The hwasan runtime has stubs, make the function to add all version
variants of a shared library available to the sanitizer mutator.
Test: m checkbuild
Change-Id: Ie4124022140f8520834d7bf7b59f0499cd381083
There were multiple stacked issues with prebuilt library stubs that
cancelled eachother out. Prebuilts were never considered to be
DirectlyInAnyApex by the AndroidMk logic to handle stubs libraries
because it looked it up in the global list of modules in apexes
using the name with the "prebuilt_" prefix. Fixing that to use
ctx.BaseModuleName() exposed a second issue, that stubs variants
for prebuilt libraries were never created, so there was no latest
version to expose to Make.
Making the *prebuiltLibraryLinker type work with all of the
methods that handle stubs should really be done with an interface
and methods implemented on *libraryDecorator, but that would
also cause other types like that embed libraryDecorator to
participate in stubs that may trigger more issues. I'd like
to replace those methods anyways, so just manually handle
*prebuiltLibraryLinker for now.
Test: m checkbuild
Change-Id: I1267ee01659ad9ab11d75318c6c6bdbf8f72a061
This will be used to track who should be recognized for any CVEs
resulting from bugs found by a fuzz target.
Bug: 145745999
Test: n/a
Change-Id: I21e065ab8013d013b1d9f42981c729b996163387
For global ThinLTO, don't implicitly turn on LTO for static libraries,
but instead rely on mutator to generate correct variants.
Bug: 169004486
Test: GLBOAL_THINLTO=true m
Change-Id: I9cdeea706ec6dd4ad31f55b9e12a96b42176aa89
This is a quick fix to avoid build failures due to prebuilts lacking
stubs versions that the sources have, which causes dependencies on
older version stubs fail on master-art-host where the sources of e.g.
Bionic are no longer available.
More work is required to make the version handling of stubs DTRT - a
TODO comment outlines it.
Test: m nothing
Test: `m runtime-module-sdk`, check that the generated Android.bp
has the full stubs.versions list like the sources.
Bug: 169373910
Change-Id: I0ebfaf94f9d45a0e58d9785c40e7cea626f0ab83
I2954bb21c1cfdeb305f25cfb6c8711c930f6ed50 switched normalizeVersions
to work on ApiLevels, which inadvertantly caused it to return "current"
instead of "10000" for libraries that specify "current" in their stubs
property. ChooseSdkVersion couldn't handle "current" because it was
manually converting the version to an int. Switch ChooseSdkVersion
to use ApiLevels instead so that it can handle "current".
Test: m checkbuild
Change-Id: Id412359e092483ba419118dd03bc206fae702a96
Whole_static_libs required custom error checking when
AllowMissingDependencies was set because it could end up depending
on an empty list of objects, which would leave nothing in the
dependency tree that had been replaced with an ErrorRule.
Reuse the prebuilts case to depend on the .a file when there
are no objects and remove the custom error handling.
Test: TestEmptyWholeStaticLibsAllowMissingDependencies
Change-Id: Ic3216235f7e5ae8b5b6ab31ef2ca35c3994d82aa
rust_bindgen modules can't inherit properties in cc_defaults that would
be useful for generating bindings (such as cflags). This CL moves these
common properties out into a new struct in cc and adds that struct to
cc_default.
Additionally, Cppflags is added to rust_bindgen to make sure that these
get picked up as well from cc_defaults.
Bug: 163598610
Test: rust_bindgen module uses cflags in cc_defaults.
Test: New Soong test passes
Change-Id: I702442a355244dc01954083f98a2eebbcea12e47
liblog snapshot needs to sanitizer.address=false to avoid cycle in asan
builds. Adding that separately in library_sdk_member.go would start to
feel like whack-a-mole, so the snapshot generation is instead extended
to handle nested property structs.
This uses the BpPropertySet.AddProperty extension in
https://r.android.com/1423510, and common value optimisation now
recurses into non-anonymous structs, instead of comparing them as a
whole.
Test: m nothing
Test: `m SANITIZE_TARGET=address nothing` with prebuilts/runtime
present in the manifest and a fresh snapshot made with this
Bug: 151303681
Change-Id: I472554117a488e6c800045cb2ed59377778571a4
Experimental global ThinLTO mode for compilation time evaluation.
Bug: 169004486
Test: GLBOAL_THINLTO=true m
Change-Id: I2ff06397be8b65c7972f1e5f09ab1496949567ee
Pass the value of system_shared_libs to the Android.mk world, so that
prebuilt ELF check can exclude them from fix suggestions.
Bug: 141925662
Test: Write a bad cc_prebuilt_library module and check fix suggestions
Change-Id: I0cc61821765507180ce6a582bf8125a192f83a57
A host target is considered as being cross-compiled when the target
can't run natively on the build machine. For example, linux_glibc/x86_64
is a non-cross target on a standard x86/Linux machine, but is a cross
host on Mac. Previously, whether cross or not was a static attribute of
an OsType. For example, Windows was always considered as cross host,
while linux_bionic was not. This becomes a problem when we support more
host targets like linux_bionic/arm64 which should be cross-host on
standard x86/Linux machines.
This change removes HostCross from the OsClass type and instead adds a
property HostCross to the Target type. When a target is being added, it
is initialized to true when the target can't run natively on the current
build machine.
Bug: 168086242
Test: m
Change-Id: Ic37c8db918873ddf324c86b12b5412952b0f2be2
Keeping the int constant around for now as FutureApiLevelInt because
it's still useful in places that haven't adopted ApiLevel yet for
testing if their non-ApiLevel API level is current or not.
Test: treehugger
Bug: http://b/154667674
Change-Id: I47a7012703f41fdeb56f91edf9c83afa93042deb
Bug: http://b/119560349
Bug: http://b/168925229
To build with PGO instrumentation, the link step also needs
-fprofile-generate flag. Omitting this will result in a linker error.
This surfaces if module A depends on module B but PGO is used only for
module B.
Previous uses of PGO were avoiding or working around the issue but to
accommadate recent changes, this CL propagates the PGO linker flag to
module A.
Test: pgo-coral-config1 in release branch, TH
Change-Id: I12f6243152d4f5ea5526fbce03c25dba232ddab7
-fwhole_program_vtables depends on LTO being enabled. User manually
adding this through ldflag will break no LTO build configuration.
Test: TreeHugger
Change-Id: I557ec57831f01856a467f9567c3e7e25f8a6c566
We previously disabled inlining and unrolling completely during ThinLTO
in absense of PGO profile. For global ThinLTO, we want to better balance
binary size and performance.
We evaluated a number of combination of heuristics with global ThinLTO
configuration:
binary size change
no LTO baseline
no inline, no unroll -0.54%
no inline, unroll -0.50%
import-instr-limit=5, unroll +0.02%
import-instr-limit=10, unroll +0.13%
Loop unrolling does not contribute much to the binary size, therefore
it is re-enabled.
import-instr-limit=5 balances the binary size savings from ThinLTO and
size incrase due to additional optimisation.
Bug: 78485207
Bug: 169004486
Test: TreeHugger
Change-Id: I1c21153605e2ae42daa8857b06e27c081ee8ad85
Can happen if OUT_DIR is set on an absolute path.
Test: env RBE_remote_disabled=true OUT_DIR=/home_2/tmp/out m
Bug: 169048540
Fixes: 169048540
Change-Id: Ia6b91b13142d84e54d0dee8b4e04a91ad2cecc36
-fwhole-program-vtables cflag requires LTO to be enabled. Some projects
add this flag unconditionally, breaking builds that turn off LTO.
Make this an LTO property so that build system will handle adding the
flag when LTO is on.
Test: build
Bug: 169004486
Change-Id: Ifce12c09e1d6b8c2807314c3bb982b076efcaaa4
