Create a new test for lldb launched without a script interpreter and
move it under a new `ScriptInterpreter` directory. Also move
crashlog.test there for consistency.

Summary:
The vector pattern `(a + b + 1) / 2` was previously selected to an
avgr_u instruction regardless of nuw flags, but this is incorrect in
the case where either addition may have an unsigned wrap. This CL
changes the existing pattern to require both adds to have nuw flags
and adds builtin functions and intrinsics for the avgr_u instructions
because the corrected pattern is not representable in C.

Reviewers: aheejin

Subscribers: dschuff, sbc100, jgravelle-google, hiraditya, sunfish, cfe-commits, llvm-commits

Tags: #clang, #llvm

Differential Revision: https://reviews.llvm.org/D71648

This reverts commit 298e183e.

This commit caused some build failures -- reverting while I investigate.

These changes caused LibcxxVariantDataFormatterTestCase in lldb to fail
with an assert:

Assertion failed: (Idx < size() && "Out-of-bounds Bit access."),
function operator[], file
/Users/buildslave/jenkins/workspace/lldb-cmake/llvm-project/llvm/include/llvm/ADT/SmallBitVector.h,
line 452.

In:

7  clang-10                 0x00000001094b79d9 isAtLeastAsSpecializedAs(clang::Sema&, clang::SourceLocation, clang::FunctionTemplateDecl*, clang::FunctionTemplateDecl*, clang::TemplatePartialOrderingContext, unsigned int) + 1865
8  clang-10                 0x00000001094b7111 clang::Sema::getMoreSpecializedTemplate(clang::FunctionTemplateDecl*, clang::FunctionTemplateDecl*, clang::SourceLocation, clang::TemplatePartialOrderingContext, unsigned int, unsigned int) + 97
9  clang-10                 0x000000010939bf88 clang::isBetterOverloadCandidate(clang::Sema&, clang::OverloadCandidate const&, clang::OverloadCandidate const&, clang::SourceLocation, clang::OverloadCandidateSet::CandidateSetKind) + 1128

Revert "[Concepts] Fix incorrect move out of temporary in D41910"

This reverts commit 11d5fa6e.

Revert "[Concepts] Fix crash in D41910"

This reverts commit 12038be2.

Revert "[Concepts] Constrained partial specializations and function overloads."

This reverts commit fc0731b9.

[X86] Add a simple hack to IsProfitableToFold to prevent vselect+strict fp operations from being folded into masked instructions.

We really need to update the isel patterns to prevent this, but
that requires some tablegen de-tangling. So this hack will work
for correctness in the short term.

The masked operation will suppress exceptions on masked out lanes.
But that's not what vselect+strict_fadd means.

Reviewers: aaron.ballman

Subscribers: mgorny, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D70613

When clang is invoked with a source file without -c or -S, it creates a
cc1 job, a linker job and if debug info is requested, a dsymutil job. In
case of remarks, we should also create a dsymutil job to avoid losing
the remarks that will be generated in a tempdir that gets removed.

Differential Revision: https://reviews.llvm.org/D71675

Move the function to Driver.h so that it can be re-used in other places.

MallocChecker warns when memory is passed into -[NSData initWithBytesNoCopy]
but isn't allocated by malloc(), because it will be deallocated by free().
However, initWithBytesNoCopy has an overload that takes an arbitrary block
for deallocating the object. If such overload is used, it is no longer
necessary to make sure that the memory is allocated by malloc().

LLJITBuilder will now use JITLink on supported platforms even if a custom
JITTargetMachineBuilder is supplied, provided that neither the code model,
nor the relocation model, nor the ObjectLinkingLayerCreator is set.

implementing the resolution of CWG2352.

No functionality change, except that we now convert the referent of a
reference binding to the underlying type of the reference in more cases;
we used to happen to preserve the type sugar from the referent if the
only type change was in the cv-qualifiers.

This exposed a bug in how we generate code for trivial assignment
operators: if the type sugar (particularly the may_alias attribute)
got lost during reference binding, we'd use the "wrong" TBAA information
for the load during the assignment.

This is NFC since none of the constructor calls in trunk pass
hasOffsetApplied=true.

TypeSourceInfo is a thin wrapper around TypeLocs. Notionally, the best
place for it to live would be TypeLoc.h, but Decl.h requires it to be
complete, so it needs to be lower in the dependency graph. Type.h seems
like the next best place.

By itself, this change has no impact on build time, because it doesn't
remove a single dependency edge from a .cpp file to a .h file, but it is
an incremental step towards making the AST headers less interdependent.

Reviewers: rsmith

Differential Revision: https://reviews.llvm.org/D71427

Summary: Also add an error case when targetting an unimplement architecture.

Subscribers: nemanjai, jsji, shchenz, #sanitizers, llvm-commits

Tags: #sanitizers, #llvm

Differential Revision: https://reviews.llvm.org/D71670

Use LLVM_LINK_COMPONENTS to link the FrontendOpenMP library
instead of passing it explicitly to LINK_LIBS.  This fixes duplicating
the library when clang-tidy is linked to LLVM dylib.

Differential Revision: https://reviews.llvm.org/D71674

pass.
This is fixes changes from a85f5efd.

Revert D70315, as it breaks gfx8 for some reason.

This reverts commit 65f94b33.

Use `[tag]` as a prefix and "port" instead of "merge" to
avoid confusing with "git merge".

Remove the hack that populates the cpsr register in the gpr struct by
writing past the end of the array. This was tripping up ASan.

Patch by: Reva Cuthbertson

Add new intrinsics
   llvm.experimental.constrained.minimum
   llvm.experimental.constrained.maximum
as strict versions of llvm.minimum and llvm.maximum.

Includes SystemZ back-end support.

Reviewed By: craig.topper

Differential Revision: https://reviews.llvm.org/D71624

reg is unsigned type and used here for getting array element from the end by
negating it. negation of unsigned can result in large number and array access
with that index will result in segmentation fault.

Fixes: https://bugs.llvm.org/show_bug.cgi?id=43872

Patched by: kamlesh kumar

Differential Revision: https://reviews.llvm.org/D69893

It's a singleton in CoreFoundation that always contains a non-null CFNullRef.

Summary:
[libomptarget][nfc] Introduce atomic wrapper function

Wraps atomic functions in a template prefixed __kmpc_atomic that
dispatches to cuda or hip atomic functions. Intended to be easily extended
to dispatch to OpenCL or C++ atomics for a third target.

Reviewers: ABataev, jdoerfert, grokos

Reviewed By: jdoerfert

Subscribers: Anastasia, jvesely, mgrang, dexonsmith, llvm-commits, mgorny, jfb, openmp-commits

Tags: #openmp, #llvm

Differential Revision: https://reviews.llvm.org/D71404

Loop fusion previously had a method to check whether a loop was in rotated form. This method has
been moved into the LoopInfo class. This patch removes the old isRotated method from loop fusion,
in favour of the new one in LoopInfo.

Summary:
This patch adds instructions to the InstCombine worklist after they are properly inserted. This way we don't get `<badref>`s printed when logging added instructions.
It also adds a check in `Worklist::Add` that ensures that all added instructions have parents.

Simple test case that illustrates the difference when run with `--debug-only=instcombine`:

```
define i32 @test35(i32 %a, i32 %b) {
  %1 = or i32 %a, 1135
  %2 = or i32 %1, %b
  ret i32 %2
}
```

Before this patch:
```
INSTCOMBINE ITERATION #1 on test35
IC: ADDING: 3 instrs to worklist
IC: Visiting:   %1 = or i32 %a, 1135
IC: Visiting:   %2 = or i32 %1, %b
IC: ADD:   %2 = or i32 %a, %b
IC: Old =   %3 = or i32 %1, %b
    New =   <badref> = or i32 %2, 1135
IC: ADD:   <badref> = or i32 %2, 1135
...
```

With this patch:
```
INSTCOMBINE ITERATION #1 on test35
IC: ADDING: 3 instrs to worklist
IC: Visiting:   %1 = or i32 %a, 1135
IC: Visiting:   %2 = or i32 %1, %b
IC: ADD:   %2 = or i32 %a, %b
IC: Old =   %3 = or i32 %1, %b
    New =   <badref> = or i32 %2, 1135
IC: ADD:   %3 = or i32 %2, 1135
...
```

Reviewers: fhahn, davide, spatel, foad, grosser, nikic

Reviewed By: nikic

Subscribers: nikic, lebedev.ri, hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D71093

Moves out of temporaries caused warnings that failed builds.

Summary:
[libomptarget][nfc] Extract function from data_sharing, move to common

Finding the first active thread in the warp is different on nvptx and amdgcn,
mostly due to warp size and the desire for efficiency.

Reviewers: ABataev, jdoerfert, grokos

Reviewed By: jdoerfert

Subscribers: jvesely, mgorny, openmp-commits

Tags: #openmp

Differential Revision: https://reviews.llvm.org/D71643

The VFS/subframework-symlink.m test is still XFAIL on Windows, but for
a different reason than those fixed in PR43272, so I've updated the
PR number.

Since VFS paths can be in either Posix or Windows style, we have to use
a more flexible definition of "absolute" path.

The key here is that FileSystem::makeAbsolute is now virtual, and the
RedirectingFileSystem override checks for either concept of absolute
before trying to make the path absolute by combining it with the current
directory.

Differential Revision: https://reviews.llvm.org/D70701

Differential Revision: https://reviews.llvm.org/D41910

This reverts commit 830e08b9 and eb1857ce.

This commit leads to an unexpected failure on test/CodeGen/AArch64/sve-gather-scatter-dag-combine.ll.

The review will need more changes before its re-commited.

Let the "mnop-mcount" function attribute simply be present or non-present.
Update SystemZ backend as well to use hasFnAttribute() instead.

Review: Ulrich Weigand
https://reviews.llvm.org/D71669