llvm-svn: 137412

llvm-svn: 137411

eliminating a pile of redundant code (and probably some bugs in the
process). The variation between chained and non-chained PCH is fairly
small now anyway.

llvm-svn: 137410

llvm-svn: 137409

llvm-svn: 137408

llvm-svn: 137407

called on another global variable. Use ManagedStatic to ensure that the global
we register with actually exists when we need it.

llvm-svn: 137406

The initialization of 'replicated_element' to 0 is needed, otherwise we get a garbage value to start with.

llvm-svn: 137405

Fix by Ivan Baev. Sorry I don't have a unit test, but the fix is obvious so I don't want to delay it.

llvm-svn: 137404

llvm-svn: 137403

is returned through a bitcast.

llvm-svn: 137402

1) Change SymbolDependTy map to keep pointers as data. And other small tweaks like making the DenseMap smaller 64->16 elements; remove removeSymbolDependencies() as it will probably not be used.

2) Do not mark dependents live more then once.

llvm-svn: 137401

llvm-svn: 137400

the retains and releases all use the same SSA pointer value.

Also, don't let CFG hazards disrupt nested retain+release pair
optimizations.

llvm-svn: 137399

rather than plain postorder, so that CFG constructs like single-exit loops
are reliably visited in a sensible order.

llvm-svn: 137398

This is possible now that we now longer provide an interface to iterate
the interference overlaps.

llvm-svn: 137397

llvm-svn: 137396

declaration that never actually gets serialized. Instead, serialize
the various kinds of update records (lexical decls, visible decls, the
addition of an anonymous namespace) for the translation unit, even if
we're not chaining. This way, we won't have to deal with multiple
loaded translation unit declarations.

llvm-svn: 137395

llvm-svn: 137394

llvm-svn: 137393

causing problems with printing the values of persistent
variables with struct types.

llvm-svn: 137392

llvm-svn: 137389

llvm-svn: 137388

llvm-svn: 137387

llvm-svn: 137386

checked in at r137276 and r137341.

llvm-svn: 137385

SCEV unrolling can unroll loops with arbitrary induction variables. It
is a prerequisite for -disable-iv-rewrite performance. It is also
easily handles loops of arbitrary structure including multiple exits
and is generally more robust.

This is under a temporary option to avoid affecting default
behavior for the next couple of weeks. It is needed so that I can
checkin unit tests for updateUnloop.

llvm-svn: 137384

No test case, sorry. It's one of those bugs where it's really really hard to make one. rdar://9910862.

llvm-svn: 137383

llvm-svn: 137382

llvm-svn: 137381

The Query class now holds two iterators instead of an InterferenceResult
instance. The iterators are used as bookmarks for repeated
collectInterferingVRegs calls.

llvm-svn: 137380

warning. 

llvm-svn: 137378

The current warning in -Wnull-arithmetic for comparisons between NULL and non-pointers is not very helpful.  This patch will update the wording to be more helpful to users.

Old warning:

warning: use of NULL in arithmetic operation [-Wnull-arithmetic]
  return 10 <= NULL;
            ^  ~~~~

New warning:

warning: comparison between NULL and non-pointer ('int' and NULL) [-Wnull-arithmetic]
  return 10 <= NULL;
         ~~ ^  ~~~~
llvm-svn: 137377

llvm-svn: 137376

llvm-svn: 137375

llvm-svn: 137374

either "special" type has already been initialized. Previously, we did
this check based on just the first special type (__builtin_va_list),
but now we have some NULL special type entries to content with.

llvm-svn: 137373

llvm-svn: 137372

llvm-svn: 137371

llvm-svn: 137370