This pass performs if-conversion on SSA form machine code by
speculatively executing both sides of the branch and using a cmov
instruction to select the result. This can help lower the number of
branch mispredictions on architectures like x86 that don't have
predicable instructions.

The current implementation is very aggressive, and causes regressions on
mosts tests. It needs good heuristics that have yet to be implemented.

llvm-svn: 159694

c-function parsing when a declaration with
C++0x braced-init-list is inside an @implementation.

llvm-svn: 159693

Our current inlining support (specifically RegionStore::enterStackFrame)
doesn't know that calls to overloaded operators may be calls to non-static
member functions, and that in these cases the first argument should be
treated as 'this'. This caused incorrect results and sometimes crashes.

The long-term fix will be to rewrite RegionStore::enterStackFrame to use
CallEvent and its subclasses, but for now we can just disable these
problematic calls by classifying them under a new CallEvent,
CXXMemberOperatorCall.

llvm-svn: 159692

defined in class implementations.

llvm-svn: 159691

c-function defined in objc class
implementation for now.

llvm-svn: 159690

objective-c's fast enumeration statement,
for more work to come.

llvm-svn: 159689

Make the interactive command interpreter build into the "crashlog" command work correctly when run inside LLDB. Before this fix the "cmd.Cmd" object was trying to read from stdin itself and it was competing without command interpreter for the bytes.

llvm-svn: 159688

Thanks Duncan for the idea

llvm-svn: 159687

Fixed a crasher in the "thread continue" code. There were many logic errors in the DoExecute function where thread index IDs were being used where the actual zero based thread index should have been used. This could cause crashes to happen since looking up a thread by index ID, when the zero based index of a thread should be used would return an empty thread shared pointer and cause a NULL deref.

llvm-svn: 159686

llvm-svn: 159685

runtime to gnustep from gnu.  Fix EH for the GCC runtime.

llvm-svn: 159684

By default on OS X 10.8, we don't link with a crt1.o file and the linker
knows to use _main as the entry point.  But, when compiling with -pg, we
need to link with the gcrt1.o file, and the linker needs to be told to use
the "start" symbol as the entry point.  The -no_new_main linker option does
that last part.  <rdar://problem/11491405>

llvm-svn: 159683

llvm-svn: 159682

llvm-svn: 159681

These give quite a bit more information about the DebugInfo and makes it more
readable.

llvm-svn: 159680

llvm-svn: 159679

values:

- Return integer vectors in integer registers.
- Pass vector arguments in integer registers.
- Set an upper bound for argument alignment. The largest alignment is 8-byte
  for O32 and 16-byte for N32/64.

llvm-svn: 159676

e.g. ExprWithCleanups.

llvm-svn: 159674

llvm-svn: 159673

(LLVM optimizers cannot do this optimization by themselves)

llvm-svn: 159668

Unfortunately this change requires the cache map to hold WeakVHs instead

llvm-svn: 159667

llvm-svn: 159666

Patch by Andy Gibbs.

llvm-svn: 159665

may be destroying an ASTUnit while cleanupOnDiskMapAtExit is
getting called.

rdar://11781241

llvm-svn: 159664

llvm-svn: 159662

llvm-svn: 159661

  IntegersSubsetMapping
  Added new methods
  - add(self& RHS, SuccessorClass *S)
  - detachCase
  - removeCase
  - findSuccessor
  - getCases
  - getCaseSingleNumber
  - isOverlapped
  

llvm-svn: 159660

  IntegersSubsetMapping
  - Replaced type of Items field from std::list with std::map. In neares future I'll test it with DenseMap and do the correspond replacement
    if possible.

llvm-svn: 159659

Optimized diff operation: implemented the case when LHS and RHS subsets contains single numbers only.

llvm-svn: 159658

llvm-svn: 159657

llvm-svn: 159655

[Sanitizer] Extend a symbolizer code. Implemented for Linux only. Use dl_iterate_phdr to get virtual addresses of mapped module sections. To symbolize an address from a module, map this module to memory and obtain pointers to debug info sections. Later these pointers can be passed to constructor of DWARF context-in-memory from LLVM DebugInfo lib.

llvm-svn: 159652

llvm-svn: 159650

Contributed by: Yabin Hu <yabin.hwu@gmail.com>

llvm-svn: 159649

some, and allows the routine to be inlined into common callers. The
various bits that hit this code in their hotpath seem slightly lower on
the profile, but I can't really measure a performance improvement as
everything seems to still be bottlenecked on likely cache misses. =/

llvm-svn: 159648

llvm-svn: 159647

llvm-svn: 159646

llvm-svn: 159645

Add aliases for pblendvb, blendvpd, and blendvps instructions with the implicit xmm0 operand specified. Fixes PR13252.

llvm-svn: 159644

llvm-svn: 159643