Summary:
This change makes CMake scan for lit suites and generate a target for each lit test suite. The targets follow the format check-<project>-<suite path>.

For example:
check-llvm-unit - Runs the LLVM unit tests
check-llvm-codegen-arm - Runs the ARM codeine tests

Note: These targets are not generated during multi-configuration generators (i.e. Xcode and Visual Studio) because target clutter impacts UI usability.

Reviewers: chandlerc

Subscribers: aemerson, llvm-commits

Differential Revision: http://reviews.llvm.org/D8380

llvm-svn: 232671

appears to have broken tests/bots.

This reverts commit r232660.

llvm-svn: 232670

I deleted lldb-log.cpp, but the Makefile build lists this file
explicitly, so it is removed in this patch.

llvm-svn: 232669

llvm-svn: 232668

Committed too early.

This reverts commit r232666.

llvm-svn: 232667

they can be used without a subtarget in constructing subtarget
independent passes.

llvm-svn: 232666

as we don't necessarily need to do this yet - though we could move
the base class to the TargetMachine as it isn't subtarget dependent.

This reverts commit r232103.

llvm-svn: 232665

No outlining is necessary for SEH catch blocks. Use the blockaddr of the
handler in place of the usual outlined function.

Reviewers: majnemer, andrew.w.kaylor

Differential Revision: http://reviews.llvm.org/D8370

llvm-svn: 232664

llvm-svn: 232663

The MSVC linker won't produce a .lib file for an executable that doesn't
export anything, and LLVM doesn't maintain dllexport annotations or .def
files listing all C++ symbols. It also doesn't support exporting all
symbols, like binutils ld.

CMake 3.2 changed the Ninja generator to list both the .exe and .lib
files as outputs of executable build targets. Ninja would always re-link
executables with ENABLE_EXPORTS because the .lib output file was not
present, and therefore the target was out of date.

llvm-svn: 232662

Should bring the bots back.

llvm-svn: 232661

Currently v2i64 vectors shifts (non-equal shift amounts) are scalarized, costing 4 x extract, 2 x x86-shifts and 2 x insert instructions - and it gets even more awkward on 32-bit targets.

This patch separately shifts the vector by both shift amounts and then shuffles the partial results back together, costing 2 x shuffles and 2 x sse-shifts instructions (+ 2 movs on pre-AVX hardware).

Note - this patch only improves the SHL / LSHR logical shifts as only these are supported in SSE hardware.

Differential Revision: http://reviews.llvm.org/D8416

llvm-svn: 232660

[Objdump] DumpBytes of uint8_t from ArrayRef<uint8_t> instead of char from StringRef.  Removing reinterpret_casts.

llvm-svn: 232659

This removes duplicated code from backends that don't need to do anything
fancy.

llvm-svn: 232658

llvm-svn: 232657

This reverts r232650.

Missed a piece of code in the previous commit.

llvm-svn: 232656

llvm-svn: 232655

llvm-svn: 232654

So that we don't have to update every single #include in the entire
codebase to #include this new header (which used to get included by
lldb-private-log.h, we automatically #include "Logging.h" from
within "Log.h".

llvm-svn: 232653

When calculating the lanemask of a register class we have to include the
masks of subregisters supported by any of the class members, not just
the ones supported by all class members.

This fixes problems when coalescing towards a subclass with additional
subregisters available.

The attached testcase works fine as is, but does crash if you enable
subregister liveness on x86 without this change applied.

llvm-svn: 232652

We can get there with .code64.

Fixes pr22349.

llvm-svn: 232651

llvm-svn: 232650

Always add some fake threads with x86_64 registers no matter what the architecture since this is just for testing that we can add new threads with completely different registers contexts to a process.

llvm-svn: 232649

A previous attempt to make the unit tests link properly on
Linux broke it for Windows.  This patch fixes it for both platforms.

llvm-svn: 232648

The checks here were so vague that we could nuke intrinsics
from existence and still pass the test because we'd match
the function name.

llvm-svn: 232647

llvm-svn: 232646

llvm-svn: 232645

The 'vmovntdq' was only passing due to a fluke in
SandyBridge codegen that splits 32-byte stores in half, 
but that meant that the test was not correctly checking
for the 32-byte store that we thought we were generating.

The lax checking in this file will be addressed in
another commit. There are bigger problems here.

llvm-svn: 232644

llvm-svn: 232643

This test started passing without warning, so in theory, we shouldn't mark
is as XPASS, but the buildbots are red and there's been an uncaught regression
meanwhile. Since it's passing on {ARM,Thumb2} x {NEON,VFPv3}, I'll risk future
failures for the benefit of getting the bots green again.

llvm-svn: 232642

1) <cstdlib> header should define std::abs([int|long|long long])
functions. They use "using ::abs" to import these functions (which are
declared in <stdlib.h>) into std namespace.
2) <cmath> header should define std::abs([float|double|long double])
function. If we try define new functions in std namespace, then it
will cause compile error in <cstdlib> because "using ::abs" will try
import not only [int|long|long long] functions, but also
[float|double|long double] which are defined in <math.h> header on
solaris.

Patch by C Bergstrom.

llvm-svn: 232641

It is already skipped on the other platforms anyhow.

llvm.org/pr19246

llvm-svn: 232640

llvm-svn: 232639

llvm-svn: 232638

llvm-svn: 232636

Before:
  void aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa(aaaaaaaaaaaaaaaaaaaaaaaaa*
                                              const aaaaaaaaaaaa) {}

After:
  void aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa(
      aaaaaaaaaaaaaaaaaaaaaaaaa* const aaaaaaaaaaaa) {}

llvm-svn: 232635

This moves the conversion of the open options to the target platform. On mac fcntl.h has different values for O_CREAT and O_TRUNC than on linux so by transmitting the standardized lldb open options we can correctly convert them on the target platform.

Test Plan:
On linux:
  lldb-server p --listen *:1234
On mac:
  lldb
  platform select remote-linux
  platform connect connect://ip-of-linux-box:1234
  target create ~/path/to/linux/binary
  b main
  process launch
Binary is successfully pushed to linux remote, process successfully launches and break in the main method.

Differential Revision: http://reviews.llvm.org/D8395

llvm-svn: 232634

llvm-svn: 232633

spacing also fixed by r230473.

The fix in r230473 was done to enable fixing the spacing for
  std::function<void( int, int )>.
I did not realized that it also fixed this
issue.  Since it is fairly different from
  Deleted &operator=(const Deleted &)& = default;
fixed in r230473, it seems sensible to add the regression test for it.

Also cleaned up the test by removing duplicated code and comment, and kept
repeated test set consistent.

Result of running the new tests with r230473 backed out:
  [ RUN      ] FormatTest.ConfigurableSpacesInParentheses
  Actual: "std::function<void(int, int)> callback;"
  Expected: "std::function<void( int, int )> callback;"
  Actual: "std::function<void( int, int )> callback;"
  Expected: "std::function<void(int, int)> callback;"
  Actual: "std::function<void( int, int ) > callback;"
  Expected: "std::function<void(int, int)> callback;"
  [  FAILED  ] FormatTest.ConfigurableSpacesInParentheses (402 ms)

Result of new tests with r230473:
  [ RUN      ] FormatTest.ConfigurableSpacesInParentheses
  [       OK ] FormatTest.ConfigurableSpacesInParentheses (209 ms)

Review: http://reviews.llvm.org/D7922

Patch by Jean-Philippe Dufraigne. Thanks!

llvm-svn: 232632

OpenCL C Spec v2.0 Section 6.13.11

- Made c11 _Atomic being not accepted for OpenCL

- Implemented CL2.0 atomics by aliasing them to the corresponding c11 atomic types using implicit typedef

- Added diagnostics for atomics Khronos extension enabling

llvm-svn: 232631