Rust only allows llvm-args flags to be passed in once however, so
this broke compilation.
Bug: 204915322
Test: SANITIZE_TARGET='hwaddress fuzzer' m <fuzzer>
Change-Id: I4f6c35e04bc4c891d9c66e8a53c3dcb88e8670cf
Refactor Rust prebuilts to support the new rust_stdlib_prebuilt_host
module type, and change the format for depending on the prebuilt host
stdlibs.
Bug: 140642453
Test: m
Change-Id: Ifbc4741818777934e917631c788b20911856c44a
rust_bindgen calls are missing a large set of -D cflags which control
how headers are interpretted for certain variations (such as
-D__ANDROID_VNDK__). This CL introduces the full set of these flags.
Bug: 205609024
Test: rust_bindgen vendor variant builds correctly.
Change-Id: Id781c1f88352e9c238c49619f0dce20f804cfc77
Rust installed files reside in "$MODULE_OUT/stripped/" when they are
stripped, otherwise they reside in "$MODULE_OUT". However, other parts
of Soong assume that installed files are always in $MODULE_OUT
(cc_modules place *unstripped* files in $MODULE_OUT/unstripped).
This notably causes problems when adding Rust modules as test data in
AndroidMkDataPaths. When Rust modules are parsed by AndroidMkDataPaths,
if they are stripped then they incorrectly get installed as test data
with the path:
<install_root>/<relative_install_path>/stripped/file.
This CL refactors how we handle Rust stripped output such that the
installed file always resides in $MODULE_OUT.
Bug: 171710847
Test: Installed files now always reside in $MODULE_OUT
Change-Id: I53a6ff57a0a5a55cd95ea78ae592ce22abfa20c9
This CL adds support for the MTE memtag_heap sanitizer. This is
controlled via inclusion of an ELF note.
Bug: 170672854
Test: Heap MTE-enabled Rust test binary triggers MTE
Change-Id: I2619818785e86a94667d02b30d102c83456b7925
Write toc files that list the exported symbols so dependents are
only rebuilt if the exported symbols change.
This exports the CC function TransformSharedObjectToToc, and also
removes an unused arg from its signature.
Bug: 178185435
Test: New Soong test passes.
Test: m <toc file path>
Change-Id: I7ab69bf7e7f32f25eb4c7ca9d18d877dac1511db
This CL adds Rust support for the SANITIZE_TARGET options.
This CL includes a couple small fixes to related to HWASAN, ASAN,
ensuring that the Never sanitize property is respected. Notably,
additional llvm-args are passed to ensure that HWASAN-ified Rust/C
interop works correctly.
Bug: 170672854
Bug: 204915322
Test: SANITIZE_TARGET globally applies hwasan to Rust targets
Change-Id: Ia904d07b4618f72cdc95c51f88961905c240ac53
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
Structs embedding binaryDecorator (rust_test, rust_benchmark, rust_fuzz)
are binaries as well, but won't pass checks against *binaryDecorator,
such as the check in StaticExecutable().
Add a binaryInterface that can be checked instead to simplify these
checks and ensure we catch all binaries.
Bug: 170672854
Test: rust_test, rust_benchmark return true StaticallyLinked
Change-Id: I2373d3663373a6977260785602a02d39a41320fe
This CL allows binaries to depend on whole static libraries which
don't begin with the 'lib' prefix.
Bug: 170672854
Test: Whole static library that doesn't have lib prefix can be linked
Change-Id: I908496d9369c7bec3232e2feed0599f6cf6d9383
Bug: http://b/202032071
Thin LTO provides similar performance gains to full LTO but takes less
time during build.
Thin LTO also doesn't cause the error with `-C lto` and `-C
llvm-args=--runtime-counter-relocation` in b/202032071.
Test: presubmit
Test: verify b/202032071
Change-Id: Id5749aaea598d4b555a9d959ed753a312ae36f26
Rust fuzzers were not packaging up their CC shared dependencies.
This would lead to fuzzers using the shared libraries included on
system, which may not be sanitized, leading to incorrect behavior.
This refactors the relevant code from CC and calls it from the Rust
build logic.
Bug: 202282599
Test: output rust fuzzer zip file includes shared dependencies.
Change-Id: I92196eb0141733797a67eae24f8e9aedea94c3bc
It's an error if a rust module's srcs is empty. Prior to this change
the output was a panic that complains loudly about a slice error and
nothing else. This change doesn't stop the panic, but at least adds
a bit of context so devs who make a simple mistake can more easily
fix it.
Test: SOONG_GEN_RUST_PROJECT=1 m nothing
Change-Id: Id7d8465d533413c3000699661222a53a7c8678f3
These crates were skipped before, leaving dependencies missing in
rust-project.json. Include them and mark them as `"is_proc_macro": true`
so that rust-analyzer can process them.
Fixes: 202290038
Test: SOONG_GEN_RUST_PROJECT=1 m nothing
Change-Id: Ia80e6f5e2f56a76608ba057075600e6b4424281b
This change makes the installation routine for rust be similar to that
of cc. Previously, rust.baseCompiler.install() (which internally calls
android.ModuleContext.InstallFile()) was not called when the module is
not installable. Although this may sound right at first glance, it is a
behavior different from that of cc and prevents an uninstallable rust
module from being packaged into a packaging module like
android_filesystem. This is because the packaging happens inside
InstallFile().
Fixing the issue by following the behavior of cc modules; Call
HideFromMake() or SkipInstall() when a rust module is not installable,
but call InstallFile() as long as the output file is valid.
InstallFile() internally skips the installation (but the packaging) when
HideFromMake() or SkipInstall() was called.
Bug: N/A
Test: atest MicrodroidHostTestCases
Change-Id: I15f4adc8544dac53647647d8bc4273f9f4acbeb2
This is actually required for some code.
Test: Modify the new test so it fails and see it fail.
Test: Build crates that fail without it.
Change-Id: I527752b765e5552aa2de7e201f056955e053e1f3
Move the cfg and feature flag calculation out of compilerFlags so that
it's a separate step.
The previous arrangement resulted in overridden compilerFlags which
must to set any additional cfgs/features before calling the base.
This is a bit confusing and undocumented behavior, so instead break
it out into a separate call that can itself be overriden.
Bug: N/A
Test: Soong tests pass
Change-Id: I28e4f707b3b3ca6eb621b7613c3737817f877bb8
Adds the require_root property to rust_test modules so that tests can
run as root. Also adds an install base to allow vendor tests to install
to a different path, otherwise installs to 'unrestricted' like cc_test.
Bug: 200602232
Test: test with require_root: true runs as root
Test: test installed to new base directory
Change-Id: Ifa716871ff87471f0c0caff7af558d47db85ccd1
