Refactor the get_llvm_lit_path() logic to respect LLVM_EXTERNAL_LIT,
and require the fallback to be defined explicitly
as LLVM_DEFAULT_EXTERNAL_LIT. This fixes building libcxx standalone
after r346888.
The old logic was using LLVM_EXTERNAL_LIT both as user-defined cache
variable and an optional pre-definition of default value from caller
(e.g. libcxx). It included a hack to make this work by assigning
the value back and forth but it was fragile and stopped working
in libcxx.
The new logic is simpler and more transparent. Default value is
provided in a separate variable, and used only when user-specified
variable is empty (i.e. not overriden).
Differential Revision: https://reviews.llvm.org/D57282
llvm-svn: 352374
The check_library_exists CMake uses a custom symbol definition. This
is a problem when checking for C library symbols because Clang
recognizes many of them as builtins, and returns the
-Wbuiltin-requires-header (or -Wincompatible-library-redeclaration)
error. When building with -Werror which is the default, this causes
the check_library_exists check fail making the build think that C
library isn't available.
To avoid this issue, we should use a symbol that isn't recognized by
Clang and wouldn't cause the same issue. __libc_start_main seems like
reasonable choice that fits the bill.
Differential Revision: https://reviews.llvm.org/D57142
llvm-svn: 352341
This is useful when the static libc++abi library is being linked into
shared libraries that may be used in with other shared libraries that
use different C++ library. We want to avoid avoid exporting libc++abi
or libc++ symbols in those cases. This achieved by a new CMake option
which can be enabled by libc++abi vendors as needed.
Differential Revision: https://reviews.llvm.org/D56026
llvm-svn: 352017
When built within the llvm runtimes directory, the runtimes
CMakeLists.txt adds the same.
Differential Revision: https://reviews.llvm.org/D56979
llvm-svn: 351873
all missed!
Thanks to Alex Bradbury for pointing this out, and the fact that I never
added the intended `legacy` anchor to the developer policy. Add that
anchor too. With hope, this will cause the links to all resolve
successfully.
llvm-svn: 351731
to reflect the new license. These used slightly different spellings that
defeated my regular expressions.
We understand that people may be surprised that we're moving the header
entirely to discuss the new license. We checked this carefully with the
Foundation's lawyer and we believe this is the correct approach.
Essentially, all code in the project is now made available by the LLVM
project under our new license, so you will see that the license headers
include that license only. Some of our contributors have contributed
code under our old license, and accordingly, we have retained a copy of
our old license notice in the top-level files in each project and
repository.
llvm-svn: 351648
to reflect the new license.
We understand that people may be surprised that we're moving the header
entirely to discuss the new license. We checked this carefully with the
Foundation's lawyer and we believe this is the correct approach.
Essentially, all code in the project is now made available by the LLVM
project under our new license, so you will see that the license headers
include that license only. Some of our contributors have contributed
code under our old license, and accordingly, we have retained a copy of
our old license notice in the top-level files in each project and
repository.
llvm-svn: 351636
This installs the new developer policy and moves all of the license
files across all LLVM projects in the monorepo to the new license
structure. The remaining projects will be moved independently.
Note that I've left odd formatting and other idiosyncracies of the
legacy license structure text alone to make the diff easier to read.
Critically, note that we do not in any case *remove* the old license
notice or terms, as that remains necessary until we finish the
relicensing process.
I've updated a few license files that refer to the LLVM license to
instead simply refer generically to whatever license the LLVM project is
under, basically trying to minimize confusion.
This is really the culmination of so many people. Chris led the
community discussions, drafted the policy update and organized the
multi-year string of meeting between lawyers across the community to
figure out the strategy. Numerous lawyers at companies in the community
spent their time figuring out initial answers, and then the Foundation's
lawyer Heather Meeker has done *so* much to help refine and get us ready
here. I could keep going on, but I just want to make sure everyone
realizes what a huge community effort this has been from the begining.
Differential Revision: https://reviews.llvm.org/D56897
llvm-svn: 351631
With this patch, the copies of the files ItaniumDemangle.h,
StringView.h, and Utility.h are kept byte-for-byte in sync between
libcxxabi and llvm. All differences (namespaces, fallthrough, and
unreachable macros) are defined in each copies' DemanglerConfig.h.
This patch also adds a script to copy changes from libcxxabi
(cp-to-llvm.sh), and a README.txt explaining the situation.
Differential revision: https://reviews.llvm.org/D53538
llvm-svn: 351474
Summary:
std::bad_array_length was added by n3467, but this never made it into C++.
This commit removes the definition of std::bad_array_length from the headers
AND from the shared library. See the comments in the ABI changelog for details
about the ABI implications of this change.
Reviewers: mclow.lists, dexonsmith, howard.hinnant, EricWF
Subscribers: christof, jkorous, libcxx-commits
Differential Revision: https://reviews.llvm.org/D54804
llvm-svn: 347903
Summary:
This (very specialized) function was added to enable an LLDB use case.
Now that a more generic interface (overriding of parser functions -
D52992) is available, and LLDB has been converted to use that (D54074),
the function is unused and can be removed.
Reviewers: erik.pilkington, sgraenitz, rsmith
Subscribers: mgorny, hiraditya, christof, libcxx-commits, llvm-commits
Differential Revision: https://reviews.llvm.org/D54893
llvm-svn: 347670
This is needed when cross-compiling for a different target since
CFLAGS may contain additional flags like -resource-dir which
change the location in which compiler-rt builtins are found.
Differential Revision: https://reviews.llvm.org/D54371
llvm-svn: 346820
Summary:
This silences the two -Wimplicit-fallthrough warnings clang finds in
ItaniumDemangle.h in libc++abi.
Clang does not have a GNU attribute spelling for this attribute, so this
is necessary.
I will commit the same change to the LLVM demangler soon.
Reviewers: EricWF, ldionne
Subscribers: christof, erik.pilkington, cfe-commits
Differential Revision: https://reviews.llvm.org/D53985
llvm-svn: 345870
Summary:
This uses CRTP (for performance reasons) to allow a user the override
demangler functions to implement custom parsing logic. The motivation
for this is LLDB, which needs to occasionaly modify the mangled names.
One such instance is already implemented via the TypeCallback member,
but this is very specific functionality which does not help with any
other use case. Currently we have a use case for modifying the
constructor flavours, which would require adding another callback. This
approach does not scale.
With CRTP, the user (LLDB) can override any function it needs without
any special support from the demangler library. After LLDB is ported to
use this instead of the TypeCallback mechanism, the callback can be
removed.
More context can be found in D50599.
Reviewers: erik.pilkington, rsmith
Subscribers: christof, ldionne, llvm-commits, libcxx-commits
Differential Revision: https://reviews.llvm.org/D52992
llvm-svn: 344607
Summary:
I copied the sanitizer-related logic in libcxx/lib/CMakeLists.txt. In
the future, it would be great to avoid duplicating this logic in the
compiler, libc++ and libc++abi.
Reviewers: EricWF
Subscribers: mgorny, christof, dexonsmith, libcxx-commits, davide
Differential Revision: https://reviews.llvm.org/D53028
llvm-svn: 344191
Summary:
This was committed back in september (D51463), but it seems it never
made it into the libcxxabi copy.
The original commit message was:
The hash computed for an ArrayType was different when first constructed
versus when later profiled due to the constructor default argument, and
we were not tracking constructor / destructor variant as part of the
mangled name AST, leading to incorrect equivalences.
Reviewers: erik.pilkington, rsmith, EricWF
Subscribers: christof, ldionne, libcxx-commits
Differential Revision: https://reviews.llvm.org/D53063
llvm-svn: 344121
The "dynamic_cast error 2" error can apparently happen when the same
type (with RTTI) is defined in more than one translation unit, and
those translation units are linked together. This is technically an
ODR violation, but making the error message more obvious is still
helpful.
llvm-svn: 344052
This patch fixes a bug where exceptions in 32 bit builds
would be incorrectly aligned because malloc only provides 8 byte aligned
memory where 16 byte alignment is needed.
This patch makes libc++abi correctly use posix_memalign when it's
available. This requires defining _LIBCPP_BUILDING_LIBRARY so that
libc++ only defines _LIBCPP_HAS_NO_ALIGNED_ALLOCATION when libc doesn't
support it and not when aligned new/delete are disable for other
reasons.
This bug somehow made it into the 7.0 release, making it a regression.
Therefore this patch should be included in the next dot release.
llvm-svn: 342815
r340663 - Allow Allocator::make to make a node of a different type than that
requested.
r340664 - Add documentation comment to ForwardTemplateReference.
r340665 - Fix ExpandedSpecialSubstitution demangling for Sa and Sb.
r340670 - Allow demangler's node allocator to fail, and bail out of the entire
demangling process when it does.
llvm-svn: 340671
Move Itanium demangler implementation into a header file and add visitation support.
Summary:
This transforms the Itanium demangler into a generic reusable library that can
be used to build, traverse, and transform Itanium mangled name trees.
This is in preparation for adding a canonicalizing demangler, which
cannot live in the Demangle library for layering reasons. In order to
keep the diffs simpler, this patch moves more code to the new header
than is strictly necessary: in particular, all of the printLeft /
printRight implementations can be moved to the implementation file.
(And indeed we could make them non-virtual now if we wished, and remove
the vptr from Node.)
All nodes are now included in the Kind enumeration, rather than omitting
some of the Expr nodes, and the three different floating-point literal
node types now have distinct Kind values.
As a proof of concept for the visitation / matching mechanism, this
patch implements a Node dumping facility on top of it, replacing the
prior mechanism that produced the pretty-printed output rather than a
tree dump. Sample dump output:
FunctionEncoding(
NameType("int"),
NameWithTemplateArgs(
NestedName(
NameWithTemplateArgs(
NameType("A"),
TemplateArgs(
{NameType("B")})),
NameType("f")),
TemplateArgs(
{NameType("int")})),
{},
<null>,
QualConst, FunctionRefQual::FrefQualLValue)
As a next step, it would make sense to move the LLVM high-level interface to
the demangler (the itaniumDemangler function and ItaniumPartialDemangler class)
into the Support library, and implement them in terms of the Demangle library.
This would allow the libc++abi demangler implementation to be an identical copy
of the llvm Demangle library, and would allow the LLVM implementation to reuse
LLVM components such as llvm::BumpPtrAllocator, but we'll need to decide how to
coordinate that with the MS ABI demangler, so I'm not doing that in this patch.
No functionality change intended other than the behavior of dump().
Reviewers: erik.pilkington, zturner, chandlerc, dlj
Subscribers: aheejin, llvm-commits
Differential Revision: https://reviews.llvm.org/D50930
llvm-svn: 340207
This function calls a callback whenever a <type> is parsed.
This is necessary to implement FindAlternateFunctionManglings in LLDB, which
uses a similar hack in FastDemangle. Once that function has been updated to use
this version, FastDemangle can finally be removed.
Differential revision: https://reviews.llvm.org/D50586
llvm-svn: 339580
Stack overflow on invalid. While collapsing references, we were skipping over a
cycle check in ForwardTemplateReference leading to a stack overflow. This commit
fixes the problem by duplicating the cycle check in ReferenceType.
llvm-svn: 338190
Summary:
rL337867 introduced two new cmake_dependent_option options:
- LIBCXXABI_INSTALL_STATIC_LIBRARY
- LIBCXXABI_INSTALL_SHARED_LIBRARY
They depend on LIBCXXABI_ENABLE_STATIC and LIBCXXABI_ENABLE_SHARED
and so therefore need to (it seems) come after the declaration of
these two options.
Subscribers: mgorny, aheejin, christof, ldionne, cfe-commits
Differential Revision: https://reviews.llvm.org/D49825
llvm-svn: 337982
Currently it's only possible to control whether shared or static library
build of libc++, libc++abi and libunwind is enabled or disabled and
whether to install everything we've built or not. However, it'd be
useful to have more fine grained control, e.g. when static libraries are
merged together into libc++.a we don't need to install libc++abi.a and
libunwind.a. This change adds this option.
Differential Revision: https://reviews.llvm.org/D49573
llvm-svn: 337867
We really should set *status to memory_alloc_failure, but we need to refactor
the demangler a bit to properly propagate the failure up the stack. Until then,
its better to explicitly terminate then rely on a null dereference crash.
rdar://31240372
llvm-svn: 337759
This allows handling SEH based exceptions, with unwind functions
provided by libgcc.
Differential Revision: https://reviews.llvm.org/D49638
llvm-svn: 337754
Currently it's possible to select whether to statically link unwinder
or the C++ ABI library, but this option applies to both the shared
and static library. However, in some scenarios it may be desirable to
only statically link unwinder and C++ ABI library into static C++
library since for shared C++ library we can rely on dynamic linking
and linker scripts. This change enables selectively enabling or
disabling statically linking only to shared or static library.
Differential Revision: https://reviews.llvm.org/D49502
llvm-svn: 337668
ItaniumDemangle had a small NFC refactor to make some of its
code reusable by the newly added Microsoft demangler. To keep
the libcxxabi demangler as close as possible to the master copy
this refactor is being merged over.
Differential Revision: https://reviews.llvm.org/D49575
llvm-svn: 337582
Do not use LLVM_RUNTIMES_LIBDIR_SUFFIX variable which is an internal
variable used by the runtimes build from individual runtimes, instead
set per-runtime librarhy directory suffix variable which is necessary
for the sanitized runtimes build to install libraries into correct
location.
Differential Revision: https://reviews.llvm.org/D49121
llvm-svn: 336713
The alignment specified by a constant for the field
`BumpPointerAllocator::InitialBuffer` exceeded the alignment
guaranteed by `malloc` and `new` on Windows. This change set
the alignment value to that of `long double`, which is defined
by the used platform.
It fixes https://bugs.llvm.org/show_bug.cgi?id=37944.
Differential Revision: https://reviews.llvm.org/D48889
llvm-svn: 336312
vector is a generic C++ header, whereas __config is libc++-specific, so
we can look for it instead to guarantee we're finding a libc++
installation. This was suggested by Eric in https://reviews.llvm.org/D48694.
This is less important now that we're limiting the header search to the
specified directories (which definitely shouldn't have any other C++
library's headers anyway), but it shouldn't hurt either. There's a
chance some other library could also be providing a __config header, so
there's still a trade-off there. It would be ideal if we could check for
the presence of both __config and vector in the same directory, but
there doesn't seem to be any easy way to do that in CMake.
llvm-svn: 336034
Right now, when libc++abi is locating libc++ headers, it specifies
several search locations, but it also doesn't prevent CMake from looking
for those headers in system directories. I don't know if this was
intentional or an oversight, but it has several issues:
* We're looking specifically for the vector header, which could just as
easily be found in a libstdc++ (or other C++ library) installation.
* No system I know of places their C++ headers directly in system
include directories (they're always under a C++ subdirectory), so the
system search will never succeed.
* find_path searches system paths before the user-specified PATHS, so
if some system does happen to have C++ headers in its system include
directories, those headers will be preferred, which doesn't seem
desirable.
It makes sense to me to limit this header search to the explicitly
specified paths (using NO_DEFAULT_PATH, as is done for the other
find_path call in this file), but I'm putting it up for review in case
there's some use case I'm not thinking of.
Differential Revision: https://reviews.llvm.org/D48694
llvm-svn: 336032
This change adds a support for multiarch style runtimes layout, so in
addition to the existing layout where runtimes get installed to:
lib/clang/$version/lib/$os
Clang now allows runtimes to be installed to:
lib/clang/$version/$target/lib
This also includes libc++, libc++abi and libunwind; today those are
assumed to be in Clang library directory built for host, with the
new layout it is possible to install libc++, libc++abi and libunwind
into the runtime directory built for different targets.
The use of new layout is enabled by setting the
LLVM_ENABLE_RUNTIME_TARGET_DIR CMake variable and is supported by both
projects and runtimes layouts. The runtimes CMake build has been further
modified to use the new layout when building runtimes for multiple
targets.
Differential Revision: https://reviews.llvm.org/D45604
llvm-svn: 335809
The paths output from llvm-config --cmakedir and from clang
--print-libgcc-file-name can contain backslashes, while CMake
can't handle the paths in this form.
This matches what compiler-rt already does (since SVN r203789
and r293195).
Differential Revision: https://reviews.llvm.org/D48355
llvm-svn: 335171
Patch by Ryan Prichard
If the destination type does not derive from the static type, we can skip
the search_above_dst call, but we still need to run the
!does_dst_type_point_to_our_static_type block of code. That block of code
will increment info->number_to_dst_ptr to 2, and because dest isn't derived
from static, the cast will ultimately fail.
Fixes PR33439
Reviewed as https://reviews.llvm.org/D36447
llvm-svn: 332767
This adds the test which was mistakenly not committed in r332763.
Patch by Ryan Prichard
Propagate the found_our_static_ptr and found_any_static_type flags from
__vmi_class_type_info::search_above_dst to its caller.
Fixes PR33425 and PR33487
Reviewed as https://reviews.llvm.org/D36446
llvm-svn: 332764
Patch by Ryan Prichard
Propagate the found_our_static_ptr and found_any_static_type flags from
__vmi_class_type_info::search_above_dst to its caller.
Fixes PR33425 and PR33487
Reviewed as https://reviews.llvm.org/D36446
llvm-svn: 332763
This is a follow-up change to r331150. The CL moved the macro from individual
file to build file, but the macro is missed in a test config file.
https://reviews.llvm.org/D46385
Patch from Taiju Tsuiki <tzik@chromium.org>!
llvm-svn: 331450
_LIBCPP_ENABLE_CXX17_REMOVED_UNEXPECTED_FUNCTIONS is currently used to
bring back std::unexpected, which is removed in C++17, but still needed
for libc++abi for backward compatibility.
This macro used to define in cxa_exception.cpp only, but actually
needed for all sources that touches exceptions.
So, a build-system-level macro is better fit to define this macro.
https://reviews.llvm.org/D46056
Patch from Taiju Tsuiku <tzik@chromium.org>!
llvm-svn: 331150
This is basically part 2 of r313694.
It's a little unfortunate that I had to copy-paste atomic_support.h,
but I don't really see any alternative.
The refstring.h changes are the same as the libcxx changes in r313694.
llvm-svn: 330162
Summary:
exception_header->exceptionDestructor is a void(*)(void*) function
pointer; however, it can point to destructors like std::
exception::~exception that don't match that type signature.
Reviewers: pcc, vitalybuka
Reviewed By: vitalybuka
Subscribers: kcc, christof, cfe-commits
Differential Revision: https://reviews.llvm.org/D45455
llvm-svn: 329629
Strictly in a conversion operator's type, a <template-param> refers to a
<template-arg> that is further ahead in the mangled name. Instead of
doing a second parse to resolve these, introduce a
ForwardTemplateReference Node and back-patch the referenced
<template-arg> when we're in the right context.
This is also a correctness fix, previously we would only do a second
parse if the <template-param> was out of bounds in the current set of
<template-args>. This lead to misdemangles (gasp!) when the conversion
operator was a member of a templated struct, for instance.
llvm-svn: 328464
Rather than eagerly propagating up parameter pack sizes in Node ctors,
find the parameter pack size during printing. This is being done to
support back-patching forward referencing <template-param>s.
llvm-svn: 328463
C++17 removes `std::unexpected_handler`, but libc++abi needs it to define
`__cxa_exception`. When building against libc++, this is easily rectified by
telling libc++ we're building the library. We already do this in the other
places where we need these symbols.
Differential Revision: https://reviews.llvm.org/D42987
llvm-svn: 324542
This commit cleans up the expression parser, using a new style:
- parse* functions now return Node pointers.
- The mangled name is now held in Db and accessed with look() and consume()
- LLVM coding style
This style is meant to avoid the 2 most common types of bugs in the
old demanger, namely misusing the Names stack (ie, calling back() on
empty) and going out of bounds on the mangled name. I also think it
makes the demangler a lot cleaner.
Differential revision: https://reviews.llvm.org/D41887
llvm-svn: 324111
This commit changes how variadic templates are represented in the
demangler, in order to fix some longstanding bugs. Now instead of
expanding variadic templates during parsing, the expansion is done
during printing by reusing the unexpanded AST. This allows the
demangler to handle cases where multiple packs contribute to a single
production, and correctly handle "Dp" and "sp" productions, which
corrispond to pack expansions in type and expression contexts.
Differential revision: https://reviews.llvm.org/D41885
llvm-svn: 323906
[cmake] [libcxxabi] Call llvm_setup_rpath() when adding shared libraries.
Clang and llvm already use llvm_setup_rpath(), so this change will
help standarize rpath usage across all projects.
Differential Revision: https://reviews.llvm.org/D42460
llvm-svn: 323495
Shoaib Meenai pointed out this will break standalone builds can be built without llvm.
Differential Revision: https://reviews.llvm.org/D42460
llvm-svn: 323458
Clang and llvm already use llvm_setup_rpath(), so this change will
help standarize rpath usage across all projects.
Differential Revision: https://reviews.llvm.org/D42460
llvm-svn: 323455
This fixes:
src/cxa_default_handlers.cpp:25:13: error: unused function 'demangling_terminate_handler' [-Werror,-Wunused-function]
Patch by Thomas Anderson!
Differential Revision: https://reviews.llvm.org/D42399
llvm-svn: 323397
When CMAKE_SYSROOT or CMAKE_FIND_ROOT_PATH is set, cmake
recommends setting CMAKE_FIND_ROOT_PATH_MODE_INCLUDE=ONLY
globally which means find_path() always prepends CMAKE_SYSROOT or
CMAKE_FIND_ROOT_PATH to all paths used in the search.
However, these find_path() invocations are looking for paths in
the libcxx and libunwind projects on the host system, not the
target system, which can be done by passing
NO_CMAKE_FIND_ROOT_PATH.
Differential Revision: https://reviews.llvm.org/D41623
llvm-svn: 323145
Summary:
Don't print, possibly erroneous, warning if
LIBCXXABI_INCLUDE_TESTS is false.
This patch fixes a problem introduced in r291367.
Differential Revision: https://reviews.llvm.org/D42229
llvm-svn: 322870
This patch updates libc++abi's HandleOutOfTreeLLVM.cmake to match
libc++'s -- and more importantly, to fix a bug where llvm-lit wasn't
found/created when libc++abi was built out-of-tree. This prevented
the test suite from running.
llvm-svn: 322768
Summary:
We want to automatically copy the appropriate mailing list
for review requests to the libc++abi repository.
For context, see the proposal and discussion here:
http://lists.llvm.org/pipermail/cfe-dev/2017-November/056032.html
Similar to D40500, I set up a new Diffusion repository with callsign
"CXXA" for libc++abi:
https://reviews.llvm.org/source/libcxxabi/
This explicitly updates libcxxabi's .arcconfig to point to the new
CXX repository in Diffusion, which will let us use Herald rule H268.
Reviewers: phosek, beanz, EricWF, compnerd
Reviewed By: phosek
Subscribers: cfe-commits, klimek, sammccall, dlj, bkramer
Differential Revision: https://reviews.llvm.org/D40501
llvm-svn: 319713
LLVM is gaining install-*-stripped targets to perform stripped installs,
and in order for this to be useful for install-distribution, all
potential distribution components should have stripped installation
targets. LLVM has a function to create these install targets, but since
we can't use LLVM CMake functions in libc++abi, let's do it manually.
Differential Revision: https://reviews.llvm.org/D40681
llvm-svn: 319499
object is sufficiently aligned.
r303175 annotated field unwindHeader of __cxa_exception with attribute
'aligned' to ensure the thrown object following the __cxa_exception
header was sufficiently aligned. This caused changes in the field
offsets of __cxa_exception relative to the start of the thrown object,
which was an ABI breaking change for some clients.
Instead of annotating field unwindHeader, this commit inserts extra
space before the header. This ensures the thrown object following the
header is sufficiently aligned without changing the field offsets, thus
avoiding any ABI breakages.
rdar://problem/25364625
rdar://problem/35556163
llvm-svn: 319123
This is useful in cases where we only build static library and
libc++abi.a is combined with libc++.a into a single archive in which
case we don't want to have libc++abi.a installed separately. The same
option is already provided by libcxx CMake build.
Differential Revision: https://reviews.llvm.org/D40194
llvm-svn: 318568
When using LLVM unwinder and static unwinder option is set, merge
libunwind and libc++abi objects into a single archive. libc++ already
supports merging libc++abi.a and libc++.a into a single archive; with
this change, it is possible to also include libunwind.a in the same
archive which is useful when doing static link and using libc++ as
a default C++ library and compiler-rt as a default runtime library.
Differential Revision: https://reviews.llvm.org/D39949
llvm-svn: 318563
compiler-rt recently added the __asan_handle_no_return() function that libc++abi
needs to use, however older versions of compiler-rt don't declare this interface
publicly and that breaks the libc++abi build.
This patch attempts to fix the issues by declaring the asan function explicitly,
so we don't depend on compiler-rt to provide the declaration.
llvm-svn: 313308
Summary:
compiler-rt recently added the `__asan_handle_no_return()` function that libc++abi needs to use, however older versions of compiler-rt don't provide this interface and that breaks the libc++abi build.
This patch attempts to fix the issues by using a macro to detect if `asan_interface.h` is new enough to provide the function.
See D37871
Reviewers: phosek, vitalybuka
Reviewed By: phosek, vitalybuka
Subscribers: dberris, cfe-commits
Differential Revision: https://reviews.llvm.org/D37872
llvm-svn: 313304
The ASan runtime on many systems intercepts cxa_throw just so it
can call asan_handle_no_return first. Some newer systems such as
Fuchsia don't use interceptors on standard library functions at all,
but instead use sanitizer-instrumented versions of the standard
libraries. When libc++abi is built with ASan, cxa_throw can just
call asan_handle_no_return itself so no interceptor is required.
Patch by Roland McGrath
Differential Revision: https://reviews.llvm.org/D37229
llvm-svn: 313215
The ASan runtime on many systems intercepts cxa_throw just so it
can call asan_handle_no_return first. Some newer systems such as
Fuchsia don't use interceptors on standard library functions at all,
but instead use sanitizer-instrumented versions of the standard
libraries. When libc++abi is built with ASan, cxa_throw can just
call asan_handle_no_return itself so no interceptor is required.
This is a re-land of r311045, which has become safe after r311869
changed compiler-rt to declare __asan_handle_no_return.
Patch by Roland McGrath
Differential Revision: https://reviews.llvm.org/D37229
llvm-svn: 312606
The ASan runtime on many systems intercepts cxa_throw just so it
can call asan_handle_no_return first. Some newer systems such as
Fuchsia don't use interceptors on standard library functions at all,
but instead use sanitizer-instrumented versions of the standard
libraries. When libc++abi is built with ASan, cxa_throw can just
call asan_handle_no_return itself so no interceptor is required.
Patch by Roland McGrath
Differential Revision: https://reviews.llvm.org/D36599
llvm-svn: 311045
The demangler now demangles by producing an AST, then traverses that
AST to produce a demangled name. This is done for performance reasons,
now the demangler doesn't manuiplate std::strings, which hurt
performance and caused string operations to be inlined into the
parser, leading to large code size and stack usage.
Differential revision: https://reviews.llvm.org/D35159
llvm-svn: 309340
This patch removes the dependancy on libc++'s __refstring header,
which was only a header in the first place so that libc++abi could
build library code using it, and not because libc++ needed it in
the headers.
This patch allows libc++ to stop shipping <__refstring> publicaly
at the cost of duplicating it across projects. Ideally libc++abi
would always require the libc++ sources when building, but that's
a separate discussion I plan to start on the mailing lists shortly.
llvm-svn: 307748
This is going to be used by the runtime build in the multi-target
setup to allow using different install prefix for each target.
Differential Revision: https://reviews.llvm.org/D33761
llvm-svn: 307611
Rather than manually checking for support for the spelling of the C++
standard, indicate to CMake that we require that the compiler support
C++11 and that we compile without the GNU extensions. This simplifies
the flags handling in libc++abi itself by relying on CMake to translate
the flag and add it as appropriate.
llvm-svn: 305175
Use the POSITION_INDEPENDENT_CODE target property to indicate that we
should be building with -fPIC or the equivalent flag based on the
toolchain that we are using. This makes the check more portable and
simplifies the flags management. Because we don't want this setting to
propagate in the case of an in-tree build, set the property on the
targets we construct explicitly rather than setting
CMAKE_POSITION_INDEPENDENT_CODE to ON globally.
llvm-svn: 305174
Use the C++11 (formalised in C++17) tag to indicate a fallthrough in the
switch case. Silences a -Wimplicit-fallthrough warning with gcc:7
llvm-svn: 305173
Refactor cmake to remove dependence on LLVM's cmake modules.
This improves handling of cmake checks when cross compiling and brings
libcxxabi in line with libcxx and other project modules.
Differential revision: https://reviews.llvm.org/D33635
llvm-svn: 304374
As per r241993, libunwind_ext.h is not used anymore, and thus only the public libunwind includes are needed.
This eases distro packaging efforts and removes an unneeded requirement for out-of-tree building.
Reviewed as D33178
llvm-svn: 304359
The problem was that if base_name() was called from a context without
an actual base name, it could gulp up the entire string, which can
result in recursive duplications. The fix is to be more strict as to
what qualifies as a base name.
Differential revision: https://reviews.llvm.org/D33637
llvm-svn: 304113
Summary:
Previously if we parsed a constructor then we set parsed_ctor_dtor_cv
to true and never reseted it. This causes issue when a template argument
references a constructor (e.g. type of lambda defined inside a
constructor) as we will have the parsed_ctor_dtor_cv flag set what will
cause issues when parsing later arguments.
Reviewers: EricWF, compnerd
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D33385
llvm-svn: 303737
The problem is that multiple types could have been parsed from parse_type(),
which the lamdba parameter parsing didn't handle.
Differential revision: https://reviews.llvm.org/D33368
llvm-svn: 303718
r276215 made a change to annotate _Unwind_Exception with attribute
"aligned" so that an exception object following field __cxa_exception
is sufficiently aligned. This fix hasn't been incorporated to unwind.h
on Darwin since it is an ABI breaking change.
Instead of annotating struct _Unwind_Exception with the attribute, this
commit annotates field unwindHeader of __cxa_exception. This ensures the
exception object is sufficiently aligned without breaking the ABI.
This recommits r302978 and r302981, which were reverted in r303016
because a libcxx test was failing on an AArch64 bot. I also modified the
libcxxabi test case to check the alignment of the pointer returned by
__cxa_allocate_exception rather than compiling the test with -O1 and
checking whether it segfaults.
rdar://problem/25364625
Differential Revision: https://reviews.llvm.org/D33030
llvm-svn: 303175
For ARM EHABI, _Unwind_Exception is an alias of _Unwind_Control_Block,
which is not aligned, so we shouldn't align unwindHeader either.
rdar://problem/25364625
llvm-svn: 302981
r276215 made a change to annotate _Unwind_Exception with attribute
"aligned" so that an exception object following field __cxa_exception
is sufficiently aligned. This fix hasn't been incorporated to unwind.h
on Darwin since it is an ABI breaking change.
Instead of annotating struct _Unwind_Exception with the attribute, this
commit annotates field unwindHeader of __cxa_exception. This ensures the
exception object is sufficiently aligned without breaking the ABI.
This recommits r302763 with fixes to RUN lines in the test case.
rdar://problem/25364625
Differential Revision: https://reviews.llvm.org/D33030
llvm-svn: 302978
r276215 made a change to annotate _Unwind_Exception with attribute
"aligned" so that an exception object following field __cxa_exception
is sufficiently aligned. This fix hasn't been incorporated to unwind.h
on Darwin since it is an ABI breaking change.
Instead of annotating struct _Unwind_Exception with the attribute, this
commit annotates field unwindHeader of __cxa_exception. This ensures the
exception object is sufficiently aligned without breaking the ABI.
rdar://problem/25364625
Differential Revision: https://reviews.llvm.org/D33030
llvm-svn: 302763
libc++abi can't depend on libc++, so disable extern templates in libc++
headers project-wide. This was previously done in cxa_demangle.cpp, but
I consider it more appropriate to do at the cmake level (since none of
libc++abi's source files can use libc++ extern templates).
I also think the _LIBCPP_NO_EXCEPTIONS in cxa_demangle.cpp is
suspicious, but I'm doing one thing at a time.
Differential Revision: https://reviews.llvm.org/D32329
llvm-svn: 302739
CMake has the problem with the single dash variant because of the
space, so use the double dash with equal sign version. These flag
need to be included in compile flags to propagate correctly. We also
don't have to pass the target triple when checking for compiler-rt
since that flag is already included in compile flags now.
Differential Revision: https://reviews.llvm.org/D32069
llvm-svn: 300418
Since libc++abi is built with -nodefaultlibs, we should be using this
option even for CMake checks to avoid any inconsistency and also to
avoid dependency on a working C++ standard library just for the setting
up the build itself. The implementation is largely similar to the one
used by libc++.
Differential Revision: https://reviews.llvm.org/D31639
llvm-svn: 299797
This test fails on ARM bare-metal targets because it assumes the Itanium ABI,
whereas EHABI requires the exception address to be 8-byte aligned.
I was a bit puzzled at first because this should've failed on the public
arm-linux builder too. I think the reason it passes there is because we don't
include libunwind headers in the include path when running the libcxxabi tests,
so the system unwind.h gets picked up.
Reviewers: rengolin, EricWF
Differential revision: https://reviews.llvm.org/D31178
llvm-svn: 299435
Summary: It's now completely empty, so we can remove it entirely.
Reviewers: mclow.lists, EricWF
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D31502
llvm-svn: 299129
This was originally there for the _POSIX_THREADS define, to detect the
presence of pthreads. That went away with the externalized threading
support, so the include can go away too.
config.h is now completely empty. A follow-up commit will remove it
entirely.
llvm-svn: 299087
This patch fully reformats fallback_malloc.cpp. Previously the test
was a mess of different styles and indentations. This made it very
hard to work in and read. Therefore I felt it was best to re-format
the whole thing.
Unfortuantly this means some history will be lost, but hopefully
much of it will still be accessible after ignoring whitespace changes.
llvm-svn: 296960
Summary:
In 32 bit builds on a 64 bit system `std::malloc` does not return correctly aligned memory. This leads to undefined behavior.
This patch switches to using `posix_memalign` to allocate correctly aligned memory instead.
Reviewers: mclow.lists, danalbert, jroelofs, compnerd
Reviewed By: compnerd
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D25417
llvm-svn: 296952
Summary:
Currently both libc++ and libc++abi provide definitions for operator new/delete. However I believe this is incorrect and that one or the other should offer them.
This patch adds the CMake option `-DLIBCXXABI_ENABLE_NEW_DELETE_DEFINITIONS` which defaults to `OFF` unless otherwise specified. This means that by default
only libc++ provides the new/delete definitions.
Reviewers: mclow.lists, mehdi_amini, dexonsmith, beanz, jroelofs, danalbert, smeenai, rmaprath, mgorny
Reviewed By: mehdi_amini
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D30517
llvm-svn: 296801
Summary:
Currently both libc++ and libc++abi provide definitions for new/delete. However libc++abi's definitions haven't been updated to include aligned new/delete or sized deallocation.
I don't see any reason why libc++abi shouldn't provide these newer overloads.
This patch copies libc++'s implementation of `new/delete` into libc++abi so that it's now up to date.
After applying this patch I plan to fix a longstanding bug where both libc++ and libc++abi provide definitions for new/delete.
Reviewers: mclow.lists, mehdi_amini, dexonsmith, danalbert, smeenai, rmaprath, jroelofs
Reviewed By: mehdi_amini
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D30514
llvm-svn: 296787
This patch cleans up how libc++abi handles the definitions for new/delete.
It is in preperation for upcoming changes to fix how both libc++ and libc++abi
handle new/delete.
The primary changes in this patch are:
* Move the definitions for bad_array_length and bad_new_array_length
into stdlib_exception.cpp. This way stdlib_new_delete.cpp only
contains new/delete.
* Rename cxa_new_delete.cpp -> stdlib_new_delete.cpp for consistency
with other files.
* Add a FIXME regarding when stdlib_new_delete.cpp is actually compiled
as part of the dylib.
llvm-svn: 296715
Use the libc++abi visibility macros instead of pragmas or using
visibility attributes directly. Clean up redundant attributes on
definitions (where the declarations already have visibility attributes
applied, from either libc++ or libc++abi headers).
Introduce _LIBCXXABI_WEAK as a drive-by cleanup, which matches the
semantics of _LIBCPP_WEAK.
No functional change. Tested by building on Linux before and after this
change and verifying that the list of exported symbols is identical.
Differential Revision: https://reviews.llvm.org/D26949
llvm-svn: 296576
These tests embed calls to exceptions-related symbols from the abi library,
which are absent in the no-exceptions variant. The tests need to be marked
as unsupported for the no-exceptions configuration.
llvm-svn: 296344
Made a mistake in the condition typo because LIBCXXABI_BAREMETAL is always
defined, I should have been checking the contents to see if it's enabled.
Differential Revision: https://reviews.llvm.org/D30343
llvm-svn: 296146
We've been having issues with using libcxxabi and libunwind for baremetal
targets because fprintf is dependent on io functions, this patch disables calls
to fprintf when building for baremetal in release mode.
Differential Revision: https://reviews.llvm.org/D30339
llvm-svn: 296136
The libcxx/test/libcxx Python package has been moved into
libcxx/utils/libcxx in r294651, but libcxxabi's CMakeLists.txt still
looks for the old path.
Differential Revision: https://reviews.llvm.org/D30133
llvm-svn: 295540
When libcxxabi is built in LIBCXXABI_SILENT_TERMINATE mode, libcxx test suite reports
two failures:
std/depr/exception.unexpected/set.unexpected/get_unexpected.pass.cpp
std/depr/exception.unexpected/set.unexpected/set_unexpected.pass.cpp
This is because the default unexpected handler is set to std::abort instead of
std::terminate which these tests expect.
llvm-svn: 295175
When building as part of runtimes, there is no predefined order in
which the runtimes are loaded, so the targets from other projects
might not be available. We need to rely on HAVE_<name> variables
instead in that case.
Differential Revision: https://reviews.llvm.org/D29574
llvm-svn: 294552
When checking if the type is a r-value ref, we would not do a complete
check. This would result in us treating a trailing parameter reference
`&)` as a r-value ref, and improperly inject the cv qualifier on the
type. We now correctly demangle the type `KFvRmE` as a constant
function rather than a constant reference.
Fixes PR31741!
llvm-svn: 292973
Fix an off-by-one case which would destroy the final parameter in a
CV-qualified function type with a reference. We still get the CV
qualification incorrect, but at least we do not clobber the type name
any longer.
Partially fixes PR31741.
llvm-svn: 292963
Pending LIT changes are about to remove the REQUIRES-ANY keyword
in place of supporting boolean && and || within "REQUIRES". This
patch prepares libc++ for that change so that when applied
the bots don't lose their mind.
llvm-svn: 292906
This test contained an implicit conversion from nullptr to bool.
Clang warns about this but the test had supressed that warning.
However GCC diagnoses the same code as an error and requires
-fpermissive to accept it.
This patch fixes both the warning and the error by explicitly
converting the pointer to bool.
llvm-svn: 292638
The Itanium ABI [1] specifies that __cxa_demangle accept either:
1) symbol names, which start with "_Z"
2) type manglings, which do not start with "_Z"
r286788 erroneously assumes that it should only handle symbols, so this patch
reverts it and adds a counterexample to the testcase.
1: https://mentorembedded.github.io/cxx-abi/abi.html#demangler
Reviewers: zygoloid, EricWF
llvm-svn: 292418
Add LIBUNWIND_* directories to include path only if they were actually
found, in order to fix the CMake error. Both of the directories are
usually unnecessary since libcxxabi uses only the common part of
unwind.h that is supplied both by GCC and Clang.
Differential Revision: https://reviews.llvm.org/D25314
llvm-svn: 292018
This patch adjusts the out-of-tree CMake configuration so that
the stderr output is ignored when an old llvm-config is found
that doesn't support --cmakedir.
llvm-svn: 291993
The current std::terminate_handler pulls in some string code, some I/O
code, and more. Since it is automatically setup as the default, this
means that any trivial binary linking against libcxxabi will get this
extra bloat.
This patch allows disabling it as a build-time option, if you want to
avoid the extra bloat.
Patch by Tom Rybka!
Reviewers: EricWF
Subscribers: danalbert, llvm-commits, mgorny
Differential Revision: https://reviews.llvm.org/D28497
llvm-svn: 291946
Use the new --cmakedir option to obtain LLVM_CMAKE_PATH straight from
llvm-config. Fallback to local reconstruction if llvm-config does not
support this option.
llvm-svn: 291506
+ Now that libcxxabi shares the same threading API as libcxx, a whole
chunk of code in src/config.h is made redundant (I missed this earlier).
+ r291275 split off the externalized-thread-api libcxx configuration from the
external-thread-library libcxx configuration. libcxxabi should follow the
same approach.
llvm-svn: 291440
Tests targets will now be enabled by default when building libcxxabi out of tree
(unless turned off with LIBCXXABI_INCLUDE_TESTS=OFF).
Differential Revision: https://reviews.llvm.org/D28450
llvm-svn: 291378
Use an `extern "C" { }` block around the definition rather than doing
the inline definition. This avoids a GCC warning about a declaration
being extern and having a definition. NFC.
llvm-svn: 290937
r281179 Introduced an externally threaded variant of the libc++ library. This
patch adds support for a similar library variant for libc++abi.
Differential revision: https://reviews.llvm.org/D27575
Reviewers: EricWF
llvm-svn: 290888
The macOS thread-local variable finalizer routines do not handle the
case where a termination function registers another termination function
correctly, causing this test to fail. I've filed a radar for this;
mark the test XFAIL in the meantime. See [1] for more details.
[1] http://lists.llvm.org/pipermail/cfe-dev/2016-November/051376.html
Differential Revision: https://reviews.llvm.org/D27434
llvm-svn: 289513
It's useful to be able to disable visibility annotations entirely; for
example, if we're building libc++abi static to include in another library,
and we don't want any libc++abi functions getting exported out of that
library. This is a generalization of _LIBCXXABI_DISABLE_DLL_IMPORT_EXPORT.
Differential Revision: https://reviews.llvm.org/D26950
llvm-svn: 288692
Sink the Db initialization into the structure rather than out-of-line at the
declaration size. This just makes it easier to see what initialization is being
performed. NFC.
llvm-svn: 287364
The block invocation function uses an extension where the prefix is ___Z
as opposed to _Z. This should make the tests pass again.
Disable a negative test which was testing a crasher. The symbol being
demangled is not a valid mangled symbol and will return a nullptr.
Adjust the type info decoding test to be a valid symbol name.
llvm-svn: 286793
Ensure that we have a mangled symbol before attempting to demangle it. We would
previously treat any input as a mangled symbol rather than checking that the
symbol has the initial C++ Itanium v3 mangling prefix of `_Z`. This changes the
behaviour from the previous case which would undecorate `f` to `float` rather
than nullptr as it should.
Unfortunately, we do not have any negative testing for the demangler.
llvm-svn: 286788
GCC 4.7 or newer emits 0x90 (indirect | pcrel) as the ttype encoding.
This would hit an assertion in cxa_personality.cpp. This commit fixes
the problem by relaxing the assertion.
llvm-svn: 286760
This commit fixes libc++abi build when LLVM unwinder (libunwind_llvm) is
not enabled.
This commit fixes the problem by removing "LLVM_NATIVE_ARCH MATCHES ARM"
from CMakeLists.txt so that LIBCXXABI_USE_LLVM_UNWINDER will only be
defined when LLVM unwinder is enabled.
We need LIBCXXABI_USE_LLVM_UNWINDER becase there is a subtle difference
between the unwinder from libgcc and the one from libunwind_llvm. For
the unwinder from libgcc, we have to initialize register r12 with the
address of _Unwind_Control_Block; otherwise,
_Unwind_GetLanguageSpecificData() and _Unwind_GetRegionStart() won't
work properly. Consequently, there is an extra _Unwind_SetGR() when
LLVM unwinder is disabled. Check cxa_personality.cpp for details.
llvm-svn: 286759
The runtimes subdir is the new location for runtimes, we should
include it when looking for libcxx and libunwind headers.
Differential Revision: https://reviews.llvm.org/D26362
llvm-svn: 286614
The runtimes subdir is the new location for runtimes, we should
include it when looking for libcxx and libunwind headers.
Differential Revision: https://reviews.llvm.org/D26362
llvm-svn: 286332
Summary:
Some changes are made to cmake, especially the addition of a new
LLVM_ENABLE_PROJECTS option that makes the build system aware of
the monorepo directory structure.
Also a new script is added in llvm/utils/git-svn/. When present in
the $PATH, it enables a `git llvm` command. It is providing at this
point only the ability to push from the git monorepo: `git llvm push`.
It is intended to evolves with more features, for instance I plan on
features like `git llvm show r284955` to help working with sequential
revision numbers.
The push feature is taken from Justin Lebar's script available here:
https://github.com/jlebar/llvm-repo-tools/
Reviewers: jlebar
Subscribers: mgorny, modocache, llvm-commits
Differential Revision: https://reviews.llvm.org/D26334
llvm-svn: 286123
This patch refactors all pthread uses of libc++abi into a separate API. This
is the first step towards supporting an externlly-threaded libc++abi library.
I've followed the conventions already used in the libc++ library for the same
purpose.
Patch from: Saleem Abdulrasool and Asiri Rathnayake
Reviewed by: compnerd, EricWF
Differential revisions:
https://reviews.llvm.org/D18482 (original)
https://reviews.llvm.org/D24864 (final)
llvm-svn: 284128
Throwing an exception for the first time may lead to call calloc to
allocate memory for __cxa_eh_globals. If the memory pool is exhausted
at that moment, it results in abnormal termination of the program.
This patch addresses the issue by using fallback_malloc in that case.
In this revision, some restrictions were added into the test to not
run it in unsuitable environments.
Differential Revision: https://reviews.llvm.org/D17815
llvm-svn: 283531
The cmake files install directory has been changed to
${prefix}/lib/cmake/llvm since r259821. Searching cmake modules in
${prefix}/share/llvm/cmake will result in fatal errors.
This commit fixes the out-of-tree build by changing the CMake module
search path to: "$(llvm-config --obj-root)/lib/cmake/llvm"
llvm-svn: 283100
Throwing an exception for the first time may lead to call calloc to
allocate memory for __cxa_eh_globals. If the memory pool is exhausted
at that moment, it results in abnormal termination of the program.
This patch addresses the issue by using fallback_malloc in that case.
Differential Revision: https://reviews.llvm.org/D17815
llvm-svn: 282692
`__declspec(dllexport)` and `__declspec(dllimport)` should only be used
when building libc++abi as a DLL, but that's the more common use case,
so default to adding the annotations and add an option to opt out.
Similar to r282449, which made the corresponding change for libc++.
Differential Revision: https://reviews.llvm.org/D24945
llvm-svn: 282470
These data and text symbols were missing annotations for building with hidden
visibility. As we do not currently enable hidden visibility by default, this is
a NFC for the buildbots.
llvm-svn: 282300
is_initialized is only used in the no threads case or if on non ARM Apple
targets. Use the preprocessor to remove the function otherwise. NFC.
llvm-svn: 280286
Summary: This patch adds a check for if -std=c++11 is a supported flag, and adds it to CMAKE_CXX_FLAGS if it is supported.
Reviewers: EricWF
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D24007
llvm-svn: 280021
This patch (and commit summary) mirror r279675 by Chris B which was applied
to libc++ but not libc++abi.
Summary:
The point of this patch is to have a consistent convention for naming build,
check and install targets so that the targets can be constructed from the
project name.
This change renames a bunch of CMake components and targets from libcxxabi to
cxxabi. For each renamed target I've added a convenience target that matches
the old target name and depends on the new target. This will preserve function
of the old targets so that the change doesn't break the world. We can evaluate
if it is worth removing the extra targets later.
llvm-svn: 279938
This patch applies changes similar to those in r279515 to libc++abi.
Summary of changes in this patch:
* Renamed variable LLVM_CONFIG -> LLVM_CONFIG_PATH
* Renamed variable LIBCXXABI_BUILT_STANDALONE -> LIBCXXABI_STANDALONE_BUILD
* Add an include of AddLLVM in the tests subdirectory for add_lit_testsuite.
llvm-svn: 279936
When we're running tests under ASAN or MSAN, they're compiled with -O1, which
enables tail call elimination. This causes backtrace_test to fail: the compiler
performs tail call elimination for call3_nothrow, but it can't for call3_throw,
leading to a mismatched frame count. Disable tail call elimination (and
inlining, just to be explicit) to avoid this.
Patch by Shoaib Meenai!
llvm-svn: 279935
When using libunwind and not building as standalone project, we
can directly depend on the unwind library target.
Differential Revision: https://reviews.llvm.org/D23289
llvm-svn: 278169
When using libunwind and not building as standalone project, we
need to add LLVM library directory to the list of linker directories
to ensure it can find libunwind dependency.
Differential Revision: https://reviews.llvm.org/D23287
llvm-svn: 278076
When libcxxabi is being built standalone, unwind dependency is not
available, so do not use it even when LLVM unwinder is being
requested.
Differential Revision: https://reviews.llvm.org/D23228
llvm-svn: 278058
Michal Gorny