Part of <rdar://problem/14620988>

llvm-svn: 187710

double test(double a, double b, double c, double d) { return a<b ? c : d; }

before:
_test:
	ucomisd	%xmm0, %xmm1
	ja	LBB0_2
	movaps	%xmm3, %xmm2
LBB0_2:
	movaps	%xmm2, %xmm0

after:
_test:
	cmpltsd	%xmm1, %xmm0
	andpd	%xmm0, %xmm2
	andnpd	%xmm3, %xmm0
	orpd	%xmm2, %xmm0

Small speedup on Benchmarks/SmallPT

llvm-svn: 187706

llvm-svn: 187705

Apparently Atoms use lea for stack adjustment, which we weren't
looking for.

llvm-svn: 187704

Due to the weird and wondeful usual arithmetic conversions, some
calculations involving negative values were getting performed in
uint32_t and then promoted to int64_t, which is really not a good
idea.

Patch by Katsuhiro Ueno.

llvm-svn: 187703

llvm-svn: 187702

llvm-svn: 187701

Store them in a place that gets cleaned up properly.

llvm-svn: 187700

llvm-svn: 187699

llvm-svn: 187698

Dyld never outlives MemMgr, just put both on the stack.

llvm-svn: 187697

Using an object to do the cleanup may look like overkill, but it's safer and nicer than putting deletes everywhere.

llvm-svn: 187696

llvm-svn: 187695

Internally, the PowerPC backend names the 32-bit GPRs R[0-9]+, and names the
64-bit parent GPRs X[0-9]+. When matching inline assembly constraints with
explicit register names, on PPC64 when an i64 MVT has been requested, we need
to follow gcc's convention of using r[0-9]+ to refer to the 64-bit (parent)
registers.

At some point, we'll probably want to arrange things so that the generic code
in TargetLowering uses the AsmName fields declared in *RegisterInfo.td in order
to match these inline asm register constraints. If we do that, this change can
be reverted.

llvm-svn: 187693

Remove assertion that the verifier should catch.

llvm-svn: 187692

llvm-svn: 187687

Recent versions of the OS X linker support this but follow the existing
OS X linker convention of using an underscore in the option name, i.e.,
-export_dynamic. Rather than changing our configure scripts to check for
that alternate spelling, it is simpler to just use the compiler's -rdynamic
option and let it deal with translating that to the appropriate linker
option. One potential disadvantage of this approach is that the compiler
will typically ignore -rdynamic on platforms where it is not supported, so
the HAVE_LINK_EXPORT_DYNAMIC in config.h will not necessarily show whether
that option has any effect or not. I don't see any in-tree uses of that
macro, so I'm assuming it is OK.

llvm-svn: 187686

This is for the benefit of those of us with inferior debuggers which
do not permit member function calls on value types.

llvm-svn: 187685

attribute list is ordered by index.

Differential Revision: http://llvm-reviews.chandlerc.com/D1265

llvm-svn: 187682

Everything that comes after -- should be treated as a filename. This
enables passing in filenames that would otherwise be conflated with
command-line options.

This is especially important for clang-cl which supports options
starting with /, which are easily conflatable with Unix-style
path names.

Differential Revision: http://llvm-reviews.chandlerc.com/D1274

llvm-svn: 187675

The ExtractLoops function tries to reduce the failing test case by extracting
one or more loops from the misoptimized piece of the program. In doing this,
ExtractLoops must keep the MiscompiledFunctions vector up-to-date by ensuring
that the pointers refer to functions in the current failing program.
Unfortunately, this is not trivial because:

 - ExtractLoops is iterative, and there are several early exits (and the
   MiscompiledFunctions vector must be consistent with the current program at
every non-fatal exit point).
 - Several of the utility functions used by ExtractLoops (such as
   TestOptimizer, some of which are called through the TestFn callback
parameter, and Linker::LinkModules) delete their inputs upon success.

This change adds several updates of the MiscompiledFunctions vector at
different points. The first is after the initial call to TestMergedProgram
which checks that the loop-extracted program still works. The second is after
the call to TestFn (TestOptimizer, for example). This function will delete its
inputs (which is why the existing ExtractLoops logic cloned the inputs first).

llvm-svn: 187674

llvm-svn: 187671

llvm-svn: 187667

llvm-svn: 187666

SLPVectorizer: Fix PR16777. PHInodes may use multiple extracted values that come from different blocks.
Thanks Alexey Samsonov.

llvm-svn: 187663

llvm-svn: 187662

This patch fixes the multiple breakages on ARM test-suite after the SLP
vectorizer was introduced by default on O3. The problem was an illegal
vector type on ARMTTI::getCmpSelInstrCost() <3 x i1> which is not simple.

The guard protects this code from breaking (cause of the problems) but
doesn't fix the issue that is generating the odd vector in the first
place, which also needs to be investigated.

llvm-svn: 187658

Bugfix for making the DWARF debug strings and labels to code emitted as secrel32 instead of long opcodes (only for coff). This makes them debuggable with GDB (with fix for 64bits msvc)

llvm-svn: 187656

Patch by Daniel Sanders.

llvm-svn: 187651

llvm-svn: 187649

llvm-svn: 187646

Revert r187597, "Bugfix for making the DWARF debug strings and labels to code emitted as secrel32 instead of long opcodes (only for coff). This makes them debuggable with GDB."

It broke x86_64-win32 builder in llvm/test/DebugInfo.

llvm-svn: 187642

Patch by Benjamin Scarlet!

llvm-svn: 187641

instructions." in an attempt to bring back some bots.

This reverts commit r187609.

llvm-svn: 187638

llvm-svn: 187635

remove lines that are setting DecoderNamespace for pseudo atomic instructions.

No intended functionality change.

llvm-svn: 187632

llvm-svn: 187629

llvm-svn: 187628

llvm-svn: 187623

Function attributes are the future! So just query whether we want to realign the
stack directly from the function instead of through a random target options
structure.

llvm-svn: 187618