Enable PPC CTR loop formation by default.

Thanks to Jakob's help, this now causes no new test suite failures!

Over the entire test suite, this gives an average 1% speedup. The largest speedups are:
SingleSource/Benchmarks/Misc/pi - 108%
SingleSource/Benchmarks/CoyoteBench/lpbench - 54%
MultiSource/Benchmarks/Prolangs-C/unix-smail/unix-smail - 50%
SingleSource/Benchmarks/Shootout/ary3 - 32%
SingleSource/Benchmarks/Shootout-C++/matrix - 30%

The largest slowdowns are:
MultiSource/Benchmarks/mediabench/gsm/toast/toast - -30%
MultiSource/Benchmarks/Prolangs-C/bison/mybison - -25%
MultiSource/Benchmarks/BitBench/uuencode/uuencode - -22%
MultiSource/Applications/d/make_dparser - -14%
SingleSource/Benchmarks/Shootout-C++/ary - -13%

In light of these slowdowns, additional profiling work is obviously needed!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158223 91177308-0d34-0410-b5e6-96231b3b80d8
diff --git a/lib/Target/PowerPC/PPCTargetMachine.cpp b/lib/Target/PowerPC/PPCTargetMachine.cpp
index bcab83c..b9c703e 100644
--- a/lib/Target/PowerPC/PPCTargetMachine.cpp
+++ b/lib/Target/PowerPC/PPCTargetMachine.cpp
@@ -23,8 +23,8 @@
 using namespace llvm;
 
 static cl::
-opt<bool> EnableCTRLoops("enable-ppc-ctrloops", cl::Hidden,
-                        cl::desc("Enable CTR loops for PPC"));
+opt<bool> DisableCTRLoops("disable-ppc-ctrloops", cl::Hidden,
+                        cl::desc("Disable CTR loops for PPC"));
 
 extern "C" void LLVMInitializePowerPCTarget() {
   // Register the targets
@@ -103,9 +103,7 @@
 }
 
 bool PPCPassConfig::addPreRegAlloc() {
-  // FIXME: Once this can be enabled by default, this condition should read:
-  // if (!DisableCTRLoops && getOptLevel() != CodeGenOpt::None)
-  if (EnableCTRLoops)
+  if (!DisableCTRLoops && getOptLevel() != CodeGenOpt::None)
     PM->add(createPPCCTRLoops());
 
   return false;