Allow ARM if-converter to be run after post allocation scheduling.

- This fixed a number of bugs in if-converter, tail merging, and post-allocation
  scheduler. If-converter now runs branch folding / tail merging first to
  maximize if-conversion opportunities.
- Also changed the t2IT instruction slightly. It now defines the ITSTATE
  register which is read by instructions in the IT block.
- Added Thumb2 specific hazard recognizer to ensure the scheduler doesn't
  change the instruction ordering in the IT block (since IT mask has been
  finalized). It also ensures no other instructions can be scheduled between
  instructions in the IT block.

This is not yet enabled.

llvm-svn: 106344

5 files changed, 57 insertions, 41 deletions

diff --git a/llvm/lib/CodeGen/BranchFolding.cpp b/llvm/lib/CodeGen/BranchFolding.cpp
index e440e40f05a..e17e47a662d 100644
--- a/llvm/lib/CodeGen/BranchFolding.cpp
+++ b/llvm/lib/CodeGen/BranchFolding.cpp
@@ -358,24 +358,10 @@ static unsigned ComputeCommonTailLength(MachineBasicBlock *MBB1,
 }
 
 /// ReplaceTailWithBranchTo - Delete the instruction OldInst and everything
-/// after it, replacing it with an unconditional branch to NewDest.  This
-/// returns true if OldInst's block is modified, false if NewDest is modified.
+/// after it, replacing it with an unconditional branch to NewDest.
 void BranchFolder::ReplaceTailWithBranchTo(MachineBasicBlock::iterator OldInst,
                                            MachineBasicBlock *NewDest) {
-  MachineBasicBlock *OldBB = OldInst->getParent();
-
-  // Remove all the old successors of OldBB from the CFG.
-  while (!OldBB->succ_empty())
-    OldBB->removeSuccessor(OldBB->succ_begin());
-
-  // Remove all the dead instructions from the end of OldBB.
-  OldBB->erase(OldInst, OldBB->end());
-
-  // If OldBB isn't immediately before OldBB, insert a branch to it.
-  if (++MachineFunction::iterator(OldBB) != MachineFunction::iterator(NewDest))
-    TII->InsertBranch(*OldBB, NewDest, 0, SmallVector<MachineOperand, 0>(),
-                      OldInst->getDebugLoc());
-  OldBB->addSuccessor(NewDest);
+  TII->ReplaceTailWithBranchTo(OldInst, NewDest);
   ++NumTailMerge;
 }
 
diff --git a/llvm/lib/CodeGen/IfConversion.cpp b/llvm/lib/CodeGen/IfConversion.cpp
index 36e6fc727bf..16d324735c8 100644
--- a/llvm/lib/CodeGen/IfConversion.cpp
+++ b/llvm/lib/CodeGen/IfConversion.cpp
@@ -235,6 +235,12 @@ bool IfConverter::runOnMachineFunction(MachineFunction &MF) {
   TRI = MF.getTarget().getRegisterInfo();
   if (!TII) return false;
 
+  // Tail merge tend to expose more if-conversion opportunities.
+  BranchFolder BF(true);
+  bool BFChange = BF.OptimizeFunction(MF, TII,
+                                   MF.getTarget().getRegisterInfo(),
+                                   getAnalysisIfAvailable<MachineModuleInfo>());
+
   DEBUG(dbgs() << "\nIfcvt: function (" << ++FnNum <<  ") \'"
                << MF.getFunction()->getName() << "\'");
 
@@ -376,6 +382,7 @@ bool IfConverter::runOnMachineFunction(MachineFunction &MF) {
                         getAnalysisIfAvailable<MachineModuleInfo>());
   }
 
+  MadeChange |= BFChange;
   return MadeChange;
 }
 
diff --git a/llvm/lib/CodeGen/MachineVerifier.cpp b/llvm/lib/CodeGen/MachineVerifier.cpp
index 8baf01c9073..2297c908b1e 100644
--- a/llvm/lib/CodeGen/MachineVerifier.cpp
+++ b/llvm/lib/CodeGen/MachineVerifier.cpp
@@ -390,7 +390,8 @@ MachineVerifier::visitMachineBasicBlockBefore(const MachineBasicBlock *MBB) {
         report("MBB exits via unconditional fall-through but its successor "
                "differs from its CFG successor!", MBB);
       }
-      if (!MBB->empty() && MBB->back().getDesc().isBarrier()) {
+      if (!MBB->empty() && MBB->back().getDesc().isBarrier() &&
+          !TII->isPredicated(&MBB->back())) {
         report("MBB exits via unconditional fall-through but ends with a "
                "barrier instruction!", MBB);
       }
diff --git a/llvm/lib/CodeGen/PostRASchedulerList.cpp b/llvm/lib/CodeGen/PostRASchedulerList.cpp
index f3c725da9e6..4af8e07f348 100644
--- a/llvm/lib/CodeGen/PostRASchedulerList.cpp
+++ b/llvm/lib/CodeGen/PostRASchedulerList.cpp
@@ -79,6 +79,7 @@ AntiDepBreaker::~AntiDepBreaker() { }
 namespace {
   class PostRAScheduler : public MachineFunctionPass {
     AliasAnalysis *AA;
+    const TargetInstrInfo *TII;
     CodeGenOpt::Level OptLevel;
 
   public:
@@ -181,30 +182,9 @@ namespace {
   };
 }
 
-/// isSchedulingBoundary - Test if the given instruction should be
-/// considered a scheduling boundary. This primarily includes labels
-/// and terminators.
-///
-static bool isSchedulingBoundary(const MachineInstr *MI,
-                                 const MachineFunction &MF) {
-  // Terminators and labels can't be scheduled around.
-  if (MI->getDesc().isTerminator() || MI->isLabel())
-    return true;
-
-  // Don't attempt to schedule around any instruction that defines
-  // a stack-oriented pointer, as it's unlikely to be profitable. This
-  // saves compile time, because it doesn't require every single
-  // stack slot reference to depend on the instruction that does the
-  // modification.
-  const TargetLowering &TLI = *MF.getTarget().getTargetLowering();
-  if (MI->definesRegister(TLI.getStackPointerRegisterToSaveRestore()))
-    return true;
-
-  return false;
-}
-
 bool PostRAScheduler::runOnMachineFunction(MachineFunction &Fn) {
   AA = &getAnalysis<AliasAnalysis>();
+  TII = Fn.getTarget().getInstrInfo();
 
   // Check for explicit enable/disable of post-ra scheduling.
   TargetSubtarget::AntiDepBreakMode AntiDepMode = TargetSubtarget::ANTIDEP_NONE;
@@ -265,8 +245,8 @@ bool PostRAScheduler::runOnMachineFunction(MachineFunction &Fn) {
     MachineBasicBlock::iterator Current = MBB->end();
     unsigned Count = MBB->size(), CurrentCount = Count;
     for (MachineBasicBlock::iterator I = Current; I != MBB->begin(); ) {
-      MachineInstr *MI = prior(I);
-      if (isSchedulingBoundary(MI, Fn)) {
+      MachineInstr *MI = llvm::prior(I);
+      if (TII->isSchedulingBoundary(MI, MBB, Fn)) {
         Scheduler.Run(MBB, I, Current, CurrentCount);
         Scheduler.EmitSchedule();
         Current = MI;
diff --git a/llvm/lib/CodeGen/TargetInstrInfoImpl.cpp b/llvm/lib/CodeGen/TargetInstrInfoImpl.cpp
index 1641c7162ab..53f3ee84910 100644
--- a/llvm/lib/CodeGen/TargetInstrInfoImpl.cpp
+++ b/llvm/lib/CodeGen/TargetInstrInfoImpl.cpp
@@ -13,6 +13,7 @@
 //===----------------------------------------------------------------------===//
 
 #include "llvm/Target/TargetInstrInfo.h"
+#include "llvm/Target/TargetLowering.h"
 #include "llvm/Target/TargetMachine.h"
 #include "llvm/Target/TargetRegisterInfo.h"
 #include "llvm/ADT/SmallVector.h"
@@ -27,6 +28,25 @@
 #include "llvm/Support/raw_ostream.h"
 using namespace llvm;
 
+void
+TargetInstrInfoImpl::ReplaceTailWithBranchTo(MachineBasicBlock::iterator Tail,
+                                             MachineBasicBlock *NewDest) const {
+  MachineBasicBlock *MBB = Tail->getParent();
+
+  // Remove all the old successors of MBB from the CFG.
+  while (!MBB->succ_empty())
+    MBB->removeSuccessor(MBB->succ_begin());
+
+  // Remove all the dead instructions from the end of MBB.
+  MBB->erase(Tail, MBB->end());
+
+  // If MBB isn't immediately before MBB, insert a branch to it.
+  if (++MachineFunction::iterator(MBB) != MachineFunction::iterator(NewDest))
+    InsertBranch(*MBB, NewDest, 0, SmallVector<MachineOperand, 0>(),
+                 Tail->getDebugLoc());
+  MBB->addSuccessor(NewDest);
+}
+
 // commuteInstruction - The default implementation of this method just exchanges
 // the two operands returned by findCommutedOpIndices.
 MachineInstr *TargetInstrInfoImpl::commuteInstruction(MachineInstr *MI,
@@ -316,6 +336,28 @@ isReallyTriviallyReMaterializableGeneric(const MachineInstr *MI,
   return true;
 }
 
+/// isSchedulingBoundary - Test if the given instruction should be
+/// considered a scheduling boundary. This primarily includes labels
+/// and terminators.
+bool TargetInstrInfoImpl::isSchedulingBoundary(const MachineInstr *MI,
+                                               const MachineBasicBlock *MBB,
+                                               const MachineFunction &MF) const{
+  // Terminators and labels can't be scheduled around.
+  if (MI->getDesc().isTerminator() || MI->isLabel())
+    return true;
+
+  // Don't attempt to schedule around any instruction that defines
+  // a stack-oriented pointer, as it's unlikely to be profitable. This
+  // saves compile time, because it doesn't require every single
+  // stack slot reference to depend on the instruction that does the
+  // modification.
+  const TargetLowering &TLI = *MF.getTarget().getTargetLowering();
+  if (MI->definesRegister(TLI.getStackPointerRegisterToSaveRestore()))
+    return true;
+
+  return false;
+}
+
 // Default implementation of CreateTargetPostRAHazardRecognizer.
 ScheduleHazardRecognizer *TargetInstrInfoImpl::
 CreateTargetPostRAHazardRecognizer(const InstrItineraryData &II) const {