Changes

[clang][codegen] Skip adding default function attributes on intrinsics.

- After loading builtin bitcode for linking, skip adding default
  function attributes on LLVM intrinsics as their attributes are
  well-defined and retrieved directly from internal definitions. Adding
  extra attributes on intrinsics results in inconsistent result when
  `-save-temps` is present. Also, that makes few optimizations
  conservative.

Differential Revision: https://reviews.llvm.org/D87761

[AArch64][GlobalISel] Make G_BUILD_VECTOR os <16 x s8> legal.

[gn build] make "all" target build

If you want to build everything, building the default target
via just `ninja` is better, but `ninja all` shouldn't give you
compile errors -- this fixes that.

GlobalISel: Lift store value widening restriction

This doesn't change the memory size and doesn't need to worry about
non-power-of-2 sizes.

[obj2yaml] - Match ".stack_size" with the original section name, and not the uniquified name.

Without this patch, obj2yaml decodes the content of only one ".stack_size" section. Other sections are dumped with their full contents.

Reviewed By: grimar, MaskRay

Differential Revision: https://reviews.llvm.org/D87727

Add '<' meta command to read in code from external file

Perform all error handling in ReadCode()

Add :help text describing “< path”, add extra line before Commands

Differential Revision: https://reviews.llvm.org/D87640

[UpdateTestChecks][NFC] Fix spelling

[OpenMP][FIX] Do not crash trying to print a missing (demangled) user condition

Reviewed By: JonChesterfield

Differential Revision: https://reviews.llvm.org/D85875

[OpenMP][FIX] Do not drop a '$' while demangling declare variant names

Reviewed By: ABataev

Differential Revision: https://reviews.llvm.org/D85876

[OpenMP] Support nested OpenMP context selectors (declare variant)

Due to `omp begin/end declare variant`, OpenMP context selectors can be
nested. This patch adds initial support for this so we can use it for
target math variants. We should improve the detection of "equivalent"
scores and user conditions, we should also revisit the data structures
of the OMPTraitInfo object, however, both are not pressing issues right
now.

Reviewed By: JonChesterfield

Differential Revision: https://reviews.llvm.org/D85877

[OpenMP] Context selector extensions for return value overloading

This extension allows to declare variants in between `omp begin/end
declare variant` that do not match the type of the existing function
with that name. Without this extension we would not find a base function
(with a compatible type), therefore create a new one, which would
cause conflicting declarations. With this extension we will not create
"missing" base functions, which basically renders these specializations
harmless. They will be generated but never called.

Reviewed By: JonChesterfield

Differential Revision: https://reviews.llvm.org/D85878

[OpenMP] Overload `std::isnan` and friends multiple times for the GPU

`std::isnan` and friends can be found in two variants in the wild, one
returns `bool`, as the standard defines it, one returns `int`, as the C
macros do. So far we kinda hoped the system versions of these functions
will work for people, e.g. they are definitions that can be compiled for
the target. We know that is not the case always so we leverage the
`disable_implicit_base` OpenMP context extension to specialize both
versions of these functions without causing an invalid redeclaration.

Reviewed By: JonChesterfield, tra

Differential Revision: https://reviews.llvm.org/D85879

[OpenMP] Context selector extensions for template functions

With this extension the effects of `omp begin declare variant` will be
applied to template function declarations. The behavior is opt-in and
controlled by the `extension(allow_templates)` trait. While generally
useful, this will enable us to implement complex math function calls by
overloading the templates of the standard library with the ones in
libc++.

Reviewed By: JonChesterfield

Differential Revision: https://reviews.llvm.org/D85735

[OpenMP] Support `std::complex` math functions in target regions

The last (big) missing piece to get "math" working in OpenMP target
regions (that I know of) was complex math functions, e.g.,
`std::sin(std::complex<double>)`. With this patch we overload the system
template functions for these operations with versions that have been
distilled from `libcxx/include/complex`. We use the same
  `omp begin/end declare variant`
mechanism we use for other math functions before, except that we this
time overload templates (via D85735).

Reviewed By: JonChesterfield

Differential Revision: https://reviews.llvm.org/D85777

[AMDGPU] gfx1030 RT support

Differential Revision: https://reviews.llvm.org/D87782

Revert "[obj2yaml] - Match ".stack_size" with the original section name, and not the uniquified name."

This reverts commit 14e55f82980cf1342d4d3eea4885a5375e829496.

[SystemZ][z/OS] Set aligned allocation unavailable by default for z/OS

Aligned allocation is not supported on z/OS. This patch sets -faligned-alloc-unavailable as default in z/OS toolchain.

Reviewed By: abhina.sreeskantharajan, hubert.reinterpretcast

Differential Revision: https://reviews.llvm.org/D87611

[gn build] unconfuse sync script about "sources = []" in clang/lib/Headers/BUILD.gn

Fix build.

[InstCombine] Add test for infinite combine loop (NFC)

Test courtesy of bkramer for the infinite combine loop introduced
by D87480.

[InstSimplify] Clarify SimplifyWithOpReplaced() return value

If SimplifyWithOpReplaced() cannot simplify the value, null should
be returned. Make sure this really does happen in all cases,
including those where SimplifyBinOp() returns the original value.

This does not matter for existing users, but does mattter for
D87480, which would go into an infinite loop otherwise.

Reapply [InstCombine] Simplify select operand based on equality condition

Reapply after fixing SimplifyWithOpReplaced() to never return
the original value, which would lead to an infinite loop in this
transform.

-----

For selects of the type X == Y ? A : B, check if we can simplify A
by using the X == Y equality and replace the operand if that's
possible. We already try to do this in InstSimplify, but will only
fold if the result of the simplification is the same as B, in which
case the select can be dropped entirely. Here the select will be
retained, just one operand simplified.

As we are performing an actual replacement here, we don't have
problems with refinement / poison values.

Differential Revision: https://reviews.llvm.org/D87480

[gn build] Port 56069b5c71c

[UpdateTestChecks] Allow $ in function names

Some compilers generation functions with '$' in their names, so recognize those
functions.

This also requires recognizing function names inside quotes in some contexts in
order to escape certain characters.

Differential Revision: https://reviews.llvm.org/D82995

[aarch64][tests] Add tests which show current lack of implicit null support

I will be posting a patch which adds appropriate target support shortly; landing the tests so that the diffs are clear.

 [flang][msvc] Work around if constexpr (false) evaluation. NFC.

MSVC tries to expand templates that are in the false-branch of a `if constexpr` construct. In this case, the condition checks whether a tuple has at least one element and then is trying to access it using `std::get<0>`, which fails when the tuple has 0 elements.

The workaround is to extract that case into a separate method.

This patch is part of the series to make flang compilable with MS Visual Studio <http://lists.llvm.org/pipermail/flang-dev/2020-July/000448.html>.

Reviewed By: klausler

Differential Revision: https://reviews.llvm.org/D87728

[X86] Add test case for a masked load mask becoming all ones after type legalization.

We should be able to turn this into a unmasked load. X86 has an
optimization to detect that the first and last element aren't masked
and then turn the whole thing into an unmasked load and a blend.
That transform is disabled on avx512 though.

But if we know the blend isn't needed, then the unmasked load by
itself should always be profitable.

[DAGCombiner] Teach visitMLOAD to replace an all ones mask with an unmasked load

If we have an all ones mask, we can just a regular masked load. InstCombine already gets this in IR. But the all ones mask can appear after type legalization.

Only avx512 test cases are affected because X86 backend already looks for element 0 and the last element being 1. It replaces this with an unmasked load and blend. The all ones mask is a special case of that where the blend will be removed. That transform is only enabled on avx2 targets. I believe that's because a non-zero passthru on avx2 already requires a separate blend so its more profitable to handle mixed constant masks.

This patch adds a dedicated all ones handling to the target independent DAG combiner. I've skipped extending, expanding, and index loads for now. X86 doesn't use index so I don't know much about it. Extending made me nervous because I wasn't sure I could trust the memory VT had the right element count due to some weirdness in vector splitting. For expanding I wasn't sure if we needed different undef handling.

Differential Revision: https://reviews.llvm.org/D87788

[lsan] Share platform allocator settings between ASan and LSan

This moves the platform-specific parameter logic from asan into
sanitizer_common so lsan can reuse it.

Patch By: mcgrathr

Differential Revision: https://reviews.llvm.org/D85930

Revert "[lsan] Share platform allocator settings between ASan and LSan"

This reverts commit c57df3dc09e8b59c55c83ba5c354569a82a5c3b8 which broke
Windows sanitizer bots.

[ORC] Add operations to create and lookup JITDylibs to OrcV2 C bindings.

Revert "Do not apply calling conventions to MSVC entry points"

This reverts commit 4cff1b40dacf6a5489b09657d94ea4757b8cd3b0.

Caused "undefined symbol: _WinMain@16" link errors.

[AMDGPU] gfx1030 test update. NFC.

[NFC][LSAN] Change SuspendedThreadsList interface

Remove RegisterCount and let GetRegistersAndSP to resize buffer as needed.

Reviewed By: morehouse

Differential Revision: https://reviews.llvm.org/D87747

[ELF] Bump the limit of thunk creation passes from 10 to 15

I have noticed that a 374MiB powerpc64le 'ld.lld' requires 11 passes to link.
There is a ThunkSection (whose parent OutputSection is ".text" of 169MiB) with 12867 thunks.

[NFC][regalloc] type LiveInterval::reg() as Register

We have the Register type which precisely captures the role of this
member. Storage-wise, it's an unsigned.

This helps readability & maintainability.

Differential Revision: https://reviews.llvm.org/D87768

[obj2yaml] - Match ".stack_size" with the original section name, and not the uniquified name.

Without this patch, obj2yaml decodes the content of only one ".stack_size" section. Other sections are dumped with their full contents.

Reviewed By: grimar, MaskRay

Differential Revision: https://reviews.llvm.org/D87727

[lldb/test] Enable faulthandler in dotest

Register the `faulthandler` module so we can see what lldb tests are doing when they misbehave (e.g. run under a test runner that sets a timeout). This will print a stack trace for the following signals:

- `SIGSEGV`, `SIGFPE`, `SIGABRT`, `SIGBUS`, and `SIGILL` (via `faulthandler.enable()`)
- `SIGTERM` (via `faulthandler.register(SIGTERM)`) [This is what our test runners sends when it times out].

The only signal we currently handle is `SIGINT` (via `unittest2.signals.installHandler()`) so there should be no overlap added by this patch.

Because this import is not available until python3, and the `register()` method is not available on Windows, this is enabled defensively.

This should have absolutely no effect when tests are passing (or even normally failing), but can be observed by running this while ninja is running:

```
kill -s SIGTERM $(ps aux | grep dotest.py | head -1 | awk '{print $2}')
```

Reviewed By: JDevlieghere

Differential Revision: https://reviews.llvm.org/D87637

[NFC] Refactor DiagnosticBuilder and PartialDiagnostic

PartialDiagnostic misses some functions compared to DiagnosticBuilder.

This patch refactors DiagnosticBuilder and PartialDiagnostic, extracts
the common functionality so that the streaming << operators are
shared.

Differential Revision: https://reviews.llvm.org/D84362

[libunwind] Support for leaf function unwinding.

Unwinding leaf function is useful in cases when the backtrace finds a
leaf function for example when it caused a signal.
This patch also add the support for the DW_CFA_undefined because it marks
the end of the frames.

Ryan Prichard provided code for the tests.

Reviewed By: #libunwind, mstorsjo

Differential Revision: https://reviews.llvm.org/D83573

[flang] Substrings with lower bound greater than upper bound

According to section 9.4.1, paragraph 3,
 If the starting point is greater than the ending point, the substring has
 length zero

But the compilers code for substring processing was failing a call to `CHECK()`
in this case.  I fixed this by just setting the number of items in the
resulting string to 0 for this situation.

Differential Revision: https://reviews.llvm.org/D87799

Disable a large test for EXPENSIVE_CHECKS and debug build

Summary:
When running a large test in LLVM_ENABLE_EXPENSIVE_CHECKS=ON mode,
buildbot could hit timeout.
Disable the test when this mode is on.
Also disable it for debug so that the test won't hang for too long.

Reviewed By: hubert.reinterpretcast

Differential Revision: https://reviews.llvm.org/D87794

[AArch64] Add -mmark-bti-property flag.

Writing the .note.gnu.property manually is error prone and hard to
maintain in the assembly files.
The -mmark-bti-property is for the assembler to emit the section with the
GNU_PROPERTY_AARCH64_FEATURE_1_BTI. To be used when C/C++ is compiled
with -mbranch-protection=bti.

This patch refactors the .note.gnu.property handling.

Reviewed By: chill, nickdesaulniers

Differential Revision: https://reviews.llvm.org/D81930

[gn build] (manually) port 1321160a2

[EarlyCSE] Simplify max/min pattern matching. NFC.

[Flang] Fixed installation permission of the "binary" flang

Under current configuration, the permission of `flang` after installation is 700.
This could bring a problem for system administrators who build and install flang
for other users, which only the user who builds LLVM can execute it, and others
can not. In this patch, the explicit permission setting in the `install` command
is removed, and let CMake determine what perssion to be used like other components.

Reviewed By: DavidTruby

Differential Revision: https://reviews.llvm.org/D87783

Commenting out atomics with padding to unbreak MSAN tests

Revert "[AArch64] Add -mmark-bti-property flag."

This reverts commit 95e43f84b7b9c61011aece7583c0367297dd67d8.

[AArch64] Add -mmark-bti-property flag.

Writing the .note.gnu.property manually is error prone and hard to
maintain in the assembly files.
The -mmark-bti-property is for the assembler to emit the section with the
GNU_PROPERTY_AARCH64_FEATURE_1_BTI. To be used when C/C++ is compiled
with -mbranch-protection=bti.

This patch refactors the .note.gnu.property handling.

Reviewed By: chill, nickdesaulniers

Differential Revision: https://reviews.llvm.org/D81930

Reland with test dependency on aarch64 target.

[DAGCombiner] Teach visitMSTORE to replace an all ones mask with an unmasked store.

Similar to what done in D87788 for MLOAD.

Again I've skipped indexed, truncating, and compressing stores.

[MemorySSA] Rename uses in blocks with Phis.

Renaming should include blocks with existing Phis.

Resolves PR45927.

Differential Revision: https://reviews.llvm.org/D87661

Canonicalize declaration pointers when forming APValues.

References to different declarations of the same entity aren't different
values, so shouldn't have different representations.

Recommit of e6393ee813178e9d3306b8e3c6949a4f32f8a2cb with fixed
handling for weak declarations. We now look for attributes on the most
recent declaration when determining whether a declaration is weak.

PR47555: Inheriting constructors are implicitly definable.

Don't forget to define them if they're constexpr and used inside a
template; we might try to evaluate a call to them before the template is
instantiated.

[NewPM] Port -print-alias-sets to NPM

Really it should be named print<alias-sets>, but for the sake of
changing fewer tests, added a TODO to rename after NPM switch and test
cleanup.

Reviewed By: ychen

Differential Revision: https://reviews.llvm.org/D87713

[IRSim] Adding IR Instruction Mapper

This introduces the IRInstructionMapper, and the associated wrapper for
instructions, IRInstructionData, that maps IR level Instructions to
unsigned integers.

Mapping is done mainly by using the "isSameOperationAs" comparison
between two instructions.  If they return true, the opcode, result type,
and operand types of the instruction are used to hash the instruction
with an unsigned integer.  The mapper accepts instruction ranges, and
adds each resulting integer to a list, and each wrapped instruction to
a separate list.

At present, branches, phi nodes are not mapping and exception handling
is illegal.  Debug instructions are not considered.

The different mapping schemes are tested in
unittests/Analysis/IRSimilarityIdentifierTest.cpp

Differential Revision: https://reviews.llvm.org/D86968

[mlir] expose affine map to C API

This patch provides C API for MLIR affine map.
- Implement C API for AffineMap class.
- Add Utils.h to include/mlir/CAPI/, and move the definition of the CallbackOstream to Utils.h to make sure mlirAffineMapPrint work correct.
- Add TODO for exposing the C API related to AffineExpr and mutable affine map.

Differential Revision: https://reviews.llvm.org/D87617

[gn build] Port b04c1a9d312

[libunwind][DWARF] Fix end of .eh_frame calculation

 * When .eh_frame is located using .eh_frame_hdr (PT_GNU_EH_FRAME), the
   start of .eh_frame is known, but not the size. In this case, the
   unwinder must rely on a terminator present at the end of .eh_frame.
   Set dwarf_section_length to UINTPTR_MAX to indicate this.

 * Add a new field, text_segment_length, that the FrameHeaderCache uses
   to track the size of the PT_LOAD segment indicated by dso_base.

 * Compute ehSectionEnd by adding sectionLength to ehSectionStart,
   never to fdeHint.

Fixes PR46829.

Differential Revision: https://reviews.llvm.org/D87750

[MachineSink] add one more mir case - nfc

[PowerPC] Fix store-fptoi combine of f128 on Power8

llc would crash for (store (fptosi-f128-i32)) when -mcpu=pwr8, we should
not generate FP_TO_(S|U)INT_IN_VSR for f128 types at this time. This
patch fixes it.

Reviewed By: steven.zhang

Differential Revision: https://reviews.llvm.org/D86686