For Polly the two interesting changes are short_circuit && and || AST
expressions as well as the introduction of isl_ast_build_expr_from_set,
a well defined interface to compute ast expressions from constraint sets.

llvm-svn: 230636

The latency for the WriteMULm class was set to 4, which is actually lower than the latency for WriteMULr (5). 
A better estimate would be 4 added to WriteMULr, that is, 9.

llvm-svn: 230634

llvm-svn: 230633

InstCombine has logic to convert aligned Altivec load/store intrinsics into
regular loads and stores. Unfortunately, there seems to be no regression test
covering this behavior. Adding one...

llvm-svn: 230632

Wrap the StopTheWorld call in a dl_iterate_phdr() callback. This ensures that no
other threads are holding the libdl lock, and we can safely reenter it in the
tracer thread.

llvm-svn: 230631

llvm-svn: 230630

llvm-svn: 230629

llvm-svn: 230628

Before:
  Aaaa aaaaaaaaaaa{
      {
       a, // +1 indent weird.
       b, // trailing comma signals one per line.
      },  // trailing comma signals one per line.
  };

After:
  Aaaa aaaaaaaaaaa{
      {
          a, // better!?
          b, // trailing comma signals one per line.
      },     // trailing comma signals one per line.
  };

Interesting that this apparently was entirely untested :-(.

llvm-svn: 230627

Differential revision: http://reviews.llvm.org/D7853

llvm-svn: 230626

Before:
  vector<int>(*foo_)[6];

After:
  vector<int> (*foo_)[6];

llvm-svn: 230625

  With the patches r230325, r230329 and r230340 we can handle non-affine
  control flow in (loop-free) subregions. As all LLVM test-suite tests pass and
  we get ~20% more non-trivial SCoPs, we activate it now by default.

llvm-svn: 230624

formulaic into the top v8i16 lowering routine.

This makes the generalized lowering a completely general and single path
lowering which will allow generalizing it in turn for multiple 128-bit
lanes.

llvm-svn: 230623

MVT.

llvm-svn: 230622

Fixed crash on codegen for directives like 'omp for', 'omp single' etc. inside of the 'omp parallel', 'omp task' etc. regions.

llvm-svn: 230621

This isn't generally nice, but better than violating the column limit.

llvm-svn: 230620

backedge-taken count in profiliing data.

llvm-svn: 230619

IRCE can now split the iteration space for loops like:

   for (i = n; i >= 0; i--)
     a[i + k] = 42; // bounds check on access

llvm-svn: 230618

This ordering matches that of DAG.getNode.

llvm-svn: 230617

[X86] Remove the blendps/blendpd builtins. They aren't used by the intrinsic headers. We use appropriate shuffle vector instead.

llvm-svn: 230616

llvm-svn: 230615

This fixes a linker crash (found out while testing --gc-sections,
testcase provided by Rafael Avila de Espindola).
While this behaviour was found while testing ELF, it' not necessarily
ELF specific and this change is (apparently) harmless on all the
other drivers.

Differential Revision:  D7823
Reviewed by:    ruiu

llvm-svn: 230614

Apparently we can use this now!

llvm-svn: 230613

Add `CHECK-SAME`, which requires that the pattern matches on the *same*
line as the previous `CHECK`/`CHECK-NEXT` -- in other words, no newline
is allowed in the skipped region.  This is similar to `CHECK-NEXT`,
which requires exactly 1 newline in the skipped region.

My motivation is to simplify checking the long lines of LLVM assembly
for the new debug info hierarchy.  This allows CHECK sequences like the
following:

    CHECK:      ![[REF]] = !SomeMDNode(
    CHECK-SAME: file: ![[FILE:[0-9]+]]
    CHECK-SAME: otherField: 93{{[,)]}}

which is equivalent to:

    CHECK: ![[REF]] = !SomeMDNode({{.*}}file: ![[FILE:[0-9]+]]{{.*}}otherField: 93{{[,)]}}

While this example just has two fields, many nodes in debug info have
more than that.  `CHECK-SAME` will keep the logic easy to follow.

Morever, it enables interleaving `CHECK-NOT`s without allowing newlines.
Consider the following:

    CHECK:      ![[REF]] = !SomeMDNode(
    CHECK-SAME: file: ![[FILE:[0-9]+]]
    CHECK-NOT:  unexp...

There is no supported toolchain which provides headers / libs / object
files specific to the mips32r[3|5] and mips64r[3|5] ISA. So select "r2"
specific folders when they are available.

http://reviews.llvm.org/D7879

llvm-svn: 230611

Use the newly minted `DIImportedEntity` default constructor (r230609)
rather than explicitly specifying `nullptr`.  The latter will become
ambiguous when the new debug info hierarchy is committed, since we'll
have both of the following:

    explicit DIImportedEntity(const MDNode *);
    DIImportedEntity(const MDImportedEntity *);

(Currently we just have the former.)

A default constructor is just as clear.

llvm-svn: 230610

Add a default constructor for `DIImportedEntity`, to be used in clang in
a follow-up.

llvm-svn: 230609

Also remove the somewhat misleading initializers from
VectorizationFactor and VectorizationInterleave.  They will get
initialized with the default ctor since no cl::init is provided.

llvm-svn: 230608

llvm-svn: 230607

This cleans up a set of -Wsign-conversion, -Wint-conversion, and -Wformat
warnings from GCC 4.9.2 on Linux.  NFC.

llvm-svn: 230606

http://reviews.llvm.org/D7876

llvm-svn: 230604

Fix for PR22017.  Integer template arguments are automatically bit extended to
the size of the integer type.  In template diffing, evaluated expressions were
not having their results extending, leading to comparing two APSInt's with
different widths.  Apply the proper bit extending when evaluating template
arguments.  This mainly affected bool template arguments.

llvm-svn: 230603

If we are trying to load the scripting resource for a module whose name happens to be a Python keyword, then prefix the filename with an _ (e.g. a module named def will load _def.py)

Fixes rdar://13893506

llvm-svn: 230602

When I originally committed the statepoint docs, I left placeholders for example IR fragments.  I'm finally getting around to filling those in.  

I also added IR fragments to illustrate the usage of the PlaceSafepoints pass while I was at it.

llvm-svn: 230601

are now:

  FrontendTool -> StaticAnalyzer/Frontend -> Frontend -> StaticAnalyzer/Core

The final dependency edge here is probably removable: AnalyzerOptions (and
Analyses.def) should probably live in Basic rather than StaticAnalyzer/Core.

llvm-svn: 230600

llvm-svn: 230599

They don't conflict with MSVC flags.

llvm-svn: 230598

We'd diagnose an __assume expression which contained a function call.
This would result in us wrongly returning ExprError, causing mysterious
failures later on.

llvm-svn: 230597

It still prints "Assembling path/to/X86CompilationCallback_Win64.asm",
but linking does the same thing.

llvm-svn: 230596

llvm-svn: 230595