IRBuilder has two ways of putting bundle operands on calls: the default
operand bundle, and an overload of CreateCall that takes an operand
bundle list.

Previously, this overload used a default argument of None. This made it
impossible to distinguish between the case were the caller doesn't care
about bundles, and the case where the caller explicitly wants no
bundles. We behaved as if they wanted the latter behavior rather than
the former, which led to problems with simplifylibcalls and WinEH.

This change fixes it by making the parameter non-optional, so we can
distinguish these two cases.

llvm-svn: 261258

llvm-svn: 261255

Summary: As per title. There was a lot of part missing in the C API, so I had to extend the invoke and landingpad API.

Reviewers: echristo, joker.eph, Wallbraker

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D17359

llvm-svn: 261254

llvm-svn: 261253

Using Backend_EmitLL attemps to create a file with an empty filename.
This is problematic in certain environments: an empty filename may be
illegal, or the default output path may not be writable (in the case
where an empty filename would otherwise have some non-failing
semantics). This patch switches to use Backend_EmitMCNull, which
allows CodeGen to run, but does not attempt to create or write an
output file.

Differential Revision: http://reviews.llvm.org/D17405

llvm-svn: 261252

llvm-svn: 261251

llvm-svn: 261250

llvm-svn: 261249

It explains why we can't just synthesize bodies of setters in BodyFarm.

llvm-svn: 261248

http://reviews.llvm.org/D17397

llvm-svn: 261247

Fix a typo in FormatCache.cpp such that the cache would potentially return an invalid format in some cases

llvm-svn: 261246

These atomic operations are conceptually both a load and store from the same location. As such, we can treat them as the most conservative of those two components which in practice, means we can treat them like stores. An cmpxchg or atomicrmw captures the values, but not the locations accessed.

Note: We can probably be more aggressive about the comparison value in an cmpxhg since to have it be in memory, it must already be captured, but I figured it was better to avoid that for the moment.

Note 2: It turns out that since we don't actually support cmpxchg of pointer type, writing a negative test is impossible.

Differential Revision: http://reviews.llvm.org/D17400

llvm-svn: 261245

When modeling a call to a setter for a property that is synthesized to be
backed by an instance variable, don't invalidate the entire instance
but rather only the storage for the updated instance variable itself.

This still doesn't model the effect of the setter completely. It doesn't
bind the set value to the ivar storage location because doing so would cause
the set value to escape, removing valuable diagnostics about potential
leaks of the value from the retain count checker.

llvm-svn: 261243

llvm-svn: 261242

Differential Revision: http://reviews.llvm.org/D17088

llvm-svn: 261241

Differential Revision: http://reviews.llvm.org/D17106

llvm-svn: 261240

This patch adds a variety of different methods to query source
and line number information from PDB files.

llvm-svn: 261239

Since the behaviour is now different between Darwin and non-Darwin,
more triples are needed :-/

llvm-svn: 261238

Differential Revision: http://reviews.llvm.org/D17379

llvm-svn: 261237

llvm-svn: 261236

In r260133, LLVM was changed to no longer extend i8/i16 return values,
as it's not required by the ABI. However, code was found in the wild
that relies on the old behaviour on Darwin, so this commit reverts
back to that old behaviour for Darwin.

On other platforms, it's less likely that code would be depending on
the old behaviour, as GCC and MSVC haven't been extending such return
values.

llvm-svn: 261235

llvm-svn: 261234

We were erroneously reporting 16K as the page size on Windows because
the code that does the shadow mapping was using page size instead of
allocation granularity. After fixing that, we can resolve the FIXMEs in
the Windows implementations of GetPageSize and GetMmapGranularity by
calling GetSystemInfo instead of returning hard-coded, incorrect
answers.

llvm-svn: 261233

llvm-svn: 261232

This is one part of many of a locale refactor. See
http://reviews.llvm.org/D17146 for an idea of where this is going.

For the locale refactor, the locale management functions (newlocale,
freelocale, uselocale) are needed in a separate header from the various _l
functions. This is because some platforms implement the _l functions in terms
of a locale switcher RAII helper, and the locale switcher RAII helper needs
the locale management functions. This patch helps pave the way by getting all
the functions in the right files, so that later diffs aren't completely
horrible.

The "do-nothing" / "nop" locale functions are also useful on their own for
other lightweight platforms. Putting these nop implementations in
support/xlocale should enable code sharing.

Unfortunately, I have no access to a newlib system to build and test with, so
this change has been made blind.

Reviewed: http://reviews.llvm.org/D17382
llvm-svn: 261231

This is one part of many of a locale refactor. See
http://reviews.llvm.org/D17146 for an idea of where this is going.

For the locale refactor, the locale management functions (newlocale,
freelocale, uselocale) are needed in a separate header from the various _l
functions. This is because some platforms implement the _l functions in terms
of a locale switcher RAII helper, and the locale switcher RAII helper needs
the locale management functions. This patch helps pave the way by getting all
the functions in the right files, so that later diffs aren't completely
horrible.

Unfortunately, I have no access to an AIX machine to build with, so this change
has been made blind. Also, the original author (Xing Xue) does not appear to
have a Phabricator account.

Reviewed: http://reviews.llvm.org/D17380
llvm-svn: 261230

FreeBSD does not install a number of Clang-provided headers for the
compiler in the base system due to incompatibilities between FreeBSD's     
and Clang's versions. As a workaround do not use --sysroot=. on FreeBSD    
until this is addressed.

llvm.org/pr26651
Differential Revision:	http://reviews.llvm.org/D17383

llvm-svn: 261229

covmap needs to created as non allocatable, but not with
SHT_NOTE. The latter was needed to workaround a problem
of BFD linker with gc, which is no longer needed. (A more
proper longer term fix requires changing FE driver to force
referencing the section using linker script).

Differential Revision: http://reviews.llvm.org/D17309

llvm-svn: 261228

Commit r260721(http://reviews.llvm.org/D17182) introduced the following error when building for OSX using cmake:

Undefined symbols for architecture x86_64:

"_PyInit__lldb", referenced from:
   -exported_symbol[s_list] command line option

ld: symbol(s) not found for architecture x86_64
clang: error: linker command failed with exit code 1 (use -v to see invocation)

Adding '*' to the regex solves this problem, since it makes the symbol optional.

Reviewers: sivachandra, zturner, labath
Subscribers: lldb-commits
Differential Revision: http://reviews.llvm.org/D17384

llvm-svn: 261227

  To support non-aligned accesses we introduce a virtual element size
  for arrays that divides each access function used for this array. The
  adjustment of the access function based on the element size of the
  array was therefore moved after this virtual element size was
  determined, thus after all accesses have been created.

Differential Revision: http://reviews.llvm.org/D17246

llvm-svn: 261226

Summary:
These correspond to IMAGE_LOAD/STORE[_MIP] and are going to be used by Mesa
for the GL_ARB_shader_image_load_store extension.

IMAGE_LOAD is already matched by llvm.SI.image.load. That intrinsic has
a legacy name and pretends not to read memory.

Differential Revision: http://reviews.llvm.org/D17276

llvm-svn: 261224

llvm-svn: 261223

Llvm module object is shared between CodeGenerator and BackendConsumer,
in both classes it is stored as std::unique_ptr, which is not a good
design solution and can cause double deletion error. Usually it does
not occur because in BackendConsumer::HandleTranslationUnit the
ownership of CodeGenerator over the module is taken away. If however
this method is not called, the module is deleted twice and compiler crashes.

As the module owned by BackendConsumer is always the same as CodeGenerator
has, pointer to llvm module can be removed from BackendGenerator.

Differential Revision: http://reviews.llvm.org/D15450

llvm-svn: 261222

Add an AST matcher for real floating-point types. e.g., float, double, long double, but not complex.

llvm-svn: 261221

Compiling Hexagon target with GCC 6 produces "error: should have been
declared inside" due to GCC PR c++/69657 which was merged.

Properly wrapping operator<<() definitions within the namespace llvm
fixes the issue.

Author: domagoj.stolfa

Differential Revision: http://reviews.llvm.org/D17281

llvm-svn: 261220

Patch by Alex Sidorin!

Differential Revision: http://reviews.llvm.org/D17376

llvm-svn: 261219

Patch by Anand Kodnani.

llvm-svn: 261218

Currently we return no results when completing inside of the brackets in
a 'char foo[]' declaration. Let the generic expression completion code
handle it instead. We could get fancier here (e.g. filter non-constant
expressions in contexts where VLAs are not allowed), but it's a strict
improvement over the existing version.

llvm-svn: 261217

llvm-svn: 261216

llvm-svn: 261215