ar doesn't produce the correct results when used for linking static
archives on Apple platforms, so instead use libtool -static which is
the official way to build static archives on those platforms.
Differential Revision: https://reviews.llvm.org/D62770
llvm-svn: 362311
These seemed to have been used in the past but were since removed
by the add_compile_flags_if_supported functions that combine these
these checks and adding the flag, but the original checks were never
removed.
Differential Revision: https://reviews.llvm.org/D62566
llvm-svn: 362058
This fixes the issue introduced by r362048 where we always use
pragma comment(lib, ...) for dependent libraries when the compiler
is Clang, but older Clang versions don't support this pragma so
we need to check first if it's supported before using it.
llvm-svn: 362055
As of r360984, LLD supports dependent libraries feature for ELF.
libunwind, libc++abi and libc++ have library dependencies: libdl librt
and libpthread, which means that when libunwind and libc++ are being
statically linked (using -static-libstdc++ flag), user has to manually
specify -ldl -lpthread which is onerous.
This change includes the lib pragma to specify the library dependencies
directly in the source that uses those libraries. This doesn't make any
difference when using linkers that don't support dependent libraries.
However, when using LLD that has dependent libraries feature, users no
longer have to manually specifying library dependencies when using
static linking, linker will pick the library automatically.
Differential Revision: https://reviews.llvm.org/D62090
llvm-svn: 362048
The libc++ typeinfo implementation is being improved to better
handle non-merged type names.
This patch takes advantage of that more correct behavior by delegating
to std::type_infos default operator== instead of doing pointer equality
ourselves.
However, libc++ still expects unique RTTI by default, and so we
should still fall back to strcmp when explicitly requested.
llvm-svn: 361916
This change is a consequence of the discussion in "RFC: Place libs in
Clang-dedicated directories", specifically the suggestion that
libunwind, libc++abi and libc++ shouldn't be using Clang resource
directory. Tools like clangd make this assumption, but this is
currently not true for the LLVM_ENABLE_PER_TARGET_RUNTIME_DIR build.
This change addresses that by moving the output of these libraries to
lib/$target/c++ and include/c++ directories, leaving resource directory
only for compiler-rt runtimes and Clang builtin headers.
Differential Revision: https://reviews.llvm.org/D59168
llvm-svn: 361432
When builing the hermetic static library, the compiler switch
-fvisibility-global-new-delete-hidden is necessary to get the new and
delete operator definitions made correctly. However, when those
definitions are not included in the library, then this switch does harm.
With lld (though not all linkers) setting STV_HIDDEN on SHN_UNDEF
symbols makes it an error to leave them undefined or defined via dynamic
linking that should generate PLTs for -shared linking (lld makes this a
hard error even without -z defs). Though leaving the symbols undefined
would usually work in practice if the linker were to allow it (and the
user didn't pass -z defs), this actually indicates a real problem that
could bite some target configurations more subtly at runtime. For
example, x86-32 ELF -fpic code generation uses hidden visibility on
declarations in the caller's scope as a signal that the call will never
be resolved to a PLT entry and so doesn't have to meet the special ABI
requirements for PLT calls (setting %ebx). Since these functions might
actually be resolved to PLT entries at link time (we don't know what the
user is linking in when the hermetic library doesn't provide all the
symbols itself), it's not safe for the compiler to treat their
declarations at call sites as having hidden visibility.
Differential Revision: https://reviews.llvm.org/D61572
llvm-svn: 360004
This change introduces support for building libcxxabi. The library
build should be complete, but not all CMake options have been
replicated in GN. We also don't support tests yet.
We only support two stage build at the moment.
Differential Revision: https://reviews.llvm.org/D60372
llvm-svn: 359805
The threaded cxa guard test attempted to test multithreaded waiting
by lining up a bunch of threads at a held init lock and releasing them.
The test initially wanted each thread to observe the lock being held,
but some threads may arive too late.
This patch cleans up the test and relaxes the restrictions.
llvm-svn: 359785
The error is:
libcxxabi/src/cxa_guard_impl.h: In instantiation of ‘__cxxabiv1::{anonymous}::LibcppMutex __cxxabiv1::{anonymous}::GlobalStatic<__cxxabiv1::{anonymous}::LibcppMutex>::instance’:
libcxxabi/src/cxa_guard_impl.h:529:62: required from here
libcxxabi/src/cxa_guard_impl.h:510:23: error: ‘__cxxabiv1::{anonymous}::LibcppMutex __cxxabiv1::{anonymous}::GlobalStatic<__cxxabiv1::{anonymous}::LibcppMutex>::instance’ has incomplete type
_LIBCPP_SAFE_STATIC T GlobalStatic<T>::instance = {};
^
llvm-svn: 359175
This patch does three main things:
(1) It re-writes the cxa guard implementation to make it testable.
(2) Adds support for recursive init detection on non-apple platforms.
(3) It adds a futex based implementation.
The futex based implementation locks and notifies on a per-object basis, unlike the
current implementation which uses a global lock for all objects. Once this patch settles
I'll turn it on by default when supported.
llvm-svn: 359060
Summary:
Ensure we re-export __cxa_throw_bad_array_new_length and
__cxa_uncaught_exceptions from libc++, since they are now
provided by libc++abi.
Doing this allows us to stop linking explicitly against libc++abi in
the libc++abi tests, since libc++ re-exports all the necessary symbols.
However, there is one caveat to that. We don't want libc++ to re-export
__cxa_uncaught_exception (the singular form), since it's only provided
for backwards compatibility. Hence, for the single test where we check
this backwards compatibility, we explicitly link against libc++abi.
PR27405
PR22654
Reviewers: EricWF
Subscribers: christof, jkorous, dexonsmith, libcxx-commits
Tags: #libc
Differential Revision: https://reviews.llvm.org/D60424
llvm-svn: 358690
On ARM the hand-rolled check causes a call to __aeabi_uidiv,
which we may not have a definition for.
Using the builtin avoids the generation of any library call.
llvm-svn: 358195
Summary:
The goal is to use a descriptive name for this feature, instead of just
using __arm__.
Reviewers: EricWF
Subscribers: javed.absar, kristof.beyls, christof, jkorous, dexonsmith, libcxx-commits
Tags: #libc
Differential Revision: https://reviews.llvm.org/D60520
llvm-svn: 358106
This reverts commit r357944 and r357949.
These changes failed to account for the fact that
the guard object is under aligned for atomic operations
on 32 bit platforms (It's aligned to 4 bytes but we require 8).
llvm-svn: 357958
The read of the guard variable by the caller is atomic,
and doesn't happen under a mutex.
Our internal reads and writes were non-atomic, because they happened
under a mutex.
The writes should always be atomic since they can be observed outside
of the lock.
Making the reads atomic is not strictly necessary under the current
global mutex approach, but will be under implementations that use a
futex (which I plan to land shortly). However, they should add little
additional cost.
llvm-svn: 357944
This patch is a part of a series of patches to cleanup
our implementation of __cxa_acquire et al. No functionality
change was intended.
This patch does two primary things.
It introduces the GuardObject class to abstract the reading
and writing to the guard object. In future, it will be used
to ensure atomic accesses are used when needed.
It also introduces the GuardValue class used to represent
values of the guard object. It is an abstraction to access
and write to the various different bits of a guard.
llvm-svn: 357804
This patch is a part of a series of cleanups to cxa_guard.cpp.
It should introduce no functionality change.
This patch refactors the use of the global mutex and condvar into
a RAII lock guard class. This improves correctness (since unlocks can't
be forgotten). It also allows the unification of the non-threading and
threading implementations.
llvm-svn: 357669
This change is similar to r356150, with the same motivation. The
only difference is that the method used to merge libunwind.a and
libc++abi.a had to be changed to use the same approach as libc++
since we no longer produce object libraries that could be linked
together as we did before. We reuse the libc++ script for merging
archives to avoid duplication between the two projects.
Differential Revision: https://reviews.llvm.org/D60173
llvm-svn: 357635
This is on by default, since on many platforms and configurations
libc++abi.a gets statically linked into shared libraries and/or
PIE executables.
This change is a followup to https://reviews.llvm.org/D60005 which
allows us to default to PIC code, but disable this if needed (for
example on WebAssembly where PIC code its currently compatible with
static linking).
Differential Revision: https://reviews.llvm.org/D60049
llvm-svn: 357551
With the current WebAssembly backend, objects built with -fPIC are not
compatible with static linking. libc++abi was (mistakenly?) adding
-fPIC to the objects it was including in a static library.
IIUC this change should also mean the static build can be more efficient
on all platforms.
Differential Revision: https://reviews.llvm.org/D60005
llvm-svn: 357322
This change is a consequence of the discussion in "RFC: Place libs in
Clang-dedicated directories", specifically the suggestion that
libunwind, libc++abi and libc++ shouldn't be using Clang resource
directory. Tools like clangd make this assumption, but this is
currently not true for the LLVM_ENABLE_PER_TARGET_RUNTIME_DIR build.
This change addresses that by moving the output of these libraries to
lib/<target> and include/ directories, leaving resource directory only
for compiler-rt runtimes and Clang builtin headers.
Differential Revision: https://reviews.llvm.org/D59013
llvm-svn: 355665
This matters on OSX because static linking orders is also the order dyld
uses to search for libs (the default - Two-level namespace). If system
libs (including unwind lib) are specified before local unwind lib, local
unwind lib would never be picked up by dyld.
Before:
$ otool -L lib/libc++abi.dylib
@rpath/libc++abi.1.dylib (compatibility version 1.0.0, current version 1.0.0)
/usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 1252.200.5)
@rpath/libunwind.1.dylib (compatibility version 1.0.0, current version 1.0.0)
After:
$ otool -L lib/libc++abi.dylib
@rpath/libc++abi.1.dylib (compatibility version 1.0.0, current version 1.0.0)
@rpath/libunwind.1.dylib (compatibility version 1.0.0, current version 1.0.0)
/usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 1252.200.5)
Thanks to Yuanfang Chen for the patch.
Differential Revision: https://reviews.llvm.org/D57496
llvm-svn: 355241
This include doesn't seem to be needed for the standalone build (it's
not being used by libc++ build either), but introduces unnecessary
dependency because HandleOutOfTreeLLVM performs checks that require
a working C++ library. We shouldn't require a working C++ library to
build libc++abi or libc++ (it's what we're building after all).
Differential Revision: https://reviews.llvm.org/D58333
llvm-svn: 354284
This changes add_custom_libcxx to also build libcxxabi and merges
the two into a static and hermetic library.
There are multiple advantages:
1) The resulting libFuzzer doesn't expose C++ internals and looks
like a plain C library.
2) We don't have to manually link in libstdc++ to provide cxxabi.
3) The sanitizer tests cannot interfere with an installed version
of libc++.so in LD_LIBRARY_PATH.
Differential Revision: https://reviews.llvm.org/D58013
llvm-svn: 354212
We build libc++ and libc++abi with -nodefaultlibs, so -rtlib=compiler-rt
has no effect and results in an 'argument unused during compilation'
warning which breaks the build when using -Werror. We can therefore drop
-rtlib=compiler-rt without any functional change; note that the actual
compiler-rt linking is handled by HandleCompilerRT.
Differential Revision: https://reviews.llvm.org/D58084
llvm-svn: 353786
We're building tests with -nostdlib which means that we need to
explicitly include the builtins library. When using libgcc (default)
we can simply include -lgcc_s on the link line, but when using
compiler-rt builtins we need a complete path to the builtins library.
This path is already available in CMake as <PROJECT>_BUILTINS_LIBRARY,
so we just need to pass that path to lit and if config.compiler_rt is
true, link it to the test.
Prior to this patch, running tests when compiler-rt is being used as
the builtins library was broken as all tests would fail to link, but
with this change running tests when compiler-rt bultins library is
being used should be supported.
Differential Revision: https://reviews.llvm.org/D56701
llvm-svn: 353208
There are several changes:
- Don't stringify Pythonized bools (that's why we're Pythonizing them)
- Support specifying target and sysroot via CMake variables
- Use consistent spelling for --target, --sysroot, --gcc-toolchain
llvm-svn: 353137
CMake has a standard way of setting target triple, sysroot and external
toolchain through CMAKE_<LANG>_COMPILER_TARGET, CMAKE_SYSROOT and
CMAKE_<LANG>_COMPILER_EXTERNAL_TOOLCHAIN. These are turned into
corresponding --target=, --sysroot= and --gcc-toolchain= variables add
included appended to CMAKE_<LANG>_FLAGS.
libunwind, libc++abi, libc++ provides their own mechanism through
<PROJECT>_TARGET_TRIPLE, <PROJECT>_SYSROOT and <PROJECT>_GCC_TOOLCHAIN
variables. These are also passed to lit via lit.site.cfg, and lit config
uses these to set the corresponding compiler flags when building tessts.
This means that there are two different ways of setting target, sysroot
and toolchain, but only one is properly supported in lit. This change
extends CMake build for libunwind, libc++abi and libc++ to also support
the CMake variables in addition to project specific ones in lit.
Differential Revision: https://reviews.llvm.org/D57670
llvm-svn: 353084
When linking library dependencies, we shouldn't need to export linked
libraries to dependents. We should be explicit about this in
target_link_libraries, otherwise other targets that depend on these such
as sanitizers get repeated (and possibly even conflicting) dependencies.
Differential Revision: https://reviews.llvm.org/D57456
llvm-svn: 352688
When linking library dependencies, we shouldn't need to export linked
libraries to dependents. We should be explicit about this in
target_link_libraries, otherwise other targets that depend on these such
as sanitizers get repeated (and possibly even conflicting) dependencies.
Differential Revision: https://reviews.llvm.org/D57456
llvm-svn: 352654
Erik Pilkington