Build failed in
http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-autoconf/builds/27258

In lib/CodeGen/LiveDebugVariables.cpp:589, it uses std::prev(MBBI) to
get DebugValue's SlotIndex. however, the previous instruction may be
also a debug instruction.

llvm-svn: 338992

Generate DILabel metadata and call llvm.dbg.label after label
statement to associate the metadata with the label.

After fixing PR37395.

Differential Revision: https://reviews.llvm.org/D45045

llvm-svn: 338989

After refactoring DbgInfoIntrinsic class hierarchy, we use
DbgVariableIntrinsic as the base class of variable debug info.

In resolveTopLevelMetadata() in CGVTables.cpp, we only care about
dbg.value, so we try to cast the instructions to DbgVariableIntrinsic
before resolving variables.

Differential Revision: https://reviews.llvm.org/D50226

llvm-svn: 338985

errors on) and clean up the formattting.

This isn't actualy assembly anyways, so dropping the highlighting is
probably for the best.

llvm-svn: 338979

Summary:
Optimization remark format is slightly changed by LLVM patch D49412.
Two tests are fixed with expected messages changed.
Frankly speaking I have not tested this change yet. I will test when manage to setup the project.

Reviewers: xbolva00

Reviewed By: xbolva00

Subscribers: mehdi_amini, eraman, steven_wu, dexonsmith

Differential Revision: https://reviews.llvm.org/D50241

llvm-svn: 338971

These forward declarations for various classes in the Type
hierarchy are not needed since they are all forward declared
systematically a few lines below.

llvm-svn: 338966

Looks like the only reason we use a unique_ptr here is so that we can
conditionally construct a LogicalErrorHandler. It's a small type, and
Optional can do the same thing with 100% fewer heap allocations.

llvm-svn: 338962

llvm-svn: 338946

When a non-extended temporary object is created in a conditional branch, the
lifetime of that temporary ends outside the conditional (at the end of the
full-expression). If we're inserting lifetime markers, this means we could end
up generating

  if (some_cond) {
    lifetime.start(&tmp);
    Tmp::Tmp(&tmp);
  }
  // ...
  if (some_cond) {
    lifetime.end(&tmp);
  }

... for a full-expression containing a subexpression of the form `some_cond ?
Tmp().x : 0`. This patch moves the lifetime start for such a temporary out of
the conditional branch so that we don't need to generate an additional basic
block to hold the lifetime end marker.

This is disabled if we want precise lifetime markers (for asan's
stack-use-after-scope checks) or of the temporary has a non-trivial destructor
(in which case we'd generate an extra basic block anyway to hold the destructor
call).

Differential Revision: https://reviews.llvm.org/D50286

llvm-svn: 338945

Summary:
Added corpus of arrays to use as inputs for the functions. Check that the two
functions modify the inputted arrays in the same way.

Reviewers: kcc, morehouse

Reviewed By: morehouse

Subscribers: mgorny, cfe-commits, llvm-commits

Differential Revision: https://reviews.llvm.org/D50194

llvm-svn: 338943

llvm-svn: 338942

__builtin_memmove (in non-type-punning cases).

This is intended to permit libc++ to make std::copy etc constexpr
without sacrificing the optimization that uses memcpy on
trivially-copyable types.

__builtin_strcpy and __builtin_wcscpy are not handled by this change.
They'd be straightforward to add, but we haven't encountered a need for
them just yet.

This reinstates r338455, reverted in r338602, with a fix to avoid trying
to constant-evaluate a memcpy call if either pointer operand has an
invalid designator.

llvm-svn: 338941

Just do not generate the note at all in that case.

llvm-svn: 338935

Libc++ needs to know when aligned allocation is supported by clang, but is
otherwise unavailable at link time. Otherwise, libc++ will incorrectly end up
generating calls to `__builtin_operator_new`/`__builtin_operator_delete` with
alignment arguments.

This patch implements the following changes:

* The `__cpp_aligned_new` feature test macro to no longer be defined when
  aligned allocation is otherwise enabled but unavailable.

* The Darwin driver no longer passes `-faligned-alloc-unavailable` when the
  user manually specifies `-faligned-allocation` or `-fno-aligned-allocation`.

* Instead of a warning Clang now generates a hard error when an aligned
  allocation or deallocation function is referenced but unavailable.

Patch by Eric Fiselier.

Reviewers: rsmith, vsapsai, erik.pilkington, ahatanak, dexonsmith

Reviewed By: rsmith

Subscribers: Quuxplusone, cfe-commits

Differential Revision: https://reviews.llvm.org/D45015

llvm-svn: 338934

Differential Revision: https://reviews.llvm.org/D45712

llvm-svn: 338931

Do not crash if a CXXRecordDecl cannot be obtained for an object.

Special thanks for the reproduction to Alexander Kornienko.

llvm-svn: 338918

Encoding offload target triples onto comdat group key for offload initialization
code guarantees that it will be executed once per each unique combination of
offload targets.

Differential Revision: https://reviews.llvm.org/D50218

llvm-svn: 338916

Properly add shared locks to the initial list of locks being tracked, instead of assuming unlock functions always use exclusive locks.

Patch by Aaron Puchert.

llvm-svn: 338912

Found by KlockWorks, this variable is properly protected, however
the conditions in the test that initializes it and the one that uses
it could diverge, it seems to me that this is a 'free' init that will
prevent issues if one of the conditions is ever modified without the other.

llvm-svn: 338909

This reverts commit r338899, it was causing ASan test failures on sanitizer-x86_64-linux-fast.

llvm-svn: 338904

Ensures the statically sized alloca is not converted to DYNAMIC_STACKALLOC
later because it is not in the entry block.

Differential Revision: https://reviews.llvm.org/D50104

llvm-svn: 338899

- Testcase attempts to (not) grep 'g0' in output to ensure asm symbol is
  properly renamed, but g0 is too generic and can be part of the
  module's path in LLVM IR output.
- Changed to grep '@g0', which is what the proper global symbol name
  would be if not using asm.

llvm-svn: 338895

No longer necessary thanks to r338889 (and friends).

llvm-svn: 338893

Summary:
Previously, clang-format would avoid breaking before the first `{`
found, but then happily break before subsequent '{'s on the line. This
change fixes that by looking for the first location that has no opening
curly, if any.

This fixes the original commit by correcting the loop condition.

This reverts commit 66dc646e09b795b943668179c33d09da71a3b6bc.

Reviewers: krasimir

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D50249

llvm-svn: 338890

Clang format got the best of me... it introduced spaces around something 
in a table-genned file, so it was interpreted as an array and not a 
code block.  

llvm-svn: 338889

Add a comment in ObjCMethodDecl and ObjCContainerDecl stating that
we store some bits in ObjCMethodDeclBits and ObjCContainerDeclBits.

This was missed by the recent move in
r338641 : [AST][4/4] Move the bit-fields from ObjCMethodDecl
          and ObCContainerDecl into DeclContext

llvm-svn: 338888

Factored out from https://reviews.llvm.org/D49729
following @erichkeane comments.

* Add missing classes in the list of classes
  deriving directly from DeclContext.
* Move the friend declarations together and
  add a comment for why they are required.

Reviewed By: erichkeane

Differential Revision: https://reviews.llvm.org/D49790

llvm-svn: 338887

llvm-svn: 338886

llvm-svn: 338885

llvm-svn: 338884

This reverts commit r338837, it introduced an infinite loop on all bots.

llvm-svn: 338879

Summary: It is not necessary, but would be nice if the script run on python3 as well (as opposed to only python2, which is going to be deprecated https://pythonclock.org/)

Contributed by MarcoFalke!

Reviewers: krasimir

Reviewed By: krasimir

Subscribers: lebedev.ri, sammccall, cfe-commits

Differential Revision: https://reviews.llvm.org/D48098

llvm-svn: 338839

Summary:
Previously, clang-format would avoid breaking before the first `{`
found, but then happily break before subsequent '{'s on the line. This
change fixes that by looking for the first location that has no opening
curly, if any.

Reviewers: krasimir

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D50230

llvm-svn: 338837

Summary:
For example, when renaming `a::b::x::foo` to `y::foo` below, replacing
`x::foo()` with `y::foo()` can cause ambiguity. In such cases, we simply fully
qualify the name with leading `::`.
```
namespace a {
namespace b {
namespace x { void foo() {} }
namespace y { void foo() {} }
}
}

namespace a {
namespace b {
void f() { x::foo(); }
}
}
```

Reviewers: ilya-biryukov, hokein

Reviewed By: ilya-biryukov

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D50189

llvm-svn: 338832

llvm-svn: 338831

Summary:
This adds support to libclang for reading the flag_enum attribute.

This also bumps CINDEX_VERSION_MINOR for this patch series.

Reviewers: yvvan, jbcoe

Reviewed By: yvvan

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D49635

llvm-svn: 338820

Summary: This allows libclang to access the actual names of property setters and getters without needing to go through the indexer API. Usually default names are used, but the property can specify a different name.

Reviewers: yvvan, jbcoe

Reviewed By: yvvan

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D49634

llvm-svn: 338816

Summary:
Having access to implicit attributes is sometimes useful so users of libclang don't have to duplicate some of the logic in sema.

This depends on D49081 since it also adds a CXTranslationUnit flag.

Reviewers: yvvan, jbcoe

Reviewed By: yvvan

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D49631

llvm-svn: 338815

Summary:
This adds support to libclang for identifying ObjC related attributes that don't take arguments.

All attributes but NSObject and NSConsumed are tested.

Reviewers: yvvan, jbcoe

Reviewed By: yvvan

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D49127

llvm-svn: 338813

Summary:
Emit !llvm.mem.parallel_loop_access metadata for memory accesses even if the parallel loop is not the top on the loop stack.

Fixes llvm.org/PR37558.

Reviewers: ABataev, hfinkel, amusman, tyler.nowicki

Reviewed By: hfinkel

Subscribers: Meinersbur, hfinkel, cfe-commits

Differential Revision: https://reviews.llvm.org/D48808

llvm-svn: 338810