llvm-svn: 178364

wasn't loaded into the underlying process or
couldn't be found.

<rdar://problem/13533671>

llvm-svn: 178363

Like nearbyint, rint can be implemented on PPC using the frin instruction. The
complication comes from the fact that rint needs to set the FE_INEXACT flag
when the result does not equal the input value (and frin does not do that). As
a result, we use a custom inserter which, after the rounding, compares the
rounded value with the original, and if they differ, explicitly sets the XX bit
in the FPSCR register (which corresponds to FE_INEXACT).

Once LLVM has better modeling of the floating-point environment we should be
able to (often) eliminate this extra complexity.

llvm-svn: 178362

* Store the .block_descriptor (instead of self) in the alloca so we
  can guarantee that all captured variables are available at -O0.
* Add the missing OpDeref for the alloca.
rdar://problem/12767564

llvm-svn: 178361

* Let DIType for block-captured self to point to the completed cached
  interface type.
rdar://problem/12767564

llvm-svn: 178360

llvm-svn: 178359

llvm-svn: 178358

A9 uses itinerary classes, Swift uses RW lists. This tripped some
verification when we're expanding variants. I had to refine the
verification a bit.

llvm-svn: 178357

llvm-svn: 178356

llvm-svn: 178355

llvm-svn: 178354

rdar://problem/12767564

llvm-svn: 178353

llvm-svn: 178352

llvm-svn: 178351

The 3rd test in shrink_to_fit.pass.cpp can't possibly pass if exceptions are disabled, so #ifdef'ing out the test.

llvm-svn: 178350

It was superseded by MachineBlockPlacement and disabled by default since LLVM 3.1.

llvm-svn: 178349

Enable tab completion for regular expression commands.

llvm-svn: 178348

llvm-svn: 178346

llvm-svn: 178345

[tsan] restore performance critical inlining in tsan: remove static from ALWAYS_INLINE, use ALWAYS_INLINE USED for critical functions. 

llvm-svn: 178341

This removes all register spills from the performance-critical __tsan_{read,write}{1,2,4,8} functions when the host compiler is GCC.

llvm-svn: 178340

[libsanitizer] Put COMMON_CFLAGS/COMMON_CXXFLAGS before Tmp.CFLAGS to let the config-specific flags override common flags.

llvm-svn: 178339

These instructions are available on the P5x (and later) and on the A2. They
implement the standard floating-point rounding operations (floor, trunc, etc.).
One caveat: frin (round to nearest) does not implement "ties to even", and so
is only enabled in fast-math mode.

llvm-svn: 178337

llvm-svn: 178336

llvm-svn: 178335

This reverts commit 617330909f0c26a3f2ab8601a029b9bdca48aa61.

It broke the bots:

/home/clangbuild2/clang-ppc64-2/llvm.src/unittests/ADT/SmallVectorTest.cpp:150: PushPopTest
/home/clangbuild2/clang-ppc64-2/llvm.src/unittests/ADT/SmallVectorTest.cpp:118: Failure
Value of: v[i].getValue()
  Actual: 0
Expected: value
Which is: 2

llvm-svn: 178334

We already avoided warning for

extern "C" const char *Version_string = "2.9";

now we also don't produce any warnings for

extern "C" {
  extern const char *Version_string2 = "2.9";
}

llvm-svn: 178333

being power-of-two sized.

llvm-svn: 178332

llvm-svn: 178331

llvm-svn: 178330

llvm-svn: 178329

register classes for Mips64 and DSP-ASE.

No functionality changes.

llvm-svn: 178328

No functionality changes.

llvm-svn: 178327

No functionality changes.

llvm-svn: 178326

<rdar://problem/13527167>

llvm-svn: 178325

“process attach” should ask the same questions as process launch if there is a current process.
“process connect” then “process launch” or “process attach” should actually work.

<rdar://problem/13524210>
<rdar://problem/13524208>
<rdar://problem/13488919>

llvm-svn: 178324

llvm-svn: 178323

[cfg] Always guard (when AddStaticInitBranches == true) DeclStmts for static variables, not just ones with explicit initializers

llvm-svn: 178322

llvm-svn: 178321

llvm-svn: 178320