Summary:
This is a follow up to r355253, which inadvertently broke Visual
Studio builds by trying to copy files from CMAKE_CFG_INTDIR.
See https://reviews.llvm.org/D58537#inline-532492
Reviewers: smeenai, vzakhari, phosek
Reviewed By: smeenai
Subscribers: mgorny, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D61054
llvm-svn: 359257
When the arguments to tuple cat were const, the const was incorrectly
propagated into the type of the resulting tuple. For example:
const std::tuple<int> t(42);
auto r = std::tuple_cat(t, t);
// Incorrect! should be std::tuple<int, int>.
static_assert(is_same_v<decltype(r), std::tuple<const int, const int>>);
llvm-svn: 359255
Summary:
Avoids an MSan false positive if the SIGINT comes while the user
callback is running. The false positive happens when the interrupt
handler calls opendir() to remove some temporary files, which is
intercepted by MSan.
Fixes https://github.com/google/oss-fuzz/issues/2332.
Reviewers: kcc
Reviewed By: kcc
Subscribers: llvm-commits, Dor1s, metzman
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D61163
llvm-svn: 359254
isValidCandidateForColdCC is much more expensive than
TTI.useColdCCForColdCall, which by default just returns false. Avoid
doing this work if we're not going to look at the answer anyway.
This change is NFC, but I see significant compile time improvements on
some code with pathologically many functions.
llvm-svn: 359253
When failing materialization of a symbol X, remove X from the dependants list
of any of X's dependencies. This ensures that when X's dependencies are
emitted (or fail themselves) they do not try to access the no-longer-existing
MaterializationInfo for X.
llvm-svn: 359252
This reverts r359250 (git commit 4730604bd3)
The newly added test should use -cc1 and -emit-llvm and there are other
test failures that need fixing.
llvm-svn: 359251
Statically link certain runtime library functions for MSVC/GNU Windows
environments. This is consistent with MSVC behavior.
Fixes LNK4286 and LNK4217 warnings from link.exe when linking the static
CRT:
LINK : warning LNK4286: symbol '__std_terminate' defined in 'libvcruntime.lib(ehhelpers.obj)' is imported by 'ASAN_NOINST_TEST_OBJECTS.asan_noinst_test.cc.x86_64-calls.o'
LINK : warning LNK4286: symbol '__std_terminate' defined in 'libvcruntime.lib(ehhelpers.obj)' is imported by 'ASAN_NOINST_TEST_OBJECTS.asan_test_main.cc.x86_64-calls.o'
LINK : warning LNK4217: symbol '_CxxThrowException' defined in 'libvcruntime.lib(throw.obj)' is imported by 'ASAN_NOINST_TEST_OBJECTS.gtest-all.cc.x86_64-calls.o' in function '"int `public: static class UnitTest::GetInstance * __cdecl testing::UnitTest::GetInstance(void)'::`1'::dtor$5" (?dtor$5@?0??GetInstance@UnitTest@testing@@SAPEAV12@XZ@4HA)'
Reviewers: mstorsjo, efriedma, TomTan, compnerd, smeenai, mgrang
Subscribers: abdulras, theraven, smeenai, pcc, mehdi_amini, javed.absar, inglorion, kristof.beyls, dexonsmith, cfe-commits
Differential Revision: https://reviews.llvm.org/D55229
llvm-svn: 359250
These builtins provide access to the new integer and
sub-integer variants of MMA (matrix multiply-accumulate) instructions
provided by CUDA-10.x on sm_75 (AKA Turing) GPUs.
Also added a feature for PTX 6.4. While Clang/LLVM does not generate
any PTX instructions that need it, we still need to pass it through to
ptxas in order to be able to compile code that uses the new 'mma'
instruction as inline assembly (e.g used by NVIDIA's CUTLASS library
https://github.com/NVIDIA/cutlass/blob/master/cutlass/arch/mma.h#L101)
Differential Revision: https://reviews.llvm.org/D60279
llvm-svn: 359248
All of the new instructions are still handled mostly by tablegen. I've slightly
refactored the code to drive intrinsic/instruction generation from a master
list of supported variants, so all irregularities have to be implemented in one place only.
The test generation script wmma.py has been refactored in a similar way.
Differential Revision: https://reviews.llvm.org/D60015
llvm-svn: 359247
PTX 6.3 requires using ".aligned" in the MMA instruction names.
In order to generate correct name, now we pass current
PTX version to each instruction as an extra constant operand
and InstPrinter adjusts its output accordingly.
Differential Revision: https://reviews.llvm.org/D59393
llvm-svn: 359246
Generalized constructions of 'fragments' of MMA operations to provide
common primitives for construction of the ops. This will make it easier
to add new variants of the instructions that operate on integer types.
Use nested foreach loops which makes it possible to better control
naming of the intrinsics.
This patch does not affect LLVM's output, so there are no test changes.
Differential Revision: https://reviews.llvm.org/D59389
llvm-svn: 359245
Adds a representation of the section header table to XCOFFObjectFile,
and implements enough to dump the section headers with llvm-obdump.
Differential Revision: https://reviews.llvm.org/D60784
llvm-svn: 359244
libc++ ABI v1 provides three valarray symbols as part of the shared library:
valarray<size_t>::valarray(size_t)
valarray<size_t>::~valarray()
valarray<size_t>::resize(size_t, size_t)
The first two of these are intended to be removed in V2 of the ABI: they're
attributed _LIBCPP_HIDE_FROM_ABI_AFTER_V1, and it appears that the intention
is that these symbols from the library are not used even when building using
the V1 ABI. However, there are explicit instantiation declarations for all
three symbols in the header, which are not correct as we do not intend to find
an instantiation of these functions that is provided elsewhere.
(A recent change to clang to properly diagnose explicit instantiation
declarations of internal linkage functions -- required by [temp.explicit]p13 --
had to be rolled back because it diagnosed these explicit instantiations.)
Remove the explicit instantiation declarations, and remove the explicit
instantiation definitions for V2 of the libc++ ABI onwards.
llvm-svn: 359243
Summary:
This value is derived from the host triple, which on the machine
I'm currently using is `ppc64le-linux-redhat`. This change makes
LLVM compile.
Reviewers: nemanjai
Differential Revision: https://reviews.llvm.org/D57118
llvm-svn: 359242
I added a diagnostic along the lines of `-Wpessimizing-move` to detect `return x = y` suppressing copy elision, but I don't know if the diagnostic is really worth it. Anyway, here are the places where my diagnostic reported that copy elision would have been possible if not for the assignment.
P1155R1 in the post-San-Diego WG21 (C++ committee) mailing discusses whether WG21 should fix this pitfall by just changing the core language to permit copy elision in cases like these.
(Kona update: The bulk of P1155 is proceeding to CWG review, but specifically *not* the parts that explored the notion of permitting copy-elision in these specific cases.)
Reviewed By: dblaikie
Author: Arthur O'Dwyer
Differential Revision: https://reviews.llvm.org/D54885
llvm-svn: 359236
(this would regress size without a corresponding LLVM change that avoids
putting other user defined types inside type units when they aren't in
their own type units - instead emitting declarations inside the type
unit and a definition in the primary CU)
Reviewers: aprantl
Differential Revision: https://reviews.llvm.org/D61079
llvm-svn: 359235
Path lists on windows should always be separated by semicolons, not
colons. Reuse llvm::sys::EnvPathSeparator for this purpose (as that's
also a path list that is separated in the same way).
Alternatively, this could just be a local ifdef _WIN32 in this function,
or generalizing the existing EnvPathSeparator to e.g. a
llvm::sys::path::PathListSeparator?
Differential Revision: https://reviews.llvm.org/D61121
llvm-svn: 359233
This case was missing before, so we couldn't legalize it.
Add it to AArch64LegalizerInfo.cpp and update select-extract-vector-elt.mir.
llvm-svn: 359231
https://bugs.llvm.org/show_bug.cgi?id=41590
For the following code snippet, UninitializedObjectChecker crashed:
struct MyAtomicInt {
_Atomic(int) x;
MyAtomicInt() {}
};
void entry() {
MyAtomicInt b;
}
The problem was that _Atomic types were not regular records, unions,
dereferencable or primitive, making the checker hit the llvm_unreachable at
lib/StaticAnalyzer/Checkers/UninitializedObject/UninitializedObjectChecker.cpp:347.
The solution is to regard these types as primitive as well. The test case shows
that with this addition, not only are we able to get rid of the crash, but we
can identify x as uninitialized.
Differential Revision: https://reviews.llvm.org/D61106
llvm-svn: 359230
std::mutex was not actually is_nothrow_default_constructible in C++98/C++03,
because the variable declaration
std::mutex M;
... could throw an exception from the mutex destructor. Fix it by marking the
destructor as non-throwing. This has no effect in C++11 onwards, because
destructors are non-throwing by default in those language modes.
llvm-svn: 359229
Use constexpr to explicitly indicate that we're dealing with integer
constants, and provoke clang to assign them straight to registers
whenever possible. Adjust input constraints in %mmN tests to "rm"
as using integer constants is apparently disallowed there. Also
use "i" for %rN tests, as we don't want clang to accidentally clobber
those general purpose registers while assigning to them (however
unlikely that is).
llvm-svn: 359228
Contrary to MSVC, MinGW compilers wants the dllexport attribute on
the declaration of an explicit template instantiation, not on the
definition.
Differential Revision: https://reviews.llvm.org/D61123
llvm-svn: 359227
it keeps track of becomes too large
ARC optimizer does a top-down and a bottom-up traversal of the whole
function to pair up retain and release instructions and remove them.
This can be expensive if the number of instructions in the function and
pointer states it tracks are large since it has to look at each pointer
state and determine whether the instruction being visited can
potentially use the pointer.
This patch adds a command line option that sets a limit to the number of
pointers it tracks.
rdar://problem/49477063
Differential Revision: https://reviews.llvm.org/D61100
llvm-svn: 359226
This adds a legalization rule for G_ZEXT, G_ANYEXT, and G_SEXT which allows
extends whenever the types will fit in registers (or the source is an s1).
Update tests. Add GISel checks throughout all of arm64-vabs.ll,
where we now select a good portion of the code. Add GISel checks to
arm64-subvector-extend.ll, which has a good number of vector extends in it.
Differential Revision: https://reviews.llvm.org/D60889
llvm-svn: 359222
All constant expressions are non-potentially-throwing in C++14, but that is *not* the case in C++17. Change these tests of the `variant`-flavored overloads of `std::get` to expect the correct behavior when the compiler is not GCC or is GCC 9+.
Credit to Jonathan Wakely for providing an improved version of my initial change that validates the incorrect behavior on GCC < 9 as well as validating the correct behavior on other compilers.
Differential Revision: https://reviews.llvm.org/D61033
llvm-svn: 359220
We had special case handling here, but it uses a scalar any_extend for the
promotion then bitcasts to the final type. This won't split up the input data
into multiple promoted elements like we need.
This patch falls back to doing the conversion through memory.
Fixes PR41594 which I believe was reflected in the bitcast-vector-bool.ll
changes. The changes to vector-half-conversions.ll are fixing a previously
unknown miscompile from this issue.
Differential Revision: https://reviews.llvm.org/D61114
llvm-svn: 359219
Summary:
Since neither compiler-rt nor the libc++ we build use exceptions, we
don't need libc++abi to have them either.
This resolves an issue where libFuzzer's private libc++ contains
implementations for __cxa_throw and friends, causing fuzz targets built
with their own C++ library to segfault during exception unwinding.
See https://github.com/google/oss-fuzz/issues/2328.
Reviewers: phosek, EricWF, kcc
Reviewed By: phosek
Subscribers: kcc, dberris, mgorny, christof, llvm-commits, metzman
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D61053
llvm-svn: 359218
Currently InstrProf lowering is not enabled for Clang PGO instrumentation in
the new pass manager. The following command
"-fprofile-instr-generate -fexperimental-new-pass-manager ..." is broken.
This CL enables InstrProf lowering pass for Clang PGO instrumentation in the
new pass manager.
Differential Revision: https://reviews.llvm.org/D61138
llvm-svn: 359215
Summary:
llvm::Regex is really slow, and regex evaluation during preamble indexing was
showing up as 25% on a profile of clangd in a codebase with large preambles.
Reviewers: ilya-biryukov
Subscribers: MaskRay, jkorous, arphaman, kadircet, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D61120
llvm-svn: 359214
Add tests covering read operations for the general-purpose and XMM
registers added in x86_64 (r8-r15 and xmm8-xmm15).
Differential Revision: https://reviews.llvm.org/D61072
llvm-svn: 359210
Remove the use of 2-element array for XMM data. It is an accidental
leftover from previous implementation attempt, and it is unnecessary
with xmm_t.
Differential Revision: https://reviews.llvm.org/D61085
llvm-svn: 359208
Tom Stellard